Smart Innovation, Systems and Technologies 202

Yuzo Iano · Rangel Arthur ·

Osamu Saotome · Guillermo Kemper ·
Ana Carolina Borges Monteiro Editors

Proceedings of the
5th Brazilian
Technology
Symposium
Emerging Trends, Issues, and Challenges
in the Brazilian Technology, Volume 2
Smart Innovation, Systems and Technologies

Volume 202

Series Editors
Robert J. Howlett, Bournemouth University and KES International,
Shoreham-by-sea, UK
Lakhmi C. Jain, Faculty of Engineering and Information Technology,
Centre for Artificial Intelligence, University of Technology Sydney,
Sydney, NSW, Australia
The Smart Innovation, Systems and Technologies book series encompasses the
topics of knowledge, intelligence, innovation and sustainability. The aim of the
series is to make available a platform for the publication of books on all aspects of
single and multi-disciplinary research on these themes in order to make the latest
results available in a readily-accessible form. Volumes on interdisciplinary research
combining two or more of these areas is particularly sought.
The series covers systems and paradigms that employ knowledge and intelligence
in a broad sense. Its scope is systems having embedded knowledge and intelligence,
which may be applied to the solution of world problems in industry, the environment
and the community. It also focusses on the knowledge-transfer methodologies and
innovation strategies employed to make this happen effectively. The combination of
intelligent systems tools and a broad range of applications introduces a need for a
synergy of disciplines from science, technology, business and the humanities. The
series will include conference proceedings, edited collections, monographs, hand-
books, reference books, and other relevant types of book in areas of science and
technology where smart systems and technologies can offer innovative solutions.
High quality content is an essential feature for all book proposals accepted for the
series. It is expected that editors of all accepted volumes will ensure that
contributions are subjected to an appropriate level of reviewing process and adhere
to KES quality principles.
Indexed by SCOPUS, EI Compendex, INSPEC, WTI Frankfurt eG, zbMATH,
Japanese Science and Technology Agency (JST), SCImago, DBLP.
All books published in the series are submitted for consideration in Web of Science.

More information about this series at http://www.springer.com/series/8767

Yuzo Iano Rangel Arthur
• •

Osamu Saotome Guillermo Kemper

• •

Ana Carolina Borges Monteiro

Editors

Proceedings of the 5th

Brazilian Technology
Symposium
Emerging Trends, Issues, and Challenges
in the Brazilian Technology, Volume 2

123
Editors
Yuzo Iano Rangel Arthur
School of Electrical and Computer School of Technology
Engineering UNICAMP
UNICAMP Campinas, São Paulo, Brazil
Limeira, São Paulo, Brazil
Guillermo Kemper
Osamu Saotome Universidad Peruana de Ciencias Aplicada
Divisão de Engenharia Eletrônica Lima, Peru
Instituto Tecnológico de Aeronáutica
São José dos Campos, São Paulo, Brazil

Ana Carolina Borges Monteiro

School of Electrical and Computer
Engineering
UNICAMP
Campinas, Brazil

Associate-Editors
Reinaldo Padilha França Diego Pajuelo Castro
School of Electrical and Computer School of Electrical and Computer
Engineering Engineering
UNICAMP UNICAMP
Campinas, São Paulo, Brazil Campinas, São Paulo, Brazil

ISSN 2190-3018 ISSN 2190-3026 (electronic)

Smart Innovation, Systems and Technologies
ISBN 978-3-030-57565-6 ISBN 978-3-030-57566-3 (eBook)
https://doi.org/10.1007/978-3-030-57566-3
© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature
Switzerland AG 2021
This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of
illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and
transmission or information storage and retrieval, electronic adaptation, computer software, or by similar
or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, expressed or implied, with respect to the material contained
herein or for any errors or omissions that may have been made. The publisher remains neutral with regard
to jurisdictional claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Committees

Organizing and Executive Committee

Yuzo Iano, LCV/DECOM/FEEC/UNICAMP, General Chair BTSym & WSGE

Osamu Saotome, ITA, Associate-General Chair BTSym
Rangel Arthur, FT/UNICAMP, Vice-General Chair BTSym
Carlos Arturo Raymundo Ibáñez, UPC, Associate-General Chair BTSym
Guillermo Leopoldo Kemper Vásquez, UPC, Associate-General Chair BTSym
Fernando Hashimoto, DMU/UNICAMP, Associate-General Chair BTSym
Rogério Bazi, PUC-Campinas, Associate-General Chair BTSym
Gabriel Gomes de Oliveira, LCV/DECOM/FEEC/UNICAMP, Technical Program
and Finance Chair
Ana Carolina Borges Monteiro, LCV/DECOM/FEEC/UNICAMP, Proceedings
Chair
Reinaldo Padilha França, LCV/DECOM/FEEC/UNICAMP, Proceedings Chair
Telmo Cardoso Lustosa, LCV/DECOM/FEEC/UNICAMP, Local Arrangements
Chair
Alysson Gomes de Oliveira, LCV/DECOM/FEEC/UNICAMP, Marketing Chair
Diego Arturo Pajuelo Castro, LCV/DECOM/FEEC/UNICAMP, Proceedings Chair
David Minango, LCV/DECOM/FEEC/UNICAMP, Registration Chair
Lisber Arana, LCV/DECOM/FEEC/UNICAMP, Institutional Relationship Chair
Alex R. Ruelas, LCV/DECOM/FEEC/UNICAMP, Institutional Relationship Chair
Abel Dueñas Rodríguez, LCV/DECOM/FEEC/UNICAMP, Media Chair

Scientific and Academic Committee

Osamu Saotome, ITA

Luiz Cézar Martini, DECOM/FEEC/UNICAMP
David Bianchini, PUC-Campinas

v
vi Committees

Luis Geraldo Pedroso Meloni, DECOM/FEEC/UNICAMP

Ana Cláudia Seixas, PUC-Campinas Cláudia Seixas
Cristiano Akamine, Universidade Presbiteriana Mackenzie PUC-Campinas
Celso Iwata Frison, PUC/Minas-Poços de Caldas
Luiz Vicente Figueira de Mello Filho, Universidade Presbiteriana Mackenzie
Guillermo Leopoldo Kemper Vásquez, USMP & UNI-INICTEL
Lia Toledo Moreira Mota, PUC-Campinas
Lucas Heitzmann Gabrielli, DECOM/FEEC/UNICAMP
Edgard Luciano Oliveira da Silva, EST/UEA
Talía Simões dos Santos, FT/UNICAMP
Janito Vaqueiro Ferreira, DMC/FEM/UNICAMP
Vlademir de Jesus Silva Oliveira, UNEMAT/Sinop
Navid Razmjooy, Tafresh University, Tafresh, Iran
Hugo Enrique Hernandez Figueroa, DECOM/FEEC/UNICAMP
Antônio José da Silva Neto, IPRJ/UERJ
Bruno Sanches Masiero, DECOM/FEEC/UNICAMP
Marcos Antonio do Nascimento Guimarães, UNIP/CAMPINAS, JUNDIAÍ
Alessandra Akkari, Universidade Presbiteriana Mackenzie
Maria Thereza de Moraes Gomes Rosa, Universidade Presbiteriana Mackenzie
Angela del Pilar Flores Granados, FEA/UNICAMP
Paolo Bacega, Faculdade Anhanguera
Marcos Fernando Espindola, IFSP São Paulo
Polyane Alves Santos, Instituto Federal Da Bahia
Vicente Idalberto Becerra Sablón, USF/UNIMEP
Grimaldo Wilfredo Quispe Santivañez, UNAAT, Tarma, Perú
Jessie Leila Bravo Jaico, UNPRG, Lambayeque, Perú
Ernesto Karlo Celi Arévalo, UNPRG, Lambayeque, Perú
Nestor Adolfo Mamani Macedo - UNMSM, Lima, Perú
Luz María Paucar-Menacho, UNS, Ancash, Perú
Edwin Alberto Valencia Castillo, UNC, Cajamarca, Perú
Christian Del Carpio, UPC, Lima, Perú
Gustavo Huaman Bustamante, Inictel/UNI, Lima, Perú
Itamar Salazar Reque, Inictel/UNI, Lima, Perú

Technical Reviewers Committee

Rangel Arthur, INOVA/FT/UNICAMP

Ana Carolina Borges Monteiro, LCV/DECOM/FEEC/UNICAMP
Reinaldo Padilha França, LCV/DECOM/FEEC/UNICAMP
Adão Boava - Universidade Federal de Santa Catarina, UFSC
Amilton Lamas, PUC-Campinas
Agord de Matos Pinto Júnior, DESIF/FEEC/UNICAMP
Angela del Pilar Flores Granados, FEA/UNICAMP
Committees vii

Joaquim Marcos Santa Rita da Silva, Instituto Nacional de Telecomunicações

Alysson Gomes de Oliveira, LCV/DECOM/FEEC/UNICAMP
José Alexandre Nalon, Centro Universitário Salesiano São Paulo, UNISAL
Murilo Cesar Perin Briganti, LCV/DECOM/FEEC/UNICAMP
Luigi Ciambarella Filho, Universidade Veiga de Almeida/Develop Biotechnology
Lucas Alves, LCV/DECOM/FEEC/UNICAMP
Ingrid Araujo Sampaio, Universidade Estadual de Campinas
Denise Helena Lombardo Ferreira, PUC-Campinas
Hermes José Loschi, LCV/DECOM/FEEC/UNICAMP
Daniel Rodrigues Ferraz Izario, LCV/DECOM/FEEC/UNICAMP
Mariana Carvalho, DCA/FEEC/UNICAMP
Diego Pajuelo, LCV/DECOM/FEEC/UNICAMP
Douglas Aguiar do Nascimento, LCV/DECOM/FEEC/UNICAMP
Edson José Gonçalves, LCV/DECOM/FEEC/UNICAMP
Alessandra Akkari, Universidade Presbiteriana Mackenzie
Polyane Alves Santos, Instituto Federal Da Bahia
Kelem Christine Pereira Jordão, LCV/DECOM/FEEC/UNICAMP
Josué Marcos de Moura Cardoso, LCV/DECOM/FEEC/UNICAMP
Giovanni Moura de Holanda, FITec, Technological Innovations, Brazil
Euclides Lourenço Chuma, LCV/DECOM/FEEC/UNICAMP
Gabriel Gomes de Oliveira, LCV/DECOM/FEEC/UNICAMP
José Yauri, DCA/FEEC/UNICAMP
Julío Humberto León Ruiz, LCV/DECOM/FEEC/UNICAMP
Navid Razmjooy, Tafresh University, Tafresh, Iran
Antônio José da Silva Neto, IPRJ/UERJ
Raphael Ronald, CTI (Centro de Tecnologia da Informação)
Raquel Jahara Lobosco, UFRJ/Macaé
Necesio Gomes Costa, UFRJ
Claudia Maria Moraes Santos, UNIVAP
Vitor Chaves de Oliveira, IFSP São Paulo/Mackenzie
Gustavo Mesones, UPC, Lima, Perú
Julio Ronceros, UPC, Lima, Perú
Kalun Lau, UPC, Lima, Perú
Manuel Márquez, UPC, Lima, Perú
Mirko Klusmann, UPC, Lima, Perú
Nikolai Vinces Ramos, UPC, Lima, Perú
Sergio Salas, UPC, Lima, Perú

Sponsor

Beta Telecommunications
Foreword

It is with great satisfaction that I write this foreword to the book, Proceedings of the
5th Brazilian Technology Symposium (BTSym’19), held at the Mackenzie
Presbyterian University, SP, Brazil, in October 2019. This event is in its fifth
edition and has consolidated to become an excellent opportunity for researchers,
teachers, and students to present and discuss the results of their research works.
The 2019 edition of BTSym is characterized by the broad scope of exposed
areas, with papers dealing with current and important topics for Brazilian and world
technological development, including subjects related to the various branches of
human, engineering, smart and sustainable future of cities, architecture, biomedical,
and computer science.
Events such as BTSym are an essential part of the research and innovation
process. Firstly, these events contribute to the promotion of research activities,
which are key to a country's technological development. The dissemination of
research results, as promoted by BTSym, contributions to the transformation of
research findings into technological innovation. In addition, these events facilitate
the sharing of findings, eventually leading to the formation of research networks,
which accelerate the achievement of new results.
Therefore, I congratulate the BTSym General Chair, Prof. Dr. Yuzo Iano, and his
group of collaborators for the important initiative of organizing the BTSym 2019
and for providing the opportunity for authors to present their work to a wide
audience through this publication.
Finally, I congratulate the authors for the high-quality research papers and
scientific experiments presented in this proceedings.

Reinaldo Padilha França

Researcher and Associate-Editor
University of Campinas
Campinas, Brazil

ix
Preface

This book contains the Proceedings of the 5th Brazilian Technology Symposium,
an event held in Campinas, SP, Brazil, in October 2019. This book was divided into
two parts: the first part on Emerging Trends in Health, Natural and Physical
Sciences and the second part on Emerging Trends in Humans and Urban Sciences.
The Brazilian Technology Symposium is an excellent forum for presentations
and discussions of the latest results of projects and development research, in several
areas of knowledge, in scientific and technological scope, including smart designs,
sustainability, inclusion, future technologies, architecture and urbanism, computer
science, information science, industrial design, aerospace engineering, agricultural
engineering, biomedical engineering, civil engineering, control, and automation
engineering, production engineering, electrical engineering, mechanical engineer-
ing, naval and oceanic engineering, nuclear engineering, chemical engineering,
probability, and statistics.
This event seeks to bring together researchers, students, and professionals from
the industrial and academic sectors, seeking to create and/or strengthen the linkages
between issues of joint interest. Participants were invited to submit research papers
with methodologies and results achieved in scientific level research projects,
completion of course work for graduation, dissertations, and theses.
The 56 full chapters accepted for this book were selected from 210 submissions,
and in each case, the authors were shepherded by an experienced researcher, with a
rigorous peer review process. Among the main topics covered in this book, we can
highlight health, medicine, chemistry, physics, physiology, cytology, histology,
hematology, gynecology, virology, dermatology, pathology, health care, molecular
biology, microbiology, biochemistry, biomedicine, clinical analysis, diagnostic
methods, people management, ergonomics, agriculture, agronomy, food, fuel,
transport, bioeconomics, natural resources, environment, energy, hydraulics, tech-
nology, languages, statistical methods, digital image processing, digital signal

xi
xii Preface

processing, CNN, deep learning, e-Health, m-Health, simulation, IoT, telemedicine,

wearable technology, artificial intelligence, and much more.
We hope you enjoy and take advantage of this book and feel motivated to submit
your papers in future to Brazilian Technology Symposium.

Campinas, Brazil Ana Carolina Borges Monteiro

Proceedings Chair of Brazilian Technology Symposium
Acknowledgments

Our appreciation goes to a lot of colleagues and friends who assisted in the
development of this book, Proceedings of the 5th Brazilian Technology Symposium
(BTSym’19).
First of all, I would like to thank all the members of the Organizing and
Executive Committee, for the commitment throughout the year, several meetings
were held, and many challenges were overcome for the accomplishment of the
BTSym 2019.
Also, and with great merit, I would like to thank all the members of the Scientific
and Academic Committee and Technical Reviewers Committee, for their excellent
work, which was essential to ensure the quality of our peer review process, and
collaborating with the visibility and technical quality of the BTSym 2019.
The Brazilian Technology Symposium is an event created by the Laboratory of
Visual Communications of the Faculty of Electrical and Computer Engineering
of the University of Campinas (UNICAMP). In this way, I would like to thank the
Mackenzie Presbyterian University, especially for the support and hosting of the
BTSym 2019, which was critical for the success of their accomplishment.
Beta Telecommunications played a crucial role in holding the BTSym 2019, in
the same way as the Pro-Rectory of Extension and Culture at the University of
Campinas, especially Prof. Fernando Augusto de Almeida Hashimoto, due to
financial support from it, and it was possible to consolidate with quality many
BTSym 2019 organization aspects, which ensured the quality to support the authors
and speakers.
Finally, I thank all the authors for their participation in the BTSym 2019.
I sincerely hope to have provided an experience that was very useful and enriching
in the personal and professional lives of everyone, and my special thanks go to Prof.
Rangel Arthur. Due to the continuous support in activities that enrich the academic
environment through research directed to innovative technological development.

xiii
xiv Acknowledgments

He is a generous, charismatic, dedicated, determined, and fair person, being an

excellent fellow in academic research that certainly contributes a lot to BTSym,
who has a lot to thank Prof Rangel Arthur for.

Prof. Yuzo Iano

General Chair of Brazilian Technology Symposium
Contents

Emerging Trends in Health, Natural and Physical Sciences

Algorithm Oriented to the Detection of the Level of Blood Filling
in Venipuncture Tubes Based on Digital Image Processing . . . . . . . . . . 3
Jorge Castillo, Nelson Apfata, and Guillermo Kemper
Computational Fluid Dynamics Analysis of Blood Rheology . . . . . . . . . 17
Guilherme S. Souza, Enzo D. Giustina, Marcus V. Carvalho,
and Raquel J. Lobosco
Wearable Technology for Presumptive Diagnosis of High Blood
Pressure Based on Risk Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Eithel Josue Meza Prada, Helgar Miguel Angel Herrera Agullar,
Jimmy Armas-Aguirre, and Paola A. Gonzalez
Cytology and Hematology: A Review of the Fundamental Principles . . . 35
Ana Carolina Borges Monteiro, Yuzo Iano, Reinaldo Padilha França,
and Rangel Arthur
Hematology: A Review of the Main Methodologies of Clinical
Analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
Ana Carolina Borges Monteiro, Yuzo Iano, Reinaldo Padilha França,
and Rangel Arthur
An Algorithm Oriented at Obtaining the Molecular Weight
and Concentration of DNA Samples in Agarose Gel Images . . . . . . . . . 61
Christian del Carpio, Marcelo Garcia, Julio Elías, David Laván,
and Guillermo Kemper
Classification of Dermoscopy Skin Images with the Application
of Deep Learning Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
Pablo David Minango Negrete, Yuzo Iano, Ana Carolina Borges Monteiro,
Reinaldo Padilha França, Gabriel Gomes de Oliveira, and Diego Pajuelo

xv
xvi Contents

Association Between Human Papillomavirus Infection and Anus

Cancer Precursor Lesions: A Review . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
Ana Carolina Borges Monteiro, Yuzo Iano, Reinaldo Padilha França,
Rangel Arthur, Pablo David Minango Negrete, Diego Pajuelo,
Gabriel Gomes de Oliveira, and Lisber Arana Hisnostroza
A New Approach for ECG Recording in Rats: An Autonomic Nervous
System Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
Raphael Santos do Nascimento, Fernando da Silva Fiorin,
Adair Roberto Soares Santos, Luiz Fernando Freire Royes,
and Jefferson Luiz Brum Marques
Algorithm for Detection of Raising Eyebrows and Jaw Clenching
Artifacts in EEG Signals Using Neurosky Mindwave Headset . . . . . . . . 99
Luis Vélez and Guillermo Kemper
Correspondence Between TOVA Test Results and Characteristics
of EEG Signals Acquired Through the Muse Sensor in Positions
AF7–AF8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
Ober Castillo, Simy Sotomayor, Guillermo Kemper,
and Vincent Clement
Mobile Robot to Assist in Therapies for Children with Autism . . . . . . . 121
Madai Keila Taype Mateo and Grimaldo Wilfredo Quispe Santivañez
A Smart Home Control Prototype Using a P300-Based
Brain–Computer Interface for Post-stroke Patients . . . . . . . . . . . . . . . . 131
Sergio A. Cortez, Christian Flores, and Javier Andreu-Perez
Electronic Ecomap as an Instrument to Improve Physician–Family
Interaction in Preventive Medicine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
Néstor Mamani-Macedo, Julio Raime, Luzmila Pro, Jaime Pariona,
and Olga Solano
Intelligent Management Model in an IoT Environment Applied for
the Healthcare Sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
Jessie Bravo and Roger Alarcón
Evaluation Method of Variables and Indicators for Surgery Block
Process Using Process Mining and Data Visualization . . . . . . . . . . . . . . 159
Piero Rojas-Candio, Arturo Villantoy-Pasapera, Jimmy Armas-Aguirre,
and Santiago Aguirre-Mayorga
A Technological Solution to Identify the Level of Risk to Be Diagnosed
with Type 2 Diabetes Mellitus Using Wearables . . . . . . . . . . . . . . . . . . . 169
Daniela Nuñovero, Ernesto Rodríguez, Jimmy Armas, and Paola Gonzalez
Mamey Apple Peel for Cr3+ Removal from Contaminated Waters . . . . 177
Nathali Cotrina and Ricardo Vejarano
Contents xvii

Method of Estimating River Levels with Reflective Tapes Using

Artificial Vision Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
Lidia E. López Huamán, Marco Paul E. Apolinario Lainez,
and Samuel G. Huamán Bustamante
Design and Implementation of a Hydro-Energetic System in Water
Distribution Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
Lenin Jiménez, Cecilia Luna, Manuel Quiñones-Cuenca,
and Holger Benavides
Hydraulic Fill Assessment Model Using Weathered Granitoids Based
on Analytical Solutions to Mitigate Rock Mass Instability in
Conventional Underground Mining . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
Cristhian Portocarrero-Urdanivia, Angela Ochoa-Cuentas,
Luis Arauzo-Gallardo, and Carlos Raymundo
A Study on Investigations Carried Out in Dams from the Perspective
of Risk Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
Gabriel Gomes de Oliveira, Yuzo Iano, Ana Carolina Borges Monteiro,
Reinaldo Padilha França, Diego Pajuelo,
and Pablo David Minango Negrete
Numerical Simulation of Hydrodynamic Conditions in Rivers Facing
Extreme Events Due to the “El Niño” Phenomenon . . . . . . . . . . . . . . . . 235
Gabriela Alvarez, Alvaro Moreno, Emanuel Guzmán, and Sissi Santos
Dynamic Amplification Factor Proposal for Seismic Resistant Design
of Tall Buildings with Rigid Core Structural System . . . . . . . . . . . . . . . 245
Eder Quezada, Yaneth Serrano, and Guillermo Huaco
A Mechanical Development of a Dry Cell to Obtain HHO
from Water Electrolysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257
Gustavo Salazar, Wilmer Solis, and Leonardo Vinces
An Automated System for the Stage of Hydrolysis and Filtration in the
Extraction of Pectin from the Cocoa Shell . . . . . . . . . . . . . . . . . . . . . . . 267
Maritza Ccencho, Valeria Quijada, and Leonardo Vinces
Automatic Leaf Segmentation from Images Taken Under
Uncontrolled Conditions Using Convolutional Neural Networks . . . . . . 277
Itamar Franco Salazar-Reque and Samuel Gustavo Huamán Bustamante
Method of Anomalies Detection in Persea Americana Leaves with
Thermal and NGRDI Imagery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
Bruno D. Rivadeneyra Bustamante and Samuel G. Huamán Bustamante
Development of a Machine to Control the Level of Washing in Panca
Chili Seeds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
Anthony De La Cruz, Jaime Cardenas, and Leonardo Vinces
xviii Contents

Bacteriological Contaminants Detection in Some Consumed Raw

Vegetables Based on Spectrometry in Ayacucho Town, Peru . . . . . . . . . 309
José Yauri, Manuel Lagos, Rousell Vidalón, Julio Cárdenas, Kay Jerí,
Lisber Arana, and Yuzo Iano
Feasibility in Using Banana Flour in Bread Production: Centesimal
and Sensory Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
Nathália Rafaela Santos de Carvalho, Maria Thereza de Moraes Gomes
Rosa, Isabela Francabandiera Tornizielo,
and Daniela Helena Pelegrine Guimarães
Spectral Behavior of Maize, Rice, Soy, and Oat Crops Using
Multi-spectral Images from Sentinel-2 . . . . . . . . . . . . . . . . . . . . . . . . . . 327
Lady Angulo and Sergio Pamboukian

Emerging Trends in Humans and Urban Sciences

Production Process of Bioethanol Fuel Using Supported
Saccharomyces Cerevisiae Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
L. C. Fardelone, T. S. Bella de Jesus, G. P. Valença, J. R. Nunhez,
J. A. R. Rodrigues, and P. J. S. Moran
An Automatic Biodiesel Decanting System for the Optimization
of Glycerin Separation Time by Applying Electric Field and
Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349
Kevin Bulnes, Diego Paredes, and Leonardo Vinces
Prototype Proposal for Control and Inspection of Gas Stations . . . . . . . 357
Lucas Eder de Araújo Sousa, Felipe Silveira Stopiglia,
Luiz Vicente Figueira de Mello Filho,
Daniela Helena Pelegrine Guimarães, and Maria Thereza de Moraes
Gomes Rosa
Modeling of Solid-Liquid Equilibrium in Nitrobenzene
and Para-Xylene Mixtures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 367
Michelli Fontana, Caroline dos Santos Ribas, Paulo Henrique Schuck,
Rafael Lopes Turino, and André Zuber
X-Ray Characterization of TiO2 Obtained by Chemical Route for
Photocatalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 377
Tiago Siqueira Lima, Wesley Rafael Penteado,
André Luis de Castro Peixoto, and Ademir Geraldo Cavallari Costalonga
Co-authors Networks in Adsorption Refrigeration and
Air-Conditioning with Solar Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . 383
Carla Abregú Marcos, Abraham Sopla Maslucán, Miguel Cano Lengua,
and José C. Alvarez
Contents xix

Microsimulation of Public Transport Stops for the Optimization

of Waiting Times for Users Using the Social Force Model . . . . . . . . . . . 391
Francis Mendoza, Mayling Tong, Manuel Silvera,
and Fernando Campos
Mathematical Model for the Improvement of Vertical Traffic
at Universities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 401
Julio C. Lazo Lozano and Grimaldo Wilfredo Quispe Santivañez
A Novel Dataset for the Transport Sector in a Province of Peru . . . . . . 409
Miguel Arango Guerrero and Pedro Shiguihara Juárez
Development of a Mobile Application for English Language Learning
Through Corrective Feedback . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 417
Adriana Guanuche, Gustavo Caiza, Osana Eiriz, and Roberto Espí
Influence of the Chinese Culture on the Negotiation Process
Peru-China . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 427
Nicole Alexandra Acevedo Cueva
Phone Calls Speech-to-Text: A Comparison Between APIs
for the Portuguese Language . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 437
Nilton M. Iinuma and Massaki de O. Igarashi
Production of Cantonese Lexical Tones by Native Speakers
of Brazilian Portuguese: A Comparative Analysis . . . . . . . . . . . . . . . . . 447
João Vítor Possamai de Menezes and Adriano Vilela Barbosa
Smart Campus IoT Guidance System for Visitors Based
on Bayesian Filters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 463
Alvaro Aspilcueta Narvaez, Dennis Núñez Fernández,
Segundo Gamarra Quispe, and Domingo Lazo Ochoa
Research of the Academic Performance of University Students
Through Statistical Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 473
Olga Solano Dávila, Luis Nuñez Lira, Olga Bolaños Solano,
and Néstor Mamani-Macedo
Model of Sustainable Digital Transformation Focused on
Organizational and Technological Culture for Academic Management
in Public Higher Education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 483
Jessie Bravo, Janet Aquino, Roger Alarcón, and Nilton Germán
University of Brasilia’s Potential for Bioeconomy Services
and Innovation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 493
Flávio Duque Estrada Soares Pereira and Wagner Augusto Fischer
A Proposal for Modeling of the Management of Talent Recruitment
and Training in Peruvian Sports Centers . . . . . . . . . . . . . . . . . . . . . . . . 505
Arturo Laredo, Luzmila Pro Concepcion, and Nestor Mamani-Macedo
xx Contents

Mentoring Process: An Assessment of Career, Psychosocial Functions

and Mentor Role Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 515
Neusa Maria Bastos Fernandes dos Santos, Mariana Juer,
Igor Polezi Munhoz, Alessandra Cristina Santos Akkari,
Rodrigo Guimarães Motta, and Marianna Konyosi Miyashiro
Compensation Program: A Quantitative Study of the Top
Management Team’s Level of Knowledge and Satisfaction . . . . . . . . . . 527
Neusa Maria Bastos Fernandes dos Santos, Rafael Maia Chagas,
Alessandra Cristina Santos Akkari, Igor Polezi Munhoz,
Fernando Fukunaga, and Ligia Kaori Matsumoto Hirano
Physical Ergonomics Applied to the Administrative Sector
and the Factory Floor: The Case of a Food Industry . . . . . . . . . . . . . . . 537
Kélvin Santos Aguiar, Maria Thereza de Moraes Gomes Rosa,
Luiz Vicente Figueira de Mello Filho, João Carlos Gabriel,
and Alessandra Cristina Santos Akkari
Evaluation of Illuminance in Workplace: A Case Study on the
Production of Multicolored Self-adhesive . . . . . . . . . . . . . . . . . . . . . . . . 547
Wellingtânia Domingos Dias, Taciana Ramos Luz,
and Meinhard Sesselmann
Smart City: A Qualitative Reflection of How the Intelligence Concept
with Effective Ethics Procedures Applied to the Urban Territory Can
Effectively Contribute to Mitigate the Corruption Process
and Illicit Economy Markets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 557
Kelem Christine Pereira Jordão, David Bianchini, Yuzo Iano,
Ana Carolina Borges Monteiro, and Reinaldo Padilha França
A Discussion of the Challenges Small Towns Face in Reaching the
Promising Scenario of Electronic Government Intelligent Cities . . . . . . 571
Valéria Sueli Reis, Yuzo Iano, Gabriel Gomes de Oliveira,
Telmo C. Lustosa, Mauro A. S. Miranda, Odair dos Santos Mesquita,
Ana Carolina Borges Monteiro, and Reinaldo Padilha França
Emerging Trends in Health, Natural
and Physical Sciences
Algorithm Oriented to the Detection
of the Level of Blood Filling
in Venipuncture Tubes Based on Digital
Image Processing

Jorge Castillo , Nelson Apfata , and Guillermo Kemper

Abstract This article proposes an algorithm oriented to the detection of the level of
blood filling in patients, with detection capacity in millimeters. The objective of the
software is to detect the amount of blood stored into the venipuncture tube and avoid
coagulation problems due to excess fluid. It also aims to avoid blood levels below
that required, depending on the type of analysis to be performed. The algorithm
acquires images from a camera positioned in a rectangular structure located within
an enclosure, which has its own internal lighting to ensure adequate segmentation of
the pixels of the region of interest. The algorithm consists of an image improvement
stage based on gamma correction, followed by a segmentation stage of the area
of pixels of interest, which is based on thresholding by HSI model, in addition to
filtering to accentuate the contrast between the level of filling and staining, and as
a penultimate stage, the location of the filling level due to changes in the vertical
tonality of the image. Finally, the level of blood contained in the tube is obtained from
the detection of the number of pixels that make up the vertical dimension of the tube
filling. This number of pixels is then converted to physical dimensions expressed in
millimeters. The validation results show an average percentage error of 0.96% by
the proposed algorithm.

Keywords Blood draw · Filling level · Detection · Image processing ·

Venipuncture tube

J. Castillo · N. Apfata · G. Kemper (B)

Universidad Peruana de Ciencias Aplicadas, Av. Prolongación Primavera 2390, Santiago de
Surco, Lima, Peru
e-mail: [email protected]
J. Castillo
e-mail: [email protected]
N. Apfata
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 3
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_1
4 J. Castillo et al.

1 Introduction

The extraction of blood either with the use of a syringe or a venipuncture tube is
performed manually, which generates that the amount of blood stored is calculated
empirically or visually by the extractor, which is the large percentage of clinical
decisions based in laboratory tests.
The Ministry of Health of Peru (MINSA) in its Technical Health Standard of the
Clinical Pathology Services Production Unit has as a general objective to establish the
criteria for the organization and operation of the Clinical Pathology UPS, of public
health services and private for proper management in it [1]. On the management
control of the Clinical Pathology UPS, quality management policies, programs, and
procedures must be developed where documents such as corrective actions, quality
records, among others must be used [1]. This is the reason why the development of
new ways of monitoring processes within Clinical Pathology is necessary since the
laws of the Peruvian State promote them. The problem appears when the venipuncture
tube is filled with excess or defect over its indicated capacity for each type of test to
be performed since it must be experienced enough to be precise with the amount of
blood drawn. With this problem, the entities or persons in charge of blood collection
do not have control over the failures they make during the extraction or the times in
which the patient was wrongly performed.
Some works related to the detection of liquid level by digital image processing (not
necessarily blood) have been found in the scientific literature. In [2], it is proposed
a multiple bottle filling system for industrial application, in which the liquid is filled
depending on the color of the bottle. Liquids can be filled with a single system.
The experimental setup is designed to fill two liquids, depending on the color of
the bottles, using the Siemens S7-1200 programmable logic controller [2]. On the
other hand, in [3], it is proposed a machine that estimates the automatic insulin
dosing by digital image processing. The equipment allows improving the accuracy
of insulin dosing in a syringe that is usually done manually and inaccurately by using
an enclosure that contains an endless screw, couplers, and stepper motors. Also, we
found a robot that automates blood collection by venipuncture called Veebot; this
machine keeps the registration of the complete process of blood extraction, as the
temporal data and images. It automates the process of blood extraction, meaning the
camera seeks for the vein and extracts the blood itself; it acquires the start and end
time, level of filling of the tube. This machine is not commercial and is expensive
as well [4]. VenusPro by Vasculogic is a machine that does exactly the same as the
Veebot, so it fulfills the same cost and commercial expectations [5].
The algorithm proposed in this work is entirely in the processing of images; the
results shown at the end show that the method used obtains results consistent with
those measured with a measuring instrument with sensitivity in millimeters.
Algorithm Oriented to the Detection of the Level … 5

2 Description of the Proposed Algorithm

Figure 1 shows the block diagram of the proposed detection algorithm. In general, the
algorithm summary follows the following steps. The process begins with the improve-
ment of the image of the venipuncture tube, which consists of gamma correction in
order to generate greater contrast in the red, green, and blue tones. After that, the
segmentation of the object or region of pixels of interest is performed; for this, the
original image is transformed to the HSI model, in which a component is chosen
where the background of the image and the area of pixels of interest are differ-
entiated. Notoriously, the image component is also chosen where there is greater
contrast between tube filling and bloodstain. Once segmented the object, the image
is filtered, first uniformizing it and then looking for changes of hue in the vertical,
and as a penultimate step, the horizontal projections are obtained to find the local
maximums in this and find the position of the filling level. Finally, a conversion of
the number of pixels in the vertical component into a real dimension is made.

2.1 Image Enhancement

For the detection of the level of filling of the tube, it is necessary to differentiate the
portion of the pixel area of the tube that is blood and stain on the wall; for this, it
is necessary to generate contrast in dark areas; this was obtained with the gamma
correction.

2.1.1 Gamma Correction

In this part, you get an image with greater contrast in the area of pixels that represents
blood. The variation of the image in hue depends on the gamma factor as can be seen
in Fig. 2, so an appropriate value was chosen that generates contrast in low tones.
The correction of each component is expressed as:

Fig. 1 Block diagram of the proposed algorithm

6 J. Castillo et al.

Fig. 2 Gamma correction correspondence function

I 1R (x, y) = IR (x, y)γ (1)

 
255 × I 1R (x, y)
IR (x, y) = round (2)
I 1Rmax

I 1G (x, y) = IG (x, y)γ (3)

 
255 × I 1G (x, y)
IG (x, y) = round (4)
I 1Gmax

I 1B (x, y) = IB (x, y)γ (5)

 
255 × I 1B (x, y)
IB (x, y) = round (6)
I 1Bmax

where I 1Rmax , I 1Gmax y I 1Bmax satisfy the following conditions:

I 1Rmax ≥ IR (x, y)∀(x, y) (7)

I 1Gmax ≥ I 1G (x, y)∀(x, y) (8)

I 1Bmax ≥ I 1B (x, y)∀(x, y) (9)

Algorithm Oriented to the Detection of the Level … 7

Likewise:
IR (x, y), IG (x, y), IB (x, y): Image in its three red, green, and blue components
respectively of the input image.
IR (x, y), IG (x, y), IB (x, y): Image resulting from the red, green, and blue
components.
γ : Gamma factor.
The results obtained can be visualized in Fig. 3, 4, 5 and 6 (after applying γ = 0.4).
The gamma value used response to the contrast tests applied for the segmentation of
the venipuncture tube.

Fig. 3 a Original red image component, b red component after gamma correction (0.4)

Fig. 4 a Original green image component, b green component after gamma correction (0.4)
8 J. Castillo et al.

Fig. 5 a Original blue image component, b blue component after gamma correction (0.4)

Fig. 6 a Original image, b image after gamma correction (0.4)

2.2 Segmentation of Object of Interest

After obtaining an image where a contrast between what is blood and stain is visu-
alized on the tube wall, only the part that represents blood is segmented; for this, the
following steps are followed.

2.2.1 HSI Model

In this part, the background is separated from the area of pixels that represents the
tube; for this, the transformation is carried out to the HSI model (Hue and saturation
that represent the component of color and intensity that represents the lighting of the
image) [6]. The results obtained can be visualized in Fig. 7.
Where the tube and the bottom differ more in the intensity component.
Algorithm Oriented to the Detection of the Level … 9

Fig. 7 a Intensity component, b saturation component

2.2.2 Thresholding by the Intensity and Red Component

At this point, the tube is separated from the bottom based on the appropriate model
in item Sect. 2.2.1 in the intensity component. Therefore, a histogram of the intensity
of the image is made, which can be seen in Fig. 8.
The most coherent threshold for separating the bottom of the tube is 190, which
is visualized in Fig. 8 with a marker, separating the object of interest; the bottom
tube is removed in the red component in Fig. 3b where the greatest contrast between
filling and stain is found in the tube; the result is displayed in Fig. 9.

Fig. 8 Intensity histogram of several images, with threshold indicator (190)

10 J. Castillo et al.

Fig. 9 Resulting image after tube segmentation

2.2.3 Image Filtering

For this step, the image of Fig. 8 was filtered with an averaging mask Fmean described
in the following expression [7]:
⎡ ⎤
1 ··· 1
1 ⎢ .. . . .. ⎥
Fmean = ×⎣. . .⎦ (10)
36
1 ··· 1 6×6

After that, filtering is applied through the convolution between the image in Fig. 9
and the averaging mask Fmean with the aim of standardizing the areas where shades
that do not belong to it (such as the light reflection) can occur by the curvature of the
tube. The result of the image can be visualized in Fig. 10.
Once the filtered image is obtained, the aim is to highlight the vertical changes in
the hue component. For this, a Sobel mask [8] is used, which can be expressed as:
⎡ ⎤
−1 −2 −1
Msh = ⎣ 0 0 0 ⎦ (11)
1 2 1

Figure 11 shows the result of the filtering.

Since the image in Fig. 11 is not very useful in this condition, thresholding is
performed. In this case, a threshold value of 5 was chosen since it allowed obtain the
best results (see Fig. 12).
Algorithm Oriented to the Detection of the Level … 11

Fig. 10 Image after filtering with an averaging mask

Fig. 11 Image after applying the Sobel mask

2.2.4 Horizontal Projection

Horizontal projections [9] were used to locate positions where there are changes in
hue in the vertical direction of the image, which can be seen in Fig. 13. Local maxi-
mums are considered if they exceed half the maximum vertical projection because
this represents half of the projection in the diameter of the tube.
Having obtained the local maximums, which in the case of Fig. 13 were in the
vertical positions 181, 346 and three near 467, which can be seen with orange markers.
12 J. Castillo et al.

Fig. 12 Image after binarization by several thresholds

After that, the search for the local maximum that meets the following conditions
is carried out. First, it should not be very close to the lower and upper position in the
vertical tube. Second, we look for that position that does not have other very close
local maxima such as that of Fig. 13 at the bottom; the only point that meets these
conditions is at position 346 for the case, which represents the position of the fill
level.

Fig. 13 Horizontal image projection with local maximum markers

Algorithm Oriented to the Detection of the Level … 13

3 Results

To obtain the filling level, the number of pixels in the vertical component is scaled
to a real measurement, which follows the following expression:

D(mm) = Npixels × dpixel (12)

where
D(mm) Tube filling level in millimeters.
N pixels Number of pixels in the vertical from the filling level to the bottom of the
tube.
d pixel Dimension of a pixel in millimeters.
The validation of the proposed detection algorithm was based on measurement
with a metric sensitivity instrument in millimeters since no contact can be made with
the sample.
The procedure consists of detecting at the level of a blood sample tube to be able
to determine if it is adequate in terms of its content for the respective laboratory tests.
For this, several tests were performed for the calculation of percentage error, which
is expressed as:

Vv − Va
Ep = × 100% (13)
Vv

where
Ep Percentage error
Vv True value
Va Approximate value
Vv − Va Absolute error.
The result of these tests can be visualized in Fig. 14.

4 Conclusions

Analyzing the results, a maximum percentage error of 4.26%, average of 0.96%

and a minimum of 0% were obtained, regarding the absolute error were maximum
1.96 mm, average 0.41 mm and minimum of 0 mm, based on these data, it can be
indicated that the detection algorithm sometimes fails in more than 1 mm near the
relative 5%, in addition, the average relative error indicates satisfactory, however,
that it could be improved. The results with high errors were witnessed when the
images obtained are immediately placed in an upright position, which causes the
bloodstain to be very similar to the content. In summary, a blood level detection
algorithm in venipuncture tubes was developed successfully with an average error
of 0.96% which is considered adequate for the purpose of this investigation. On the
other hand, when waiting for 4–5 s in a vertical position of the tube to obtain the
image, the results were better than those obtained in the tests; given this limitation,
14 J. Castillo et al.

it is recommended to perform a vertical position vibration of the tube before image

acquisition for better and faster results. It should be noted that this research is a small
part of a more extensive development so that only the development of the blood level
detection algorithm is explained here. The advantage of this research is that it can
be used as the basis of a more complex algorithm for better image acquisition, using
neural networks, among other methods.

Fig. 14 a Percentage error of calculation of the level of detection of the algorithm, b absolute error
Algorithm Oriented to the Detection of the Level … 15

References

1. Peru. Ministerio de Salud. Dirección General de Salud de las Personas. Dirección de Servicios de
Salud, Norma técnica de salud de la unidad productora de servicios de patología clínica. Recov-
ered from http://www.diresaapurimac.gob.pe/media/attachments/2018/09/07/nts_072_minsa_
dgsp_upss_patologia_clinica.pdf (2009)
2. Pannu, J., Kulkarni, R., Ranjana, M.: On the automated multiple liquid bottle filling system. In:
International Conference on Circuit, Power and Computing Technologies (2016)
3. Alarcon, M., Torobeo, D., Lau, K., Kemper, G.: An electronic equipment for automatic insulin
dosage based on digital image processing (2017)
4. Veebot: Robot Draws Blood. http://www.veebot.com/solutions.html (2019)
5. VascuLogic: VenousPro. http://vasculogic.com/about-vasculogic.html (2019)
6. Saravanan, G., Yamuna, G., Nandhini, S.: Real time implementation of RGB to HSV/HSI/HSL
and its reverse color space models. Commun. Signal Proces. (2016)
7. Oré, G., Vasquez, A., Soto, J., Kemper, G.: Measuring the level of mildew in quinoa plantations
based on digital image processing. In: Proceedings of the 3rd Brazilian Technology Symposium,
Junio 2019
8. Israni, S., Jain, S.: Edge detection of license plate using Sobel operator. In: International
Conference on Electrical, Electronics, and Optimization Techniques, Marzo 2016
9. Pan Z., Wang M.: A new method of shredded paper image stitching and restoration. In:
Computing Technology, Intelligent Technology, Industrial Information Integration (2017)
Computational Fluid Dynamics Analysis
of Blood Rheology

Guilherme S. Souza , Enzo D. Giustina , Marcus V. Carvalho ,

and Raquel J. Lobosco

Abstract Computational fluid dynamics can be an auxiliary tool for the diag-
nostic and treatment of cardiovascular diseases. Thrombosis and atherosclerosis are
pathologies that affect the delivery of blood to the neck and head. One of the most
affected by these pathologies is the carotid sinus. The objective of this paper is to
improve the accuracy of blood flow simulations, in order to avoid using in vivo
experiments on developments of new technologies and treatments. To reproduce
flow characteristics, the numerical model used is based on the average dimensions
of many angiograms. The geometry and the correct rheological model are crucial to
the fidelity of the physical phenomenon. The blood as a non-Newtonian fluid was
modeled with the Casson and Carreau-Yasuda Model. Using the open-source soft-
ware OpenFOAM for the numerical simulations integrating the rheoTool facilities,
non-Newtonian models and Newtonian assumptions were compared.

Keywords Blood · Rheology · OpenFOAM · Artery

1 Introduction

The study of blood behavior is a fundamental field of biotechnology development.

According to clinical observations, bifurcation regions and curvatures are more sensi-
tive to pathological alterations. The carotid sinus (a region in the carotid artery where

G. S. Souza (B) · E. D. Giustina · M. V. Carvalho · R. J. Lobosco

Federal University of Rio de Janeiro, Macaé, Brazil
e-mail: [email protected]
E. D. Giustina
e-mail: [email protected]
M. V. Carvalho
e-mail: [email protected]
R. J. Lobosco
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 17
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_2
18 G. S. Souza et al.

a bifurcation occurs) is most affected by arteriosclerosis lesions [1]. Therefore, the

hemodynamic study is important for a better understanding of arteriosclerosis and
its relation with the blood flow [2]. In addition, it is important to mention that in vivo
precise measures of some parameters are extremely difficult, such as shear rate.
Consequently, numerical simulations turn in a great ally in the development of new
technologies devices and research development in the area [2].
There is no consensus in which the rheological model better represents the blood
viscosity due to its complex rheology. The most used fluid models are the power
law, Casson, and Carreau-Yasuda [3]. Some researchers argue that the Newtonian
behavior hypothesis is a good assumption [1, 4] for large arteries and the non-
Newtonian behavior is more important for capillaries and small vessels. Since the
shear stress in the vessel walls has great relation with atherosclerosis, it would also
be relevant to investigate non-Newtonian behaviors in large arteries [2]. However,
none of those models considers its viscoelasticity.
In order to improve the computational fluid dynamics, CFD, simulations of blood
flow, this study reproduced some cases implemented by Carvalho [5] and integrated
different boundary conditions in the rheoFoam tool, an external solver for simulating
complex fluids at OpenFOAM package. This package allows a viscoelastic treatment
for non-Newtonian fluids.

2 Methodology

2.1 Mesh

The similarity between the geometry model and the real artery is a crucial parameter.
In this work, the geometry is based on the model used by Bharadvaj [6], Ku et al.
[7], Gijsen et al. [8], which was developed by fifty-seven angiograms of twenty-two
adults and thirty-five children, which such data, used in this research, come from to
the studies referenced.
The flow characteristics were analyzed in the internal carotid in the regions
depicted in Fig. 1; the planes S1 to S5 are divided by horizontal and vertical sections;
A-A and B-B, respectively, note that the diameter of the vessel increases until its
maximum value in S1, where it starts to decrease afterward.

2.2 Boundary Conditions

Two different values of inlet velocity were implemented. At first, we have used a
steady velocity of 0.09 m/s with a parabolic radial distribution, just like M. V. P.
Carvalho, in order to be able to compare the data from the rheoTool solver with
previous results.
Computational Fluid Dynamics Analysis of Blood Rheology 19

Fig. 1 Geometry used at Gijsen et al. [8]

The second condition was based on the work of Wiwatanapataphee et al. [9] to
reproduce this flow velocity variation on time. The periodical velocity disturbance
was used in this study to mimic the influence of the heart pulsation characteristics
on the flow. The Fourier transform was used to get the velocity function used in the
simulations.
The speed is assumed to be directly proportional to the flow rate, when the section’s
area is constant (Fig. 2).
The normalized function of the sign shown in Fig. 7 was implemented as an
oscillatory regime in blood flow. The interval from 0 to 0.45 s is the diastole and
0.45 to 0.8 s the systole. The sign function multiplied for the maximum velocity in
the common carotid results in the approximated velocity function that was used as
the boundary conditions.

Fig. 2 Speed profile with 0.089 m/s [10] as maximum speed

20 G. S. Souza et al.

The velocity distribution has a parabolic profile and the pressure on the inlet,
and the outlets are constant. In the outlet, the atmospheric pressure resembles the
experiments made by Ku et al. [7]. For all the simulations, it was considered a
stress-free condition at both exits and a no-slip condition in the vessel walls.

3 Results

A comparison of the relative error between the computational and experimental

results can be seen graphically in Fig. 3. Each section cut represents the sections S1,
S2, S3, S4, and S5. The relative error values, like the one shown in Fig. 3 for the
Yasuda model, are measured in Table 1.
Figure 4 shows a comparison in the velocity profile, and the dots are the
experimental results.
Table 1 presents the statistic comparison. The mean deviation of Yasuda’s rheo-
logical model using rheoTool package from the experimental result was 0.007365
while it was 0.0082734 m/s, reproducing the simulations of Carvalho and Lobosco
[5]. The standard deviation in the same case was 0.004945 on rheoFoam while it was
0.005604. Those results mean that using the rheoTool package, the precision and
accuracy were improved.

(a) Section A- A’ Yasuda (b) Section B -B’ Yasuda

(c) Section A- A’ Yasuda with rheoFoam (d) Section B-B’ Yasuda with rheoFoam

Fig. 3 Graphics comparing the deviation of simulation from the experiment on each section on
computational simulation and the experimental result
Computational Fluid Dynamics Analysis of Blood Rheology 21

Table 1 Statistic comparison of the differences in computational results and experimental

Rheologic model Solver Mean Standard deviation
Yasuda Non-NewtonianIcoFoam −0.008273 0.005604
rheoFoam −0.007365 0.004945
a rheoFoam 0.011122 0.017199
Newtonian Non-NewtonianIcoFoam −0.004181 0.006717
a The oscillatory regime was used on simulation, but the experimental was a steady flow

(a) Section A- A’ Yasuda (b) Section B-B’ Yasuda

(c) Section A-A’ Yasuda with rheoFoam (d) Section B-B’ Yasuda with rheoFoam

(e) Section A- A’ Newtonian (f) Section B-B’ Newtonian

Fig. 4 Graphics comparing the velocity on each section on computational simulation and the
experimental results
22 G. S. Souza et al.

(a) Speed profile on Yasuda’s model using (b) Speed profile on Yasuda’s model using
rheoFoam standard solver

Fig. 5 Illustration of velocity magnitude

(a) Yasuda residues with rheoFoam (b) Yasuda residues

Fig. 6 Graphics of residues of pressure (purple) and the velocity components x (blue), y (green)
and z (yellow) on a logarithm scale

An illustration of the numerical velocity profile can be seen in Fig. 5 while Fig. 6
shows the computational numerical residues.
The results which are shown in Fig. 5 allow a comparison between the oscillatory
velocity profile and the steady-state flow. Both models, the Newtonian and non-
Newtonian can be compared in the graph.
Figure 7 compares the velocity profiles obtained from simulations with different
fluid dynamic models applied. As can be seen, Newtonian and non-Newtonian steady-
state hypotheses present substantial differences, and these results agree with the
results stated at [2, 7], which non-Newtonian approaches are preferable.
Comparing the steady-state and oscillatory velocity profiles for non-Newtonian
models, it can be seen that the oscillatory condition has a more uniform profile. This
difference would be due to the periodic disturbance of the velocity in oscillatory
conditions, which do not allow the velocity to achieve the fully developed velocity
magnitude in the steady-state condition.
Computational Fluid Dynamics Analysis of Blood Rheology 23

(a) Section A-A’ velocity profile

(b) Section B-B’ velocity profile

Fig. 7 Graphics of velocity profile adopting the Newtonian model (green), Yasuda’s model (blue),
both on steady flow and Yasuda’s model (red) simulating the heart beating on systoles
24 G. S. Souza et al.

4 Conclusion

The oscillatory boundary condition highly impacts the results of the simulation,
showing that steady flow models may be inadequate. The rheoFoam improves the
accuracy of the simulation. The results shown in this paper are not definitive, because
of some simplification in the vessel wall conditions and unconsidered viscoelastic
characteristics; therefore, the result contributes to the progress of the numerical
research of complex fluids in arteries.

References

1. Resch, M., Perktold, K, Peter, R.O.: Three-dimensional numerical analysis of pulsatile flow
and wall shear stress in the carotid artery bifurcation. J. Biomech. 24(6), 409–420 (1991)
2. Boyd, J., Buick, J.M., Green S.: Analysis of the casson and carreau-yasuda non-newtonian
blood models in steady and oscillatory flows using the lattice boltzmann method. Phys. Fluids
24(6), 093–103 (2007)
3. Shibeshi, S.S., Collins, W.E.: The rheology of blood flow in a branched arterial system. Appl.
Rheol. 15(6), 398 (2005)
4. Shibeshi, S.S., Collins, W.E.: A numerical study of blood flow using mixture theory. Int. J.
Eng. Sci. 76, 56–72 (2014)
5. Carvalho, M.V.P., Lobosco, R.J.: An´alise reol´ogica do escoamento sangu´ıneo na bifurca¸cao
da art´eria car´otida com uso do openfoam
6. Mabon, R.F., Bharadvaj, B.K., Giddens, D.P.: Steady flow in a model of the human carotid
bifurcation. Part iilaser-doppler anemometer measurements. J. Biomech. 15(5), 349–378 (1982)
7. Ku, D.N., et al.: Hemodynamics of the normal human carotid bifurcation: in vitro and in vivo
studies. Ultrasound Med. Biol. 11(1), 13–26 (1985)
8. Gijsen, F.N.V., Frank, J.H., Janssen, J.D.: The influence of the non-newtonian properties of
blood on the flow in large arteries: steady flow in a carotid bifurcation model. J. Biomech.
32(6), 601–608 (1999)
9. Wiwatanapataphee, T.S.B., Wu, Y.H., Nuntadilok, B.: Effect of branchings on blood flow in
the system of human coronary arteries. Math. Biosci. Eng. 9(1), 199–214 (2012)
10. Wegent, F., Strecker, C., Weiller, C., Harloff, A., Zech, T., Markl, M.: Comparison of blood
flow velocity quantification by 4d flow mr imaging with ultrasound at the carotid bifurcation.
AJNR Am. J. Neuroradiol. 34, 14071413 (2013)
Wearable Technology for Presumptive
Diagnosis of High Blood Pressure Based
on Risk Factors

Eithel Josue Meza Prada , Helgar Miguel Angel Herrera Agullar ,

Jimmy Armas-Aguirre , and Paola A. Gonzalez

Abstract In this paper, we propose a technological solution integrated to a wearable

device that allows measuring some physiological variables such as body mass index
(BMI), steps walked in a determined day, burned calories, blood pressure and other
risk factors associated with the Framingham´s score. The objective of this article
is to identify the evolutionary pattern of the Framingham’s score each day in order
to determine a presumptive diagnosis of high blood pressure. The technological
solution was validated in the social insurance of a public hospital in Lima, Perú. The
preliminary results obtained in a diagnostic test show a sensitivity level of 83.33%,
a level of precision better than a traditional Framingham´s score for presumptive
diagnosis of high blood pressure. Our proposal contributes to the patient’s awareness
about the bad routine habits related to lifestyle and promotes the empowerment of
data in order to make some changes that influence on the reduction of cardiovascular
disease risk.

Keywords Physiological variable · Presumptive diagnosis · Risk factor ·

Framingham´s score · Wearable device

E. J. M. Prada · H. M. A. H. Agullar (B) · J. Armas-Aguirre

Universidad Peruana de Ciencias Aplicadas, Av. Prolongación Primavera 2390, Santiago de
Surco, Lima, Peru
e-mail: [email protected]
E. J. M. Prada
e-mail: [email protected]
J. Armas-Aguirre
e-mail: [email protected]
P. A. Gonzalez
Rowe School of Business, Faculty of Management, Dalhousie University, Halifax, NS, Canada
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 25
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_3
26 E. J. M. Prada et al.

1 Introduction

The modern world is characterized by the automation of processes, in the day-to-day

work of the people. This is evidenced by the marked change in the style of life of
each individual, both in the form of feed as in the sedentary lifestyle product of
technological advance that automates activities that involved in the past physical
activity. This leads to the appearance of cardiovascular diseases due to the excess of
the sedentary lifestyle.
According to the report of the institute of statistics and informatics (INEI) issued
in the year 2018, indicates that in Peru, the number of people of 15 years over-
diagnosed as hypertensive each year is approximately 13% of the population and its
growth occurs at a rate of 1% each year [1].
Hypertension is a condition in which the blood vessels have a consistently high
voltage, which can damage the walls of the arteries. An artery is each of the vessels
that carry blood with oxygen from the heart toward the capillaries of the body. High
blood pressure and cardiovascular diseases (CVD) are closely related since high
blood pressure is the nexus to identify people at high risk for heart attacks due to
the fact that there is a close relationship between blood pressure and cardiovascular
health; while more boost a person’s blood pressure, it is very likely that the health
of your heart is in trouble. It is for this reason that people with blood pressure levels
too high are candidates for heart attacks.
According to the research of the World Health Organization (WHO) with respect
to the year 2016, two of the ten leading causes of death in the world were due to
ischemic heart disease and stroke, which are classified as CVD [2]. The CVD refers
to a set of disorder of the blood vessels of the heart and today are the main cause of
death throughout the world.
According to the information of the guide for the prevention, detection, evalua-
tion and management of hypertension in adults, major causes of hypertension are
attributable to factors of genetic predisposition, overweight, physical condition and
sodium intake [3]. Despite the fact that the main risk factors are more related to genetic
predisposition and nutritional factors, it is important to know and to control some
risk factors associated with the lifestyle in order to reduce the risk of hypertensive
patients.
This article contributes to the emerging development of technologies committed
to health care through an integrated technological solution to a portable device, which
measured the variables considered as major risk factors associated with high blood
pressure according to the Framingham score, by means of a portable device in order
to analyze the behavior of that score with the passage of time in order to generate
a guide to health education that enables the user to obtain possible values for that
score in the future and can generate changes in your lifestyle to help the prevention
of arterial hypertension.
Wearable Technology for Presumptive Diagnosis … 27

2 Literary Review

During the planning stage, it is considered the review of related articles to various
fields. On the one hand, an analysis was made of the factors related to hypertension.
On the other hand, we reviewed the portable technologies emerging in the market.

2.1 Variables Related to Hypertension

According to the medical documentation, points out that there are multiple variables
that influence arterial hypertension; for our case study, the following variables were
considered to be obtained by means of the portable device: The person’s age [4], BMI
(Weight/height) [5, 6], calories burned in a determine day [7], steps walked in a deter-
mined day [8], measured arterial pressure in sedentary [9], consumption of alcohol
[10], consumption of tobacco [11], sodium level in the blood [12], predisposition in
hypertensive parents [13].
In [4], the author argues that as the years go by a person tends to increase your
blood pressure due to aging and stretches of arterial walls. In patients older than
65 years of age, the prevalence of hypertension is located between 60 and 70%.
In [5, 6], the author argues in his research that excess body fat contributes to the
increase in blood pressure. It is estimated that for every 10 kilos obtained, an increase
in blood pressure amounts to between 2 and 3 mm Hg. The appropriate levels of BMI
for each person must be between 18.5 and 24.9 kg/m2 .
In [7], the author argues that the basal metabolism is the set of chemical reactions
that causes the body to function optimally. Therefore, each action that the body can
generate is related to the consumption of calories. The number of calories burned a
day ideal for maintaining good blood pressure levels depends on a certain way of the
action of the basal metabolism but the main factors are the completion of exercises
or any other type of physical activity.
In [8], the author argues that it is important to establish an exercise program that
combines activities such as running, dancing, nothing or walking during each week.
It is recommended to combine activities such as walking daily between 20 and 30 min
a day to get to reduce blood pressure values at adequate levels.
In [9], the author describes the two types of arterial pressures and marks the
importance of the difference in values between the two, as shown in Table 1.
In [10], the author argues that alcohol is a factor that can lead to obstruction of
the blood vessels due to the release of salts and hormones in the blood, which has
an impact on the proper cardiac function. It is estimated that alcohol intake between
0.75 and 1 grams of ethanol/kg produces an increase in blood pressure.
In [11], the author is based on preliminary studies that describe the damage caused
by the components of tobacco and smoke generated physiopathological alterations
involved in complications associated with the acceleration of atherosclerosis and
vascular disorders and endothelial cells observed in a tobacco smoker.
28 E. J. M. Prada et al.

Table 1 Arterial pressure

Category Description
categories
Normal blood pressure 120–139 mmHg systolic and
80–89 mmHg diastolic
Mild hypertension 140–159 mmHg systolic and
90–99 mmHg diastolic
Medium-serious hypertension 160–179 mmHg systolic and
100–109 mmHg diastolic
Serious hypertension >180 mmHg systolic and
>110 mmHg diastolic

In [12], the author demonstrates the direct relationship between blood pressure
and salt consumption through development of an experiment taken over four weeks
in a sample of 500 hypertensive patients that consisted in the reduction of 0.5 g of
salt; the outcome of this case study obtained a decrease of 20% of the figures related
to blood pressure in 60% of patients. Therefore, it concludes that the restriction of
Salinas decreases the incidence of secondary factors such as hypocalcemia induced
by thiazide diuretics that affect blood pressure levels.
In [13], the author argues that despite the fact that the determinants of hypertension
are not known, there are epidemiological studies that have shown that about 30%
of the 40 the variation in blood pressure of the population is determined by genetic
factors.

2.2 Portable Technologies

Portable devices are all kinds of accessories are easy to integrate into the clothing
or body of each user. Among them is with smart clothing, fibers to collect elec-
tromagnetic signals, healthy watches among others. To carry out the measurements
of arterial pressures were evaluated the main portable devices whose measurement
values have been certified by specialized institutions. Figure 1 detailed the reason
for the election of the portable device ‘Smartwatch Skmei Jam,’

3 Technological Solution

This proposal consists in a technological solution based on the Framingham’s score, a

scheme was validated by the medical community for the identification of risk factors
associated with cardiovascular disease and hypertension, which unlike the traditional
score through a form, allows the collection of data more precisely through a portable
device non-invasive. Then it will be explained in detail of each one of the phases.
Wearable Technology for Presumptive Diagnosis … 29

Fig. 1 Devices portable evaluation

3.1 Phase 1: The Physiological Data Collection of the User

Through the Portable Device

Portable devices allow recording and storing physiological data and these are
obtained through the discovery of services available from each device in specific.
During the day, the portable device will record the data obtained by the daily activities
of the person to be stored and classified in an object for each variable defined. For the
distance traveled, steps walked and calories burned will evaluate the data obtained
throughout the day; in the case of blood pressure, the indicator which allows the
correct evaluation and reliability of the sample is the state at rest, so the user will
only be assessed such data when the user is at rest.

3.2 Phase 2: The Reception of the Data Through the Mobile

Application by Means of a Bluetooth Communication
Protocol

Developed a mobile application that allows communication and reception of the data
with the portable device via Bluetooth. The software consumes the services offered
by the portable device through reading requests to the services available during the
connection to the device. The acquisition of the data is carried out through a text
format for data exchange (JSON).
30 E. J. M. Prada et al.

Fig. 2 Data communication diagram

3.3 Phase 3: The Sending of Data from the Mobile

Application to the Database Engine of SQL Azure Hosted
in the Cloud

As part of the development of software, it is contemplated two communication

services developed in C# and deployed as App Services on Azure with communica-
tion to the storage services. These services enable communication software with the
database in the cloud through HTTPS requests: GET and POST. At the end of the data
collection during the day, the software performs a POST request to the service and
inserts the information from the user on our storage service. GET requests shall facili-
tate the procurement and display of user data. Figure 2 shows the data communication
diagram.

3.4 Phase 4: The Process of Generation of Reports

of Educational Health Based on the Framingham’s Score

Once registered, the data for each patient in the technological solution will generate
a daily educational health report which will show the levels obtained by the patient in
each parameter that has been analyzed. This allows the patient to see the variance of
Wearable Technology for Presumptive Diagnosis … 31

your score over time and be aware of certain changes in your lifestyle, thus reducing
the probability of being hypertensive patients in the future.

4 Validation

4.1 Test Scenario

The validation of the proposed model was carried out with social insurance patients
from EsSalud who was diagnosed as diabetic and had a high risk of acquiring high
blood pressure at some time in their lives. The population sampling included 25
patients who used the traditional model of the discard of cardiovascular diseases
based on a questionnaire of the main risk factors according to the Framingham’s
score and the new proposed model that proposes to calculate certain parameters such
as the distance traveled and the average for all arterial pressures throughout the day
from a portable device. For the traditional method, each question receives a certain
score according to the age and gender of the patient. In this way can be calculated
if the patient has a low probability, medium, or high to be hypertensive. For both
methods, most of the parameters were similar, in contrast to the question whether or
not the patient performs a physical activity, whose question is very subjective, and
blood pressure field of the traditional method considered only blood pressure at the
time that the patient receives the query on the part of the doctor.

4.2 Comparison Between Two Models

The traditional model, after applying the questionnaire, determined that 19 people
(76%) stated that if carried out physical activity into your daily routine, while six
people (24%) stated that they did not physical activity every day. Instead of that,
the proposed model dictated that 18 people (72.00%) were not engaged in physical
activity, 5 people (20.00%) had little physical activity, and 2 people (8.00%) carried
out an adequate physical activity. Table 2 shows the Framingham score in both models
of physiological risk analysis for the 25 diabetic patients surveyed.
The 25 patients analyzed in both models went through a more comprehensive
monitoring and greater reliability on the part of the medical community for the
measurement of blood pressure on how to final disposal of diagnosis of high blood
pressure.

Table 2 The number of patients according to the risk level of arterial hypertension
Method Low risk Medium risk Serious risk
Traditional model 4 (16%) 7 (28%) 14 (56%)
Proposed model 2 (8%) 8 (32%) 15 (60%)
32 E. J. M. Prada et al.

This monitoring is known as ambulatory blood pressure monitoring (MAP), a

method that involves the use of an assistive technology device oscillometric measure-
ment of high precision that the patient carries with it during the 24 h a day and it
performs multiple measurements of blood pressure throughout the day.
Finally, in order to validate our proposal, these metrics are important to foresee
whether a person is diagnosed as a hypertensive patient:
• True positives: Are events in which the patient presented a high risk of being
hypertensive and after monitoring by MAP was diagnosed as a hypertensive
patient.
• False positives: Are the events in which the patient presented a high risk of
being hypertensive, but after the monitoring by MAP was not diagnosed as a
hypertensive patient.
• False negatives: Are the events in which the patient presented a risk to medium
or low to be hypertensive, but after the monitoring by MAP was diagnosed as a
hypertensive patient.
• True negatives: Are the events in which the patient presented a risk to medium or
low high to be hypertensive, and after the monitoring by MAP was not diagnosed
as a hypertensive patient.
• Sensitivity: It is defined as the capacity of a diagnostic test to properly identify
an individual with a high probability of being hypertensive and then confirm the
existence of the disease. This is calculated as follows:
TP
Sensitivity = (1)
TP + FN
• Specificity: It is defined as the capacity of a diagnostic test to properly identify
an individual with a high probability of not being hypertensive and then confirms
the absence of the disease. This is calculated as follows:
TN
Specificity = (2)
TN + FP

4.3 Discussion of the Results

Table 3 shows the assessment of validity from both models. From a population of 25
diabetic patients, it is determined that six patients were diagnosed as hypertensive. By
using the traditional model, it was identified that the levels of sensitivity and speci-
ficity were 66.67% and 47.37%, respectively. On the other hand, the proposed model
was identified that levels of sensitivity and specificity were 83.33% and 47.37%,
respectively.
Wearable Technology for Presumptive Diagnosis … 33

Table 3 Assessment of validity of both models

Validity and reliability assessment of the test Traditional model Proposed model
True positive (TP) 4 (16%) 5 (20%)
False positive (FP) 10 (40%) 10 (40%)
False negatives (FN) 2 (8%) 1 (4%)
True negatives (TN) 9 (36%) 9 (36%)

5 Conclusions

This work proposed a technological solution to analyze the probability of being a

hypertensive patient through the use of the Framingham score. This solution proposes
to calculate some risk factors such as the number of steps walked and the average
arterial pressure value throughout the day via the portable device.
Despite the fact that the experiment was carried out with a small number of
patients, it can be concluded that the proposed model has a better degree of accuracy
in terms of sensitivity that the traditional model, although in terms of specificity both
have the same results.
On the other hand, this technological solution aims to raise awareness of patients in
the importance of lifestyle changes in order to reduce the amount of punctuation and
prevent be hypertensive patients. In addition, it allows doctors to carry out real-time
control of arterial pressures of their patients and identify whether they are complying
with the health objectives proposed.

References

1. Instituto Nacional de Estadistica e Informatica, INEI: Encuesta Demografica y de salud

familiar. https://proyectos.inei.gob.pe/endes/2018/ppr/Indicadores_de_Resultados_de_los_
Programas_Presupuestales_ENDES_Primer_Semestre_2018.pdf. Last accessed 09 Sept 2019
(2018)
2. Organización Mundial de la Salud, OMS: Las 10 principales causas de defunción. https://www.
who.int/es/news-room/fact-sheets/detail/the-top-10-causes-of-death (2018)
3. Mendis, S., Lindholm, L.H., Mancia, G., et al.: World Health Organization (WHO) and Inter-
national Society of Hypertension (ISH) risk prediction charts: assessment of cardiovascular
risk for prevention and control of cardiovascular disease in low and middle-income countries.
J. Hypertens. 25, 1578–1582 (2007)
4. Kovesdy, C.P., Alrifai, A., Gosmanova, E.O., et al.: Age and outcomes associated with BP in
patients with incident CKD. Clin. J. Am. Soc. Nephrol. 11, 821–831 (2016)
5. Neovius, M., Rasmussen, F.: Evaluation of BMI-based classification of adolescent overweight
and obesity: choice of percentage body fat cutoffs exerts a large influence. The COMPASS
Study. Eur. J. Clin. Nutr. 62, 1201–1207 (2008)
6. Dong, J.Y., Zhang, Z.L., Wang, P.Y., et al.: Effects of high-protein diets on body weight,
glycaemic control, blood lipids and blood pressure in type 2 diabetes: meta-analysis of
randomised controlled trials. Br. J. Nutr. 110, 781–789 (2013)
34 E. J. M. Prada et al.

7. Mertens, I.L., Van Gaal, L.F.: Overweight, obesity, and blood pressure: the effects of modest
weight reduction. Obes. Res. 8, 270–278 (2000)
8. Whelton, S.P., Chin, A., Xin, X., et al.: Effect of aerobic exercise on blood pressure: a meta-
analysis of randomized, controlled trials. Ann. Int. Med. 136, 493–503 (2002)
9. Thomopoulos, C., Parati, G., Zanchetti, A.: Effects of blood pressure lowering on outcome
incidence in hypertension: 7. Effects of more versus less intensive blood pressure lowering and
different achieved blood pressure levels–updated overview and meta-analyses of randomized
trials. J. Hypertens. 34, 613–622 (2016)
10. Xin, X., He, J., Frontini, M.G., et al.: Effects of alcohol reduction on blood pressure: a meta-
analysis of randomized controlled trials. Hypertension 38, 1112–1117 (2001)
11. Meade, T.W., Imeson, J., Stirling, Y.: Effects of changes in smoking and other characteristics
on clotting factors and the risk of ischaemic heart disease. Lancet ii, 986–988 (1987)
12. Zemel, M.B., Gualdoni, S.M., Sowers, J.R.: Sodium excretion and plasma renin activity in
normotensive and hypertensive black adults as affected by dietary calcium and sodium. J.
Hypertens. 4(Suppl 6), S343–S345 (1986)
13. Kunes, J., Zicha, J.: The interaction of genetic and environmental factors in the etiology of
hypertension. Physiol. Res. 58(Suppl 2), S33–S41 (2009)
Cytology and Hematology: A Review
of the Fundamental Principles

Ana Carolina Borges Monteiro , Yuzo Iano , Reinaldo Padilha França ,

and Rangel Arthur

Abstract Since ancient times, human beings have been looking for alternative
methods to detect and diagnose diseases, mainly by assessing patients’ body fluids.
The creation of the microscope contributed to the development of modern techniques
of clinical analysis, as this instrument allowed the visualization of structures invisible
to the naked eye and opened the door to various types of scientific research. Based
on this, it is essential to develop studies that synthesize the fundamental information
on hematology and cytology, in order to support future research as well as assist in
the academic training of professionals in the medical fields.

Keywords Blood cells · Hematology · Blood count · Hematological devices ·

Cellular biology

1 Introduction

The cells are described as functional and structural units defined as the smallest part
of a living organism. The cells that make up a living organism are all the same, with
the same genetic load; however, due to the natural process of cell differentiation, they
present variations in shape and performance according to the function they perform

A. C. Borges Monteiro (B) · Y. Iano · R. Padilha França

School of Electrical and Computer Engineering (FEEC), University of Campinas, UNICAMP, Av.
Albert Einstein - 400, Barão Geraldo, Campinas, SP, Brazil
e-mail: [email protected]
Y. Iano
e-mail: [email protected]
R. Padilha França
e-mail: [email protected]
R. Arthur
School of Technology (FT), School of Technology, University of Campinas, UNICAMP, R.
Paschoal Marmo - 1888, Jardim Nova Italia, Limeira, SP, Brazil
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 35
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_4
36 A. C. B. Monteiro et al.

in the organism. This division of tasks between cells is a strategy for saving energy
and increasing efficiency in carrying out biological functions [1].
The cells have the same biological function that are grouped together and form a
tissue, which in turn form the organs (brain, lungs, intestines, heart). A grouping of
organs creates a system (neurological system, respiratory system, digestive system,
cardiovascular system). The set of systems originates organisms, such as plants,
insects, parasites, and even the human body. This entire process must take place in
an extremely precise and harmonic way. Otherwise, the changes will be so drastic
that they are incompatible with life [2].
Cells are organisms susceptible to changes in the environment in which they
are. According to Darwin’s theory, over time, there may be climate changes and/or
shortages of nutrients, consequently, such factors stimulate physiological changes
to organisms [3]. In this way, only the organisms that are able to adapt are able to
survive [2].
The human body has about 100 trillion cells, consisting mainly of carbon, oxygen,
hydrogen, calcium, phosphorus, and nitrogen [1, 4]. About 70% of the human
organism is formed by hydrogen and oxygen, which are responsible for the forma-
tion of water. The other 30% are responsible for the constitution of carbon-based
compounds, which are called organic compounds [5, 6].
Cell classification is done in two distinct types: prokaryotes and eukaryotes. The
prokaryotic cells (from the Greek, pro = first; carium = nucleus) are small (1–5 μm),
have a cell wall, responsible for mechanical protection, and have genetic material
dispersed by the cytoplasm, that is, they do not present—they have a nuclear envelope
[2, 7]. These cells include bacteria and blue algae, being formed by only a single
cell, that is, they are classified as unicellular [2, 7, 8].
Eukaryotic cells (from Greek, eu = true; carium = nucleus) have more complex
genetic material, being formed by DNA and RNA, which are organized in the cyto-
plasm through the cell envelope. Eukaryotic cells are responsible for the formation
of more complex organisms, that is, those formed by more than one cell (multicel-
lular organisms). Among these organisms, it is possible to mention: vertebrate and
invertebrate animals and plants. In addition, the cytoplasm of eukaryotic cells has
compartments that have different molecules responsible for performing specialized
functions [2, 7].
These cells have a more complex structure, being composed of several structures
such as cell membranes, nuclear envelope, nucleus, cytoplasm, peroxisomes, lyso-
somes, endoplasmic reticulum, mitochondria, Golgi complex, cytoskeleton, plasma
membrane, and cytoplasmic deposit [2, 7]. However, such discoveries could only be
made through the invention of the optical microscope.
Thus, the objective of the present study is to carry out a bibliographic review
regarding cytology and hematology, in order to easily explain the synthesis of blood
cells as well as their biological functions.
Cytology and Hematology: A Review of the Fundamental Principles 37

2 Methodology

The present study is based on the collection of information from 27 bibliographic

sources in the medical literature that addresses the subject of cytology and hema-
tology. These sources were chosen because they have an emphasis on blood cell
synthesis.

3 Results and Discussion

The first notion of microscopy appeared around the year 1595, through the works
of the Dutchman Zacharias Janssen, a glasses manufacturer who, through the union
of two lenses on the same axis, noticed the presence of small structures in objects.
Later, in the year 1665, Robert Hooke, using an ocular lens and an objective lens,
built the first microscope. This work originated in the book Micrographia, where
different living beings observed under microscopy have been described [4].
For 50 years, the Dutchman Anton van Leeuwenhoek, burned lenses and built
high-quality microscopes for his time. In this way, Leeuwenhoek was the first to
describe the various forms of bacteria in a drop of water from a lake, with his
findings constantly being sent to the Royal Society of London. Later, through the
scraping of his own teeth and water from washing a pepper, Leeuwenhoek described
the presence of bacterial structures. The Royal Society, after analyzing the letters,
sent Robert Hooke, who confirmed the veracity of the observations made by him [4].
Human eyes are capable of detecting variations in intensity and in the length of
light waves. This human capacity has been expanded through advances in the field
of microscopy, which have allowed for increased resolution as well as the use of
techniques that compensate for the transparency of cells. Cellular components are
generally transparent, with the exception of some cytoplasmic pigments capable of
absorbing certain wavelengths of light. Living cells have low absorption of these
wavelengths, as they have a high concentration of water, so even when dehydrated
this problem persists. To compensate for this limitation, dyes are used that selectively
stain the different cellular structures. These dyes are made up of compounds capable
of absorbing specific wavelengths [8, 9].
The resolution limit of the microscope is responsible for providing the details
observed in the images generated from a sample. In this way, it has no relation to the
capacity of increasing the size of the analyzed image. Thus, it is determined by the
shortest distance between two points [3]. It is important to note that the increased
capacity is valid only when accompanied by a parallel increase in the equipment’s
resolving power. Thus, the resolution limit is dependent on the ocular and objective
lenses, which are unable to add details. Thus, ocular lenses are totally dependent on
objective lenses, as they are responsible for increasing the size of the image (Fig. 1)
[3, 9].
38 A. C. B. Monteiro et al.

Fig. 1 Microscope structure—sample analysis example

In Eq. 1, it is possible to note the instrument’s resolution limit (RL) of the objective
lens. In this context, the microscope resolution limit is totally dependent on the
numerical aperture (NP) coming from the objective lenses, and it is also necessary
to consider the length of light applied to the sample.

NP = k × λAN (1)

where k is a constant estimated in seconds, being numerically represented by 0.61

and λ (lamina) is the wavelength of the light used. Due to the greater sensitivity of
human eyes to the wavelength of the yellow–green band, the value of 0.55 μm is
used in the calculation of the resolution limit. Through this formula, it is notable that
the resolution limit is directly proportional to the wavelength of the light used, and it
is also inversely proportional to the numerical aperture of the objective lens [2, 3, 9].
The creation of the first microscope opened the door to the development of tech-
niques used in the areas of clinical analysis. Clinical analyses date back to the age of
Hippocrates, the father of medicine. At that time, the color, odor, taste, and texture
of biological fluids were important tools for patients’ clinical diagnoses [2, 8].
Later, with the creation of the microscope, it was possible to correlate the physical
characteristics of biological fluids with their microscopic constitution of the analyzed
samples. In other words, it was understood that the human organism is formed by
cells and can be attacked by fungi, bacteria, and parasites. The characterization of
the morphology and after of the physiology of blood cells was essential (and still is)
for the conclusion of several types of human pathologies [2, 8].
Human blood represents approximately 7% of a person’s total body weight,
with its constitution corresponding to 55% of the plasma and 45% of blood cells.
Among the diverse functions performed by the blood, it is possible to mention: basic
acid balance, excretion of metabolites, transport of gases and nutrients to organs
and tissues, distribution of hormones, osmotic balance, and temperature regulation
[2, 7–9].
The function of transporting substances such as plasma proteins, water, amino
acids, glucose, hormones, immunoglobulins (IG), albumin, components of the coag-
ulation cascade, among others, is the responsibility of the plasma. Plasma is the
liquid part of the blood, disregarding any and all types of cells. Plasma is of great
importance for coagulation analysis, while serum is characterized by being a plasma
Cytology and Hematology: A Review of the Fundamental Principles 39

with the absence of coagulation factors. The serum is widely used for biochemical
tests, such as glucose measurement, renal markers, liver chewers, pregnancy tests,
hepatitis, HIV, among others [2, 8, 9].
The solid portion of the blood is made up of blood cells, which are called erythro-
cytes, leukocytes, and platelets. The main function of these elements is homeostasis
of the body, tissue repair, blood clotting, and defense of the body against various types
of etiological agents, which are usually: viruses, bacteria, and intestinal parasites.
However, the discovery of all blood elements as well as their respective functions is
the result of the invention of instruments capable of increasing the size of cells and
ethylene agents hundreds of times [7–9].

3.1 Hematopoiesis

The synthesis of human blood, uterine life begins at the same time. During the
embryonic phase, the first blood cells are synthesized by the yolk sac. Later, when
the fetus begins to develop, the synthesis of blood elements becomes a liver and
splenic function. When the individual reaches adulthood, the bone marrow takes
on the task of producing blood cells. This is because this region is composed of
hematopoietic tissue with controlled and orchestrated conditions for the formation
of each precursor cell. It is important to note that any of the blood cells come from
the stem cell, which resides inside the bone marrow [10–13].
The entire process of blood cell formation is called hematopoiesis and occurs
within the bone marrow from a mother cell also known as a stem cell. The main
characteristic of this cell is to be of the pluripotent type, that is, it has the capacity
to originate different blood cell lines, be they erythrocytes, leukocytes, or platelets
(Fig. 2). However, for these syntheses, several coordinated actions are necessary that
generate orchestrated responses that culminate in processes of duplication, differen-
tiation, and maturation. It is important to note that many sanguineous cells end their
maturation process in the bloodstream or in lymphoid organs [14, 15].
The stem cell is a cell capable of self-duplication, giving rise to daughter cells.
The daughter cells are the Colony Forming Units (CFU), as they have the genetic
information necessary for the formation of a cell series. The hematopoiesis process
is characterized by the participation of regulatory and stimulating factors such as
cytokines, erythropoietin (regulator of erythropoiesis), thrombopoietin (regulator of
thrombocytopoiesis), Granulocyte Colony-Stimulating Factor (G-CSF), Monocyte
Colony-Stimulating Factor (M-CSF), Macrogaph Colony-Stimulating Factor (M-
CSF), IL-3 (multiple colony-stimulating factors) and interleukins IL-2, IL-4, and
IL-6 [12, 13, 16].
40 A. C. B. Monteiro et al.

Fig. 2 Hematopoiesis

During the embryonic phase, the yolk sac is responsible for the formation of
the first blood cells; later, during fetal development, this function is transferred to
the liver and spleen. During adulthood, the inside of the bone marrow consists of
hematopoietic tissue (or hematopoiesis), which resides inside long bones and in
the axial skeleton. Inside the bone marrow, there is a network of venous sinusoids
around an arteriole and central veins, which permeate the developing cells. All blood
cells are derived from bone marrow stem cells. Thus, the bone marrow presents a
suitable microenvironment for the development of hematopoietic cells and for the
proliferation of primitive cells and progenitor cells [10–12].
Cytokines are glycoproteins that act in low concentrations on re-receptors,
producing signals that control the cell cycle, maturation and cell functions. Gener-
ally, they are responsible for the regulation of more than one cell line and show an
additive and synergistic effect with other growth factors, modulating the expression
of regulatory genes producing cytokines [12, 13].
Cytology and Hematology: A Review of the Fundamental Principles 41

3.1.1 Erythrocytes

In the literature and in the medical routine, erythrocytes receive other nomenclatures
such as red blood cells (RBC) and erythrocytes. The main morphological character-
istics of these cells are the absence of a cell nucleus, a biconcave disk shape, and
formation based on a tetramer, that is, two α chains and two β chains, which carry
out the transport of oxygen through the connection of the oxygen molecule to each of
the iron molecules [17, 18]. The gas exchange process occurs when the erythrocytes
reach the lungs, at that moment there is the formation of oxyhemoglobin (oxygen
molecules linked to the iron molecules present inside each red cell). Upon reaching
the tissues, this iron–oxygen bond is broken due to the pressure difference. In this way,
red blood cells deposit oxygen molecules and remove carbon dioxide from tissues.
The binding of carbon dioxide to iron is called carboxyhemoglobin. Subsequently,
the dioxide is transported to the lungs or dissolved in the plasma [17–20].
The erythrocytes originate in the bone marrow, being formed through the cell
maturation processes of the following cells: proerythroblast, basophilic erythrob-
last, polychromatic erythroblast, orthochromatic erythroblast, and reticulocyte.
By differentiating the pluripotential cell (Stem cell) and erythropoietin stimuli,
proerythroblasts are created [13, 19, 20].
The final stages of the maturation process are characterized by the formation of
reticulocytes, which are also called young red blood cells. The reticulocytes have
mitochondria, an irregular shape, and several remaining organelles inside. The most
striking feature of cell maturation is the reticulocytes leaving the bone marrow into the
bloodstream, where within 72 h, their cell nucleus expels, giving rise to erythrocytes
[13, 20].

3.1.2 Leukocytes

The body’s innate immune response and adaptive immune response are the results of
the action of leukocytes. Both in the literature and in the medical routine, leukocytes
are also called white blood cells (WBC). These cells are classified into two major
groups: granulocytes and agranulocytes. Granulocytic leukocytes are defense cells
characterized by the presence of granules dispersed in the cytoplasm. These granules
exert anti-inflammatory and antimicrobial action. Another morphological character-
istic of these cells is the presence of a variable number of lobes. Thus, granulocyte
cells are also called polymorphonuclear cells. The leukocytes that fit these charac-
teristics are neutrophils, eosinophils, and basophils. In turn, leukocytes classified as
agranulocytes do not have granules visible under optical microscopy generally used
in the routine of several clinical analysis laboratories. Another striking morphological
feature is the presence of only one lobule. In view of this, these leukocytes are also
called monomorphonuclear nomenclature. Agranulocyte leukocytes are lymphocytes
and monocytes [17, 18].
42 A. C. B. Monteiro et al.

Fig. 3 Mitosis phases

The synthesis of leukocytes occurs during hematopoiesis, where the leukocyte

precursor cells of the granulocytic series (neutralphils, eosinophils, and basophils)
start the process of producing proteins and cytoplasmic granules. In this context,
the primary granules are characterized by blue staining and are responsible for the
conversion of precursor cells called myeloblasts to promyelocytes. Subsequently,
specific granules appear, which are responsible for the progression to myelocytes of
the type: neutrophils, eosinophils, and basophils [19, 21].
The leukocytes resulting from the process of determining the granulocytic lineage
are indivisible by mitosis (The phases of mitosis can be seen in Fig. 3). In addition,
these cells have a segmented nucleus, the ability to phagocyte particles, the destruc-
tion of microorganisms and motility. It is important to note that depending on the
leukocyte lineage, when it reaches the maturation process, it has the ability to cross
the wall of venules in order to destroy invading agents and rebuild damaged tissues
[13, 21].

Neutrophils

After determining the production of the neutrophil lineage, the cell maturation
sequence is myeloblast, promyelocyte, myelocyte, metamyelocyte, rod neutrophil,
and segmented neutrophil, with the production of neutrophils stimulated by G-
CSF. An adult human being produces more than 1 × 1011 neutrophils per day, so
neutrophils make up about 70% of blood cells. In addition, after leaving the inside
of the bone marrow, each neutrophil survives an average of 6 h in the bloodstream
[12, 13, 19].
Until the metamyelocyte phase, the cells are contained within the bone marrow;
however, from the neutrophil rod stage, cells are released into the bloodstream. The
rod neutrophils measure about 12 μm, with a horseshoe-shaped nucleus or stick,
remaining in this stage until the segmentation of the lobes undergoes. When found in
abundance in the bloodstream, these cells indicate the presence of infection and/or
acute inflammation [12, 13, 19].
Segmented neutrophils, in turn, are characterized by a spherical shape with an
approximate diameter of 12–15 μm, and the nucleus is segmented from 3 to 5
connected lobes. Mature neutrophils have primary and secondary granules; however,
Cytology and Hematology: A Review of the Fundamental Principles 43

secondary granules are more numerous, about twice as many as primary granules.
The primary granules are azurophilic, have positive peroxidase, and measure about
500 nm. These granules are made up of a large amount of lysosomal enzymes (a
lysosome, elastase, proteinase 3—and α—antitrypsin), and bactericidal factors are
formerly known as cationic proteins (defensins, bactericidal factors and bacteri-
cidal protein to increase permeability). In turn, the secondary granules are smaller,
measuring about 200 nm and contain lactoferrin, lysozyme, and protein-bound to
B12 and other proteins [13, 19].
Classified as phagocytic cells, neutrophils have the primary function of iden-
tifying, ingesting, and destroying microorganisms. The host’s defense responses
consist of sequential stages of recognition and activation of phagocytes and the
destruction of microorganisms. Neutrophils spend less than 48 h in the bloodstream
before migrating to the tissues, being the first cells to reach the sites of infection
and/or inflammation. Thus, through direct contact with protein secretion, and the
presence of receptors for antibodies and complements, these cells are able to recruit
and mediate other cells of different lineages to the site of inflammation and/or infec-
tion. After the defense action, these cells undergo programmed cell death (apoptosis)
[12, 21].

Eosinophil

The maturation phases of eosinophils follow a similar line to that of neutrophils,

is described through the following stages: eosinophil myeloblast, promyelocyte,
myelocyte, and mature eosinophil, which suffer the action of G-CSF, IL-3, and IL-5,
responsible for eosinophil maturation) [13, 21].
Mature eosinophils represent about 2–3% of leukocytes and act on innate and
acquired immunity, being present in the mucous membranes of the respiratory,
gastrointestinal, and genito-urinary tracts, and may increase in number due to the
recruitment of these blood cells in inflammatory and/or infectious conditions. Clas-
sified as granulocytic cells, eosinophils have cytoplasmic granules that contain
enzymes responsible for damaging the cell walls of parasites or acting against allergic
responses. These granules contain basic proteins that bind to acidic dyes like eosin
[21, 22].
Eosinophilic granules and their myeloid precursors are abundant in peroxidases
and lysosomal enzymes; however, eosinophil peroxidase is genetically different from
neutrophils, conferring less bactericidal action when compared to neutrophils. Thus,
the cellular function of eosinophils is directly linked to the storage of proteins, such as
major basic protein, eosinophil cationic protein, and eosinophil-derived neurotoxin
[12, 20].
The larger basic proteins represent more than half of the granule proteins and
have cytotoxic activity against parasitic infections. They also induce the release of
histamine by basophils and mast cells, this process acts in the negative regulation of
allergic and inflammatory processes. The other basic proteins can cause the formation
of transmembrane pores capable of causing tissue damage [11–13].
44 A. C. B. Monteiro et al.

Basophils

Basophils are granulocytic cells that act on innate and acquired immunity. Its granules
are filled by various inflammatory and antimicrobial mediators, thus being respon-
sible for the immune responses against allergic reactions. Despite the structural and
functional similarities with mast cells, basophils originate from other progenitor
cells: promyelocyte, basophilic myelocyte, basophilic metamyelocytes, basophilic,
and basophilic rod. The entire process of development and maturation occurs inside
the bone marrow, with only the mature basophil being released into the bloodstream
[12, 19].
These cells represent less than 1% of the leukocytes in the blood. The granules of
the basophils are abundant and dense, capable of preventing complete visualization
of the nucleus under an optical microscope and have an affinity for basic dyes.
In addition, they are able to synthesize several mediators. Although they are not
normally present in the tissues, basophils can be recruited to some inflammatory
sites. Like mast cells, basophils express receptors for immunoglobulin G (IgG) and
immunoglobulin E (IgE) and can be stimulated by binding the antigen to the IgE on
its surface [12, 15].

Monocytes

Monocytes are classified as agranulocytic cells, originating from the following

lineage: monoblast, pro monocyte, monocytes, and macrophages. The functions of
monocytes are the removal of dead or senescent cells, removal of foreign particles,
processing, and presentation of antigens in immunological reactions and participation
in acute and chronic inflammatory reactions. They also have an important role in the
formation of atheromatous plaque (related to atherosclerosis) and in the destruction
of both microorganisms and neoplastic cells and grafts [22, 23].
Monocytes are characterized by acquiring their functional maturity in the tissues
in the form of macrophages. There are two possibilities after the monocytes leave the
medulla: either the cells are housed in the marginal compartment of blood vessels
(adhered to endothelial cells) or are released into the bloodstream [22, 24].
These cells are the largest leukocytes, measuring 15–20 μm in diameter; however,
they are responsible for forming only a small portion of the total blood cell population.
The monocyte nucleus is euchromatic being relatively large and irregular, having
the characteristic of invagination on one side. Next to this nuclear invagination,
this cell contains a prominent Golgi complex and vesicles. Monocytes are actively
phagocytic cells, containing numerous lysosomes. They are also highly mobile and
with a well-developed cytoskeleton. In turn, the cytoplasm is light in color [23, 24].
Cytology and Hematology: A Review of the Fundamental Principles 45

Lymphocytes

Lymphocytes are the second most numerous type of leukocytes in adults, while
during childhood they are the most numerous type of blood leukocytes. They are
formed through the following lineage: undifferentiated lymphoid cell, lymphoblast,
pro-lymphocyte, medium lymphocyte, small lymphocyte, and plasmocyte [25, 26].
Most circulating lymphocytes have a small size of 6–8 μm in diameter and
increased cytoplasmic volume. Often, this change occurs as a result of antigenic stim-
ulation. These cells are found in extravascular tissues (including lymphoid tissue),
being classified as the only leukocyte that returns to the circulation. The lifespan of
lymphocytes varies from a few days to many years. Long-lived lymphocytes play an
important role in maintaining immune memory [25].
Blood lymphocytes are a heterogeneous collection, mainly of B and T cells, and
consist of different subsets and stages of activity and maturity. About 85% of all
lymphocytes circulating in the blood are T cells. Primary immunodeficiency diseases
can cause molecular damage to T and B lymphocytes. Included as lymphocytes are
Natural Killer Cells (NK). NK cells are morphologically similar to T cells [26].
B lymphocytes and T lymphocytes have a rounded nucleus with dense color,
surrounded by a sparse cytoplasm, which is hardly visible under optical microscopy.
In the electron microscope, few cytoplasmic organelles can be seen, except for a
small number of mitochondria, isolated ribosomes, scarce ER profiles, and occa-
sional lysosomes. These cells become mobile when they come into contact with
solid surfaces and can pass between endothelial cells to exit or reenter the vascular
system. Thus, they migrate widely within the various tissues, including the epithelia
[26, 27].

3.1.3 Platelets

Platelets are cytoplasmic fragments of megakaryocytes produced in the bone marrow.

Because they are fragments, these cells have no nucleus and are therefore anucleated.
They have a discoid shape of about 2-4 μm in diameter and a very complex internal
structure divided into four zones: peripheral zone, sol-gel zone, organelle zone, and
membrane system. The interior of the platelet is able to communicate with the external
environment due to the presence of a channel system known as the open canalicular
system. This communication is important, as it guarantees the release of molecules
stored in platelets [13, 19, 20].
Platelets have significantly reduced blood circulation time, taking an average of
10 days to remove. These cell fragments are removed from the circulation by the retic-
uloendothelial cells of the liver and spleen. These cells have important functions for
the maintenance of our bodies. When, for example, a lesion occurs in a blood vessel,
they clump together, forming a plug, and release substances that ensure that more
platelets move to the site. In addition, they participate in the coagulation cascade,
releasing important substances that guarantee the formation of a clot. It is worth
noting that platelets also have enzymes that contribute to the removal of the clot
[13, 19, 20].
46 A. C. B. Monteiro et al.

4 Conclusions

Cells are the fundamental units of life on Earth, having undergone countless changes
since their emergence. These changes allowed the evolution of living beings to the
form and function they currently have. This evolution was due to the specialization
of cell groups to perform specific functions aimed at ensuring the balance of life
of organisms. Among these developments, there was the appearance of blood cells,
which are divided into three basic types: erythrocytes, leukocytes, and platelets,
which are responsible for the transport of gases throughout the body, defense against
antigens, blood clotting, among other functions.
The importance of the classification of blood cell types is that through this knowl-
edge it is possible to detect and carry out treatment of types of blood disease in
which blood components are affected such as anemia derivations through red blood
cells, immune thrombocytopenic derivations through platelets, leukemia derivations
by white blood cells, among other several hematological diseases, such as types
of hemoglobinopathies and hereditary coagulopathies, in the same way, that it is
possible to detect several types of related disorders and diseases.

References

1. Farah, S.B.: DNA: Segredos e Mistérios. In: DNA: segredos e misterios (2007)
2. Junqueira, L.C., Carneiro, J.: Biologia Celular e Molecular, 9ª edição. Guanabara Koogam, Rio
de Janeiro (2012)
3. Koss, L.G., Gompel, C.: Introdução à Citopatologia Ginecológica com Correlações Histológ-
icas Clínicas. Editora Roca (2006)
4. Karp, G.: Biologia Celular e Molecular: Conceitos e experimentos, 3ª edição, pp. 53–54.
Barueri, SP (2005)
5. Weiss, T.F.: Cellular Biophysics. MIT Press, Cambridge, Mass (2017)
6. Darwin, C.: A Origem das Espécies. Sumaré, Martin Claret (2014)
7. Junqueira, L.C., Carneiro, J.: Histologia Básica. Rio de Janeiro, Brazil, Guarabara Koogan
(2013)
8. Murray, P.R., Rosenthal, K.S., Pfaller, M.A.: Microbiología médica. Elsevier Health Sciences
(2017)
9. De Robertis, E., Hib, J.: De Robertis Bases da Biologia Celular e Molecular. 4ª. Guanabara
Koogan, São Paulo (2006)
10. Molinaro, E., Caputo, L., Amendoeira, R.: Métodos Para a Formação de Profissionais em
Laboratórios de Saúde. Fio Cruz, Rio de Janeiro (2009)
11. Verrastro, T., Lorenzi, T.F., Neto, S.W.: Hematologia e Hemoterapia: Fundamentos de
Morfologia, Patologia e Clínica. Atheneu, Rio de Janeiro (2005)
12. Delves, P.J., et al.: Essential Immunology. Wiley (2017)
13. Monteiro, A.C.B.: Proposta de uma metodologia de segmentação de imagens para detecção e
contagem de hemácias e leucócitos através do algoritmo WT-MO. Dissertation, State University
of Campinas, UNICAMP—SP, Brazil (2019)
14. Abbot Laboratorie De México S.A.: Atlas Com Interpretacion Histogramas Y Escatergramas.
E.G, Buenos Aires (2002)
15. Fischbach, F.T., Dunning, M.B.: A Manual of Laboratory and Diagnostic Tests. Lippincott
Williams & Wilkins (2009)
Cytology and Hematology: A Review of the Fundamental Principles 47

16. Abbas, A.K., Lichtman, A.H.H., Pillai, S.: Cellular and Molecular Immunology. Elsevier Health
Sciences (2014)
17. Monteiro, A.C.B., Iano, Y., França, R.P., Arthur, R., Estrela, V.V.: A Comparative study between
methodologies based on the hough transform and watershed transform on the blood cell count.
In: Brazilian Technology Symposium, pp. 65–78. Springer, Cham (2018)
18. Monteiro, A.C.B., Iano, Y., França, R.P., Arthur, R.: Methodology of high accuracy, sensitivity
and specificity in the counts of erythrocytes and leukocytes in blood smear images. In: Brazilian
Technology Symposium, pp. 79–90. Springer, Cham (2018)
19. Souza, M.H.L., Rêgo, M.M.S.: Princípios de Hematologia e Hemoterapia. Alfa Rio (2009)
20. Da Silva, P.H., Hashimoto, Y., Alves, H.B.: Hematologia Laboratorial. Revinter, Rio de Janeiro
(2009)
21. Murphy, K., Travers, P., Walport, M.: Janeway’s Immunobiology, 8th edn. Garland Science,
New York (2012)
22. Rodak, B.F., Caar, J.H.: Clinical Hematology Atlas. Elsevier Health Sciences (2015)
23. Nemer, A.S.A., Neves, F.J., Ferreira, J.E.S.: Manual De Solicitação e Interpretação de Exames
Laboratoriais. Revinter (2010)
24. Standring, S. (ed.): Gray’s Anatomia: A Base Anatômica da Prática Clínica. Elsevier, Brasil
(2010)
25. Erichsen, E.S., et al.: Medicina Laboratorial Para o Clínico. Elsevier (2015)
26. Da Silva, P.H., et al.: Hematologia Laboratorial: Teoria e Procedimentos. Artmed Editora (2015)
27. Failace, R.: Hemograma: Manual de Interpretação. Artmed Editora (2015)
Hematology: A Review of the Main
Methodologies of Clinical Analyses

Ana Carolina Borges Monteiro , Yuzo Iano , Reinaldo Padilha França ,

and Rangel Arthur

Abstract Human blood is composed of liquid and solid elements, blood cells being
extremely important for the assessment of the patient’s health status. Blood consists
of plasma and cellular elements, which include blood cells and cellular fragments.
Erythrocytes are blood cells containing a large amount of hemoglobin which is the
pigment responsible for the transport of oxygen, being numerous and the most found
in our blood. Leukocytes are colorless cells whose main function is to defend our
organism, still emphasizing that there is not only one type of leukocyte, but it is
possible to identify five distinct types. In this scenario, the exam that analyzes blood
cells is called a blood count, and for this reason, it is the most requested laboratory
exam in the medical routine. Based on this, explaining the methodologies of this exam
as well as performing its interpretation is essential to support scientific research and
provide a background for health professionals in academic training.

Keywords Blood cells · Hematology · Blood count · Hematological devices ·

Cellular biology

A. C. Borges Monteiro (B) · Y. Iano · R. P. França

School of Electrical and Computer Engineering (FEEC), University of Campinas – UNICAMP,
Av. Albert Einstein - 400, Barão Geraldo, Campinas, SP, Brazil
e-mail: [email protected]
Y. Iano
e-mail: [email protected]
R. P. França
e-mail: [email protected]
R. Arthur
School of Technology (FT), University of Campinas – UNICAMP, R. Paschoal Marmo - 1888,
Jardim Nova Italia, Limeira, SP, Brazil
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 49
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_5
50 A. C. B. Monteiro et al.

1 Introduction

The blood count is a laboratory test that is highly requested in the medical routine,
as it directly provides the diagnosis of pathologies or is used as an indicator for the
detection of various diseases. This exam is formed by the erythrogram, leukogram
and platelet, which evaluate the quantity and morphology of erythrocytes, leukocytes,
and platelets, respectively [1].
The erythrogram is the part of the blood count responsible for evaluating the total
number of erythrocytes (RBC), hemoglobin (Hb), hematocrit (Htc), and hematimetric
indices: Mean Corpuscular Volume (MCV), Mean Corpuscular Hemoglobin (MCH)
and Hemoglobin Concentration Mean Corpuscular (CHCM). The total erythrocyte
count is used to detect hematological disorders. Reductions in the number of red
blood cells are called erythropenia, which is a strong indication of the presence
of anemia or blood loss. On the other hand, erythrocytosis is characterized by an
increase in the production of erythrocytes, either by genetic or external factors, is
responsible for the appearance of polycythemias [1, 2].
Anemias can be classified either of genetic origin (falciform anemia, thalassemia,
spherocytosis anemia, among others) or by food deficit (iron deficiency anemia,
megaloblastic anemia, among others). Both types inspire care and medical follow-
up, as food anemias cause indisposition, decrease in work performance, pallor and
vertigo. In turn, genetic anemias need lifelong care, as they are responsible for pain
crises, hospitalizations and the need for periodic blood transfusions. Some types of
polycythemia are of genetic origin and, like genetic anemias, need continuous care,
as it is characterized by the exacerbated production of erythrocytes, causing damage
to the organism such as infarction and strokes [3–5].
Hemoglobin is a protein present inside erythrocytes, being responsible for the
reddish pigmentation of cells and for the transport of O2 and CO2 to tissues, through
the association of these gases with iron molecules. The hematocrit, in turn, expresses
in percentage the amount of blood cells present in the plasma [6].
Hematimetric indices are dependent on the values of hemoglobin, hematocrit
and RBC, with the function of assessing the color and size of erythrocytes. The
size of the erythrocytes is evaluated through the MCV, which classifies these cells
into normocytic, microcytic and macrocytic. Normocytic red blood cells are those
that are normal in size, while microcytes are smaller in size and macrocytes are
larger in volume than those expressed by reference values. Both HCM and CHCM
are responsible for evaluating the erythrocyte staining according to the parameters:
normal coloration, insufficient staining or exacerbated staining, being classified into
normochromic, hypochromic and hyperchromic, respectively [7].
It is important to emphasize that the decrease of just one parameter is not enough
to conclude a diagnosis of anemia or leukemia. The confirmation of these patholo-
gies depends on biochemical or cytochemical tests for the adequate conclusion of
the hematological disorder subtype found. However, the suspicion of these patholo-
gies is often detected through the blood count, which usually presents changes in
more than one of the analyzed parameters [8]. Healthy patients have an RBC count
Hematology: A Review of the Main Methodologies … 51

within the reference values, and the erythrocytes are classified as normocytic and
normochromic. However, patients with iron deficiency anemia, for example, have a
reduction in the total RBC count associated with microcytic and hypochromic red
blood cells [9].
The leukogram is the part of the blood count with the function of evaluating the
granulocyte and agranulocyte leukocytes through the quantitative evaluation of the
total count and by the differential count of the leukocytes. The total WBC count
indicates or rules out the presence of inflammations and/or infections from different
sources. The decrease in the total leukocyte count is called leukopenia and occurs in
cases of patients undergoing cancer treatment or in cases of a sedentary lifestyle. In
turn, leukocytosis is represented by the increase in the amount of leukocytes, due to
the presence of allergies; parasitic, viral or bacterial infections, or leukemias [10–12].
The differential leukocyte count, evaluates the amount of each leukocyte:
neutrophils, eosinophils, basophils, monocytes and lymphocytes. Neutrophils are
the most abundant blood cells, representing about 70% of the total circulating blood
cells. The quantitative increase in neutrophils is called neutrophilia, indicating the
presence of acute bacterial infections. However, neutrophilia is also present in exams
of patients with burns, myocardial infarction, postoperative period, stress, use of
glucocorticoids and intense physical exercises [11, 12].
Neutrophilia is divided into two types: neutrophilia with a left shift and
neutrophilia with a right shift. The left shift is a condition present in cases of acute
infections, when there is the release of a large amount of neutrophil rods into the
bloodstream. In turn, the deviation to the right is characterized by the presence of
segmented or hypersegmented neutrophils (more than 5 lobes), being indicative of
chronic infections. The decrease in the amount of neutrophils is called neutropenia
and may indicate, spinal infiltration, immunosuppression by the human immunode-
ficiency virus (HIV), alcoholism, food deficit, radiation exposure or use of cytotoxic
drugs [13].
Eosinophils are cells that act mainly in the intestines, epithelium and the respi-
ratory system. Thus, eosinophilia is characterized by the numerical increase of
these cells, which is indicative of intestinal infections by helminths (schistosomi-
asis, hookworm, ascariasis, among others), dermatitis, dermatoses, acute myeloid
leukemia, chronic myeloid leukemia, pernicious anemia or allergic processes respi-
ratory. However, the decrease (eosinophilia) or the absence of eosinophils in the
bloodstream is normal. Thus, eosinopenia is usually due to the administration of
adrenocortical or corticoid hormones [13, 14].
Basophils are cells little present in the bloodstream; however, they play an
important role during allergic reactions through the release of IgE and histamines.
Basophilia is a process where there is a greater number of circulating basophils,
being indicative of allergic reactions or the presence of leukemias. Basophenia, in
turn, is not classified as a pathological process [3, 14].
Monocytes, on the other hand, are largely related to phagocytosis processes. Thus,
the increase in the production and release of these cells is called monocytosis. This
increase is suggestive of tissue repairs or indicates the presence of chronic infections.
However, monocytopenia is indicative of decreased monocyte production and release
52 A. C. B. Monteiro et al.

and is often associated with the lifestyle of the patient who is exposed to stress. All
leukocytes present variations in their amounts according to the patient’s age [14, 15].
Lymphocytes are blood cells active in the processes of viral infections. The
increase in the number of these cells in the bloodstream (lymphocytosis) is directly
related to the presence of viral pathologies such as hepatitis, HIV, rubella, among
others [11–14]. The decrease in the amount of lymphocytes (lymphopenia) can be
interpreted such as chronic HIV, acute stress or the presence of liver cirrhosis. It is
important to note that lymphocytosis is a normal condition during childhood. Like
lymphocytes, other blood cells undergo changes in their amounts according to the
stage of life [16].
Thrombocytopenia occurs when the patient has a reasonable drop in the normal
number of platelets. This analysis can be done after counting platelets in a blood
sample. When a patient has thrombocytopenia, bleeding occurs more easily, as well
as the appearance of purple spots on the body. The patient may also have blood in
the stool, bloody vomiting, pain in the joints and muscles and have a weakness.
Thrombocytosis occurs when a patient has a high number of platelets in the
blood, the number is greater than 1,000,000 platelets per mm3 . Thrombocytosis can
be further classified into primary and secondary. The primary is related to myelo-
proliferative diseases (related to the abnormal production of blood cells), and the
secondary is triggered by some underlying disease, such as infections and anemias
[6].
However, in order to arrive at such quantifications of red blood cells and platelets,
specific methodologies for each cell type are employed. Thus, the present study aims
to present in a clear and brief way the main methodologies for total and differential
blood cell counts.

2 Methodology

Considering the need to gather fundamental information regarding basic concepts of

hematology, this study brought together the main concepts present in the literature
regarding hematology, explaining from blood cells to the comparison between the
methodologies employed.

3 Results and Discussion

Regardless of the methodology used to perform the complete blood count, this test
must be performed by collecting 4 ml of venous blood collected in a tube containing
EDTA anticoagulant (“Ethylenediaminetetraacetic acid”) or heparin, which can be
collected with or without fasting of the patient. The samples must avoid the presence
of hemolysis (ruptured erythrocytes causing a lower count in the total amount of
red blood cells), lipemia (presence of fat in the form of triglycerides which makes it
Hematology: A Review of the Main Methodologies … 53

difficult to read results by colorimetric methods), clots (resulting in lower counts of

platelets), use of antibiotics, anti-inflammatory and antiallergic agents hours before
the exam (decreases the amount of leukocytes, being able to mask infectious or
allergic conditions) and avoid smoking 2 h before the exam (change in hemoglobin
rates, hematocrit and indexes hematimetric) [15].
Currently, the blood count can be performed using two methodologies: manual
and automated, which are used according to the needs of each clinical analysis
laboratory. Both the manual methodology and the automated methodology must
be carried out with the help of a health professional such as doctors, pharmacists,
biologists, biomedical and hematologists [1, 5].
The manual methodology is more dependent on the health professional to be
performed, so the time taken to perform the exam is greater. It is also important
to consider that mainly in underdeveloped and developing countries, health profes-
sionals are often subjected to long working hours. This reality directly impacts the
accuracy of the results released. This is because in the manual methodology it is
necessary to count blood cells one by one under optical microscopy. The biggest
advantage of the manual methodology is the cost. This technique depends on low-
cost reagents and instruments, thus being more present in small laboratories and/or
in countries with low per capita income [1, 5].
The automated methodology has a higher cost, especially when applied to low-
income populations, since the cost of acquisition, maintenance and reagents are
passed on to the patient at the time of payment for the exam. However, the investment
is worthwhile, as there is the less direct interaction between the health professional
and the patient’s blood. Thus, there is less chance of accidents involving blood
samples. When such an accident occurs, the professional is exposed to a high risk
of infection with Hepatitis types B, C, D, HIV, Syphilis, Malaria, Chagas disease,
among others. Other benefits are the speed and accuracy of the results obtained, after
all, the equipment is not subject to stress, fatigue and personal problems like a human
[1, 5, 10].

3.1 Manual Erythrogram

The total erythrocyte count in a Neubauer chamber is performed by means of a 1:200

dilution, where 20 µl of whole blood is deposited in a test tube in 4 ml Hayen’s
liquid (sodium citrate diluted in physiological solution) with subsequent homoge-
nization. Subsequently, part of this solution is transferred to the Neubauer chamber,
which is covered by a glass coverslip and analyzed under optical microscopy, where
erythrocytes will be counted under the objective lens at 400× magnification [10, 16].
It is also necessary to make a blood smear, by sliding a few microliters of blood on
a glass slide and later using dyes, which allow the visualization of cellular structures.
These dyes have the function of staining the nuclear and cytoplasmic structures of
blood cells. There is a wide variety of dyes; however, the most used for blood smears
are Leishman and panoptic. Only the final portion of the slide is used for counting,
54 A. C. B. Monteiro et al.

as the anterior portions have clustered and/or overlapping cells, preventing reliable
counting. This blood smear is used to analyze the red cell morphology and staining
[16].
Quantitative and morphological analyzes of blood cells are fundamental parame-
ters for the detection of anemia, since the vast majority of them result in anisocytosis
(variation in shape), such as sickle cell anemia (sickle-shaped red blood cells), sphe-
rocytosis anemia (red blood cells with a circle format), hereditary ellipsis (elongated
red blood cells), among others. Through the blood smear, it is also possible to identify
blood parasites in the acute phase, such as leishmaniasis, nasturtium and malaria,
among others. It is important to emphasize that it is only necessary to make a single
blood smear during the entire blood count analysis process, with the same slide used
for the analysis of red blood cells, leukocytes and platelets [17, 18].
The erythrocytes are responsible for the transport of oxygen, since inside these
cells there is a protein called hemoglobin, which in addition to giving the blood a
red color it binds to oxygen, which allows red blood cells to transport and distribute
oxygen gas from the lungs to inside the cells of the body. The importance of analyzing
these cells goes from both ends, related to the increase in the number of erythrocytes,
is considered polyglobulia (known as polycythemia), is normally not clinically rele-
vant; however, its reduction in relation to the level of red blood cells is considered
hypoglobulia, which is a sign of anemia [18, 19].
Although in addition to assessing erythrocytes and hemoglobin, the eritrogram
also performs a hematocrit check, which it is related to the volume of erythrocytes
found in the centrifuged blood, since through that there is a possibility in determining
the ratio between red blood cells and plasma, relative to the liquid part of the blood
[15].
The determination of hematocrit consists of filling 2/3 of a glass capillary with
whole blood. Subsequently, one end of the capillary must be sealed by means of heat
sources or the use of mass, in order to prevent the escape of blood. This capillary
must be placed in a microhematocrit centrifuge at a rotation of 1500 rpm (rotations
per minute) for 5 min. After centrifugation, there is the separation of whole blood
in 2 phases: a lighter one, formed by plasma, and another more dense, formed by
the sedimentation of blood cells. To read the result, it is necessary to measure in
millimeters the size of the sediment formed at the bottom of the capillary [6, 19].
Hemoglobin, in turn, can be determined using equipment called a spectropho-
tometer. Several methodologies are available in the market for the quantification
of hemoglobin, with variations both in the time spent to perform the exam and in
the amount of sample needed and reagents used during the process. The most used
methodology for the measurement of hemoglobin is the cyanomethoglobin method,
which uses Drabkin’s liquid. This liquid can be composed of potassium ferrocyanide
(K3 FeCN6 ) potassium cyanide (KCN) or anhydrous potassium phosphate (KH2 PO4 )
[5, 20, 21].
Upon contact with Drabkin’s liquid, red blood cells are hemolyzed, with the
release of free Hb, which in turn is converted to methemoglobin (iron hemoglobin in
the iron form). Due to the action of potassium cyanide, methemoglobin is transformed
into cyanomethemoglobin, a stable compound that absorbs light at 540 nm. The
Hematology: A Review of the Main Methodologies … 55

hemoglobin concentration of the sample is calculated from a commercial hemoglobin

standard, pre-determined by the manufacturer [5, 20, 21].
The hematimetric indices, in turn, are determined by means of mathematical
formulas depending on the values obtained for total RBC, hemoglobin and hematocrit
[5, 22]. Equation 1 is used to determine the MCV:

MCV = Htc × 10RBC (1)

where Htc represents the hematocrit value obtained by centrifuging whole blood, and
RBC indicates the total red blood cell count obtained by automated equipment or by
manual counting in a Neubauer chamber. The result is expressed between 80–90 fl,
which indicates the absence of pathologies [5, 22]. In Eq. 2, the HMC determination
is presented:

MCH = Hb × 10RBC (2)

where Hb indicates hemoglobin concentration and RBC is the total red blood cell
count. The reference value is given between 26 and 34 picograms. CMCH, through
Eq. 3:

CMCH = Hb × 100Htc (3)

where Hb represents hemoglobin and Htc the hematocrit [5, 22]. The reference value
is expressed as

36g
31.5 − (4)
dL

3.2 Manual Leukogram

The leukogram assesses white blood cells (WBC), called leukocytes, the defense cells
responsible for fighting invading agents. Leukocytes are actually a group of different
cells, with different functions in the immune system. Some leukocytes directly attack
the invader, others produce antibodies, others only make the identification, and so
on; this test helps in the diagnosis of viral, bacterial or parasitic infections, spinal
dysplasias, leukemias and lymphomas [4].
The counting of WBC is performed through total counting and differential
counting, which are counts with different methodologies, but complementary. It is
important to note that the counts are expressed in absolute and relative values [23].
The total count of WBC is performed in the Neubauer chamber, through a 1:20
dilution, where 0.4 ml of Turk’s liquid and 20 µl of whole blood are deposited in a
test tube. Turk’s liquid can consist of 2% glacial acetic acid or 1% hydrochloric acid,
56 A. C. B. Monteiro et al.

which can be stained with methylene blue or gentian violet in order to differentiate
it from the Hayer liquid [23].
Differential leukocyte counting is performed using a blood smear previously
prepared for the evaluation of erythrocytes. In this process, 100 leukocytes present
in the final portion of the blood smear slide should be counted. It is necessary to note
manually, how many leukocyte subtypes were counted per field. Upon reaching the
total of 100 cells counted, the professional must finish the identification and count
of the leukocytes and perform the calculations of relative WBC and absolute WBC
(value expressed as a percentage) [4, 23].

3.3 Manual Platelet Count

To manual platelet count is need still taking into account that to make thin smears, dry
and stain using the usual methods. Examine under the microscope with an immersion
objective. After count 1000 erythrocytes in different fields, noting the number of
platelets found [8]. Apply the values in the formula:

Number of erythrocytes per mm2

Number of platelets × (5)
1000

3.4 Automated Counting

The automation of the blood count implies greater agility in the performance of
the exams and in the release of the reports; however, they are a more expensive
methodology when compared to the manual methodology. This methodology is able
to count the three cellular types simultaneously. In the 1950s, Coulter Eletronic, Inc
introduced the impedance principle for cell counts [1, 15].
The impedance principle is based on the fact that cells, bad conductors of elec-
tricity are diluted in an electrically conductive solution. This cell suspension is
weighed through an orifice with a diameter around 100 µm, where an electric current
passes. This electric current originates from two electrodes: one located on the inside
of the orifice and positively charged, and the other located on the outside of the orifice,
negatively charged. In this way, each time the cell passes through the hole, it inter-
rupts the electrical current and there is a change in conductance; therefore, each
interruption is counted as a particle [1].
The impedance principle, over the years, has been enabled with counters capable
of measuring cell volume. Such evolution was the result of the color-relation of the
proportionality of the magnitude of the interruption of the electric current (pulses)
with the cell volume. Thus, it was observed that small pulses corresponded to small
volumes, while large pulses were the result of larger volumes [15].
Hematology: A Review of the Main Methodologies … 57

From this correlation between the magnitude of the electric current and the cell
volume, a new concept called the threshold concept was created. The threshold
concept is responsible for classifying cells according to their volume, thus allowing
the detection of globular volume. The globular volume corresponds to the hematocrit
performed on the manual blood count; however, it receives this name because it is
performed without the need for microcentrifugation. Both the impedance principle
and the threshold concept are responsible for the introduction of multiparameter
devices on the market. These devices are capable of performing simultaneous cell
counts using separate channels for counting [1, 15].
In the 1960s, the conductivity technique was developed, based on the high-
frequency electromagnetic current, which is responsible for providing information
regarding the cell volume, the size of the nucleus and the content of the cytoplasmic
granulations. Subsequently, in 1970, the techniques of laser beam dispersion (laser
light scatter) and hydrodynamic fluid were introduced. Both techniques preserve the
leukocyte nuclei and granulations, retracting only the cytoplasmic membrane. These
techniques are based on the principles of diffraction, refraction and reflection of the
emitted light [15].
However, in these techniques, red blood cells are undetectable. To solve this
problem, erythrocytes started to be quantified by means of flow cytometry and hydro-
dynamic focus, where these cells are counted one by one, through an extremely fine
capillary. The red blood cells are then subjected to a laser beam, where the disper-
sion of the light is analyzed at different angles of deviation. In this context, the cell
size is indicated at zero degrees, at 10° the internal structure is indicated and at 90°
the leukocytes are identified and their lobularity characteristics and their granulation
content [1, 15].
Currently, there are a large number of multiparameter devices, which use
impedance, conductivity and scattered light techniques. These technologies can also
be associated with the cytochemical characteristics of cells (such as myeloperoxi-
dase) and the use of reagents that perform the analysis of certain cell types. However,
before purchasing a hematological device, it is necessary to take into account the
following parameters: Automation device versus type of patient attended; number
of hemograms per day × samples/time spent by the device; cost of each hemogram;
quality control; technical assistance; interfacing, training of employees [1, 15].
However, even with the acquisition of hematological equipment, manual
hemogram is not a dispensed practice, being recommended for the confirmation
of hematological reports of pediatric patients, patients over 75 years of age, cancer
patients, patients with suspected leukemia or polycythemia, patients with leukocy-
tosis and patients in severe condition (mainly in a state of hospitalization in the
Intensive Care Unit—ICU) [1, 15].
58 A. C. B. Monteiro et al.

4 Conclusions

The manual blood count is an examination that is totally dependent on human perfor-
mance, which is used for the use of non-automated equipment. It is a cheaper exam,
but more time-consuming and less reliable, as it depends on counts and calculations
performed by health professionals. It is considered a good alternative for small labo-
ratories, where the demand for tests is small and the cost of acquiring hematological
equipment and reagents does not match the cost-benefit of the process.
Over time, the study of blood cells is no longer restricted to just observing the
morphology of cells and has started to develop more specific tests, but that does
not dispense with the use of conventional microscopy. This test responsible for the
analysis of blood cells is called a complete blood count. Thus, review studies are
extremely important to fundament and inspire new research, whether they refer to
studies of cell biology, molecular biology, histology, cytology, hematology, or even
for the creation of medical devices.

References

1. Failace, R.: Hemograma: Manual de Interpretação. Artmed Editora (2015)

2. Monteiro, A.C.B., Iano, Y., França, R.P.: An improved and fast methodology for automatic
detecting and counting of red and white blood cells using watershed transform. In: VIII
Simpósio de Instrumentação e Imagens Médicas (SIIM)/VII Simpósio de Processamento de
Sinais da UNICAMP (2017)
3. Verrastro, T., Lorenzi, T.F., Neto, S.W.: Hematologia e Hemoterapia: Funda-mentos de
Morfologia, Patologia e Clínica. Atheneu, Rio de Janeiro (2005)
4. Fischbach, F.T., Dunning, M.B.: A Manual of Laboratory and Diagnostic Tests. Lippincott
Williams & Wilkins (2009)
5. Monteiro, A.C.B.: Proposta de uma metodologia de segmentação de imagens para detecção e
contagem de hemácias e leucócitos através do algoritmo WT-MO. Dissertation, State University
of Campinas – UNICAMP (2019)
6. Nemer, A.S.A., Neves, F.J., Ferreira, J.E.S.: Manual De Solicitação e Inter-pretação de Exames
Laboratoriais. Revinter (2010)
7. Monteiro, A.C.B., Iano, Y., França, R.P., Arthur, R.: Development of a laboratory medical
algorithm for simultaneous detection and counting of erythrocytes and leukocytes in digital
images of a blood smear. In: Deep Learning Techniques for Biomedical and Health Informatics,
pp. 165–186. Academic Press (2020)
8. Erichsen, E.S., et al.: Medicina Laboratorial Para o Clínico. Elsevier (2015)
9. Souza, M.H.L., Rêgo, M.M.S.: Princípios de Hematologia e Hemoterapia. Alfa Rio (2009)
10. Handin, R.I.; Lux, S.E., Stossel, T.P. (ed.).: Blood: Principles and Practice of Hematology.
Lippincott Williams & Wilkins (2003)
11. Monteiro, A.C.B., Iano, Y., França, R.P., Arthur, R.: Methodology of high accuracy, sensitivity
and specificity in the counts of erythrocytes and leukocytes in blood smear images. In: Brazilian
Technology Symposium, pp. 79–90. Springer, Cham (2018)
12. Monteiro, A.C.B., Iano, Y., França, R.P., Arthur, R., Estrela, V.V.: A Comparative study between
methodologies based on the hough transform and watershed transform on the blood cell count.
In: Brazilian Technology Symposium, pp. 65–78. Springer, Cham (2018)
13. Abbas, A.K., Lichtman, A.H.H., Pillai, S.: Cellular and Molecular Immunology. Elsevier Health
Sciences (2014)
Hematology: A Review of the Main Methodologies … 59

14. Turgeon, M.L.: Clinical Hematology: Theory and Procedures. Lippincott Williams & Wilkins
(2005)
15. Da Silva, P.H., et al.: Hematologia Laboratorial: Teoria e Procedimentos. Artmed Editora (2015)
16. Mcpherson, R.A., Pincus, M.R.: Henry’s Clinical Diagnosis and Management by Laboratory
Methods E-Book. Elsevier Health Sciences (2017)
17. Keohane, E.M., Otto, C.N., Walenga, J.M.: Rodak’s Hematology-E-Book: Clinical Principles
and Applications. Elsevier Health Sciences (2019)
18. Lazarus, H.M., Schmaier, A.H. (eds.): Concise Guide to Hematology. Springer (2018)
19. Cappell, M.S.: Principles and Practice of Hospital Medicine. McGraw-Hill Education Medical
(2017)
20. Cappellini, M.D.: Advances in hemoglobin disorders. HemaSphere 2(S2), 23 (2018)
21. LABEST: Hemoglobina. Labtest, Minas Gerais (2019)
22. Emilse, L.A.M., Cecilia, H., María, T.M., Eugenia, M.M., Alicia, I.B., Lazarte, S.S.: Blood
Res. (2018)
23. Gulati, G.L., Gatalica, Z., Hyun, B.H.: Benign disorders of leukocytes. In: Handbook of
Hematologic Pathology (2019)
An Algorithm Oriented at Obtaining
the Molecular Weight and Concentration
of DNA Samples in Agarose Gel Images

Christian del Carpio , Marcelo Garcia , Julio Elías , David Laván ,

and Guillermo Kemper

Abstract This article presents an algorithm for obtaining the weight and concen-
tration of DNA samples from agarose gel images using digital image processing
techniques. The purpose of this algorithm is to obtain much more accurate results
than the results obtained by experts using their visual sense. The algorithm is based
on capturing the DNA samples and comparing them with DNA molecular weight
markers using digital image processing techniques to find the concentration, which
is the intensity of the samples, and the weight of the samples. The proposed algo-
rithm achieved a moderate “level of agreement” in the results, as calculated by kappa,
compared with the expert results of a total of six images, containing 51 DNA samples
in total. These DNA samples open source is from flies.

Keywords Agarose gel · DNA ladder · DNA · Molecular weight · Image

processing

C. del Carpio (B) · M. Garcia · J. Elías · G. Kemper

Universidad Peruana de Ciencias Aplicadas, Lima, Peru
e-mail: [email protected]
M. Garcia
e-mail: [email protected]
J. Elías
e-mail: [email protected]
G. Kemper
e-mail: [email protected]
D. Laván
Laboratorio de Microgravedad – INICTEL - UNI, Lima, Peru
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 61
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_6
62 C. del Carpio et al.

1 Introduction

Recently, interest in the investigation of the genetic expression of DNA using agarose
gel images has increased because of the increasing interest in finding explanations
about the mutations and diseases that attack genetic information. To prepare the
agarose gel using the inserted DNA samples, the gel must undergo electrophoresis.
This technique is used to separate the DNA fragments according to their size [1].
Previous studies have proposed different image processing algorithms for calculating
the molecular weight and concentration in agarose gels.
Segmentation methods have been proposed that use the histograms of rows and
columns to obtain an accurate cut of the samples [2]. “Enhanced fuzzy c-means”,
has also been proposed, which is an efficient segmentation algorithm in the case of
samples and DNA ladders [3]. Edge detection, geometric diffusion filters, and 2D
noise filters have been used for sample detection [4].
Skutkova et al. [5], propose an algorithm to remove most distortions, such as
non-uniform noise, brightness, and/or contrast, geometrical disproportion, present
in digitized electrophoresis gel images. These distortions affect the detection of
the bands present in the image. This method involves the use of signal processing
techniques such as Dynamic time warping, peak detection, filters, etc. The algorithm
was tested on 50 images, achieving acceptable results.
Abeykoon et al. [6], propose an automated system to analyze and segment the DNA
bands using signal and image processing techniques. The proposed algorithm elimi-
nates defects due to noise and double bands present in the images. Ten images were
analyzed and compared with the result obtained by the proposed system and three
experts who analyzed the images manually. Seven images were correctly detected
and three images showed an error of up to 17% of non-detection of the bands. In the
three images, the error was due to the presence of doubles, which the algorithm did
not detect.
Intarapanich et al. [7], developed an image processing tool that automatically
processes DNA gel electrophoresis images. The tool was originally conceived to
facilitate large-scale genotyping analysis of sugarcane crops. The software allows
for the proper segmentation of DNA bands, even if they are found very close together.
Ten images were analyzed to test the tool. The average accuracy of 78% was achieved
in the segmentation of the lanes in the 10 images. The software was developed in
Java and is available through the “imageJ” framework.
Salazar et al. [8], propose a methodology for processing one-dimensional gel
electrophoresis images containing information on insect and parasite antigens. It
proposes a maximum and minimum algorithm for the identification of each of the
bands contained in the image.
The methods that were reported in the literature generally focused on improving
the segmentation of the samples. However, preprocessing the image before segmen-
tation is also important. Normally, the images1 that are used to conduct these

1 The
images that were analyzed in this project were provided by the “Microgravity Laboratory of
INICTEL-UNI”.
An Algorithm Oriented at Obtaining the Molecular Weight … 63

studies are obtained from the literature because the agarose gels are expensive and
time-consuming to prepare.
The present study proposes an algorithm for the automatic and precise calcula-
tion of the molecular weight and concentration of DNA in the agarose gel images.
The results are compared with the results that were obtained by the experts from
INICTEL-UNI to assess the accuracy of the algorithm. The agreement between the
results is assessed in terms of Cohen’s kappa coefficient which will act as the method
that can be used to evaluate the results. Six images were processed by the algorithm,
yielding a total of 51 samples.

2 Description of the Proposed Algorithm

A flowchart of the proposed method is depicted in Fig. 1 and comprises the following
five steps: image acquisition, the region of interest cropping, rotation, segmentation,
and counting the molecular weights and concentrations.

2.1 Sample Preparation

Three main steps are involved in the preparation of these samples. First, the agarose
is mixed with the buffer in an Erlenmeyer flask to prepare it for electrophoresis.
This solution is heated and poured into the gel mold, where it is left to cool for
30 min. Further, the agarose gel is taken to an electrophoresis bucket. The bucket
is filled with the buffer for electrophoresis, and the DNA samples are placed in
wells. Subsequently, the bucket is electrified for approximately 20 min during the
electrophoresis process. Finally, the agarose gel is placed in a UV transilluminator
to excite the DNA samples.

Fig. 1 Proposed method flowchart

64 C. del Carpio et al.

2.2 Image Acquisition

All the agarose gel samples were digitized using a Canon 12.1 megapixel digital
camera. The images were acquired in the RGB color mode (resolution of 8 bits per
component) with a spatial resolution of 4000 × 3000 pixels and are stored in the jpg
format. Because the three RGB components exhibit an identical value for each pixel,
the grayscale image illustrates a resolution of 8 bits per pixel as shown in Fig. 2.

2.3 Region of Interest Cropping

First, the image must be cropped to include only the region that contains the agarose
gel. Because the RGB components are known to have the same value throughout the
image, only the “R” component can be expressed as I R (x, y).
Equation (1), describes the binarization operation that was applied to the image
I R (x, y). The threshold that was used in the study was obtained by conducting various
tests on the images in which a good result was observed with a threshold of 50. The
resulting image is referred to as I B (x, y), and it is illustrated in Fig. 3. The region of
interest takes the value of “1”, whereas the rest of the image takes the value of “0”.

0, I R (x, y) ≤ 50
I B (x, y) = (1)
1, I R (x, y) > 50

Further, the coordinates of the region of interest are identified, as can be observed
in Fig. 3. The center of the image is precisely the point at which the agarose gel is
observed. The cropped image I R1 (x, y) is depicted in Fig. 4.

Fig. 2 The acquired image

An Algorithm Oriented at Obtaining the Molecular Weight … 65

Fig. 3 The binarized acquired image

2.4 Rotation

The image I R1 (x, y) will be slightly inclined and must be straightened to avoid any
errors during the later stages of processing. The inclination angle θ can be measured
from the upper left part of the image, as depicted in Fig. 4.
Equations (2) and (3) are used to rotate the agarose gel images shown in Fig. 5. The
image Ip (w, z) is the result of rotating the image I R1 (x, y) using the transformation
matrix T.

Ip (w, z) = T {I R1 (x, y)} (2)

Fig. 4 The agarose gel image

66 C. del Carpio et al.

Fig. 5 Image IP (w, z) resulting from the rotation process

⎡ ⎤
cos θ sin θ 0
T =⎢ ⎥
⎢ − sin θ cos θ 0 ⎥ (3)
⎢ 0 0 1⎥

2.5 Segmentation

The rotated image, Ip (w, z), is ready to be segmented. All columns of the image are
extracted, which are the areas at which the DNA ladders and samples are located.
To improve the segmentation process, the average value of pixels in Ip (w, z) is
calculated according to (4), where r j are the gray levels and p j is the probability of
the occurrence of each gray level. Additionally, the standard derivation is calculated
using (5). A range of pixel values is returned; however, the difference between the
values can be disregarded for the present purposes. This range of values is denoted
as black pixels, which are equal to 0, as described in (6), yielding the image Iv (w, z).
This image is illustrated in Fig. 6.

Fig. 6 Image Iv (w, z) of the agarose gel with the disregarded pixels in black
An Algorithm Oriented at Obtaining the Molecular Weight … 67


255
μ= rj × pj (4)
j=0

255
σ = (r j − μ)2 × p j (5)
j=0

0, μ − σ ≤ Ip (w, z) ≤ μ + σ
Iv (w, z) = (6)
Ip (w, z), other case

Further, (7) is used to obtain the horizontal projection Ph (z), of image Iv (w, z),
where N is the height of the image.

N −1

Ph (z) = Iv (w, z) (7)
w=0

The coordinates are obtained from this horizontal projection as shown in Fig. 7,
and each lobe that is present in the project is cut out from the image. This results in
the segmented columns in Iv w  , z  , as can be observed in Fig. 8.

Fig. 7 The horizontal projection of the image

68 C. del Carpio et al.

Fig. 8 Segmented image Iv w  , z 

2.6 Calculating Weight and Concentration

To finally estimate the weight and concentration of the samples, the segmented
columns of the DNA ladders are processed. An example of a segmented DNA ladder
is depicted in Fig. 9.
Equation (8) is used to obtain the vertical projection of each DNA ladder in the
image Iv w  , z  ; Ic (g, f ) is the image of each column of Iv w  , z  , and M is the
width of each column in the image Ic (g, f ). Thus, the position of each mark in the
DNA ladder is calculated, which allows the weights of the samples to be calculated
according to the positions of each DNA sample.

Fig. 9 The segmented DNA ladder

An Algorithm Oriented at Obtaining the Molecular Weight … 69


M−1
Pt (g) = Ic (g, f ) (8)
f =0

Further, the segmented columns in all the samples from the image are used. The
vertical projection in (8) is used again, and only the part of the image that contains
the sample is segmented, as depicted in Fig. 10.
Finally, when all the samples have been segmented, the horizontal position of
each sample is compared with the position of the marks to estimate the molecular
weight of the samples. Figure 11 shows the pixel intensity of the samples and the
marks are compared to calculate the concentration.

Fig. 10 The segmented sample

Fig. 11 Image used to obtain the results

70 C. del Carpio et al.

3 Results

Six agarose gel images were processed using the algorithm. These images were
obtained from six different biopsies.
From these images, we separated all the samples of each image to obtain 51
segmented samples. These samples were compared with the DNA ladder marks
according to the image to which they belong to obtain the results. These were
compared with the results decided by the experts in molecular biology at the
INICTEL-UNI.
It is important to mention that the agarose gel images are open-source DNA of
flies.
Cohen’s kappa coefficient is used to validate the algorithm. This measure indicates
the level of agreement between two evaluators that can be used to classify a result
according to different excluding possibilities. In Table 1, the simplest example is
illustrated using only two possibilities.
The kappa coefficient [9] k is calculated using (9)

P0 − Pe
κ= (9)
1 − Pe

where, P0 (ratio of observed agreement) and Pe (ratio of the agreement by chance)

are calculated using (10) and (11), respectively.

a+d
Po = (10)
n
s3 × s1 + s4 × s2
Pe = (11)
n2
The table for the kappa coefficient was made and the results are depicted in Table 2.
According to Table 3, which lists the levels of agreement for different kappa
coefficients, the agreement obtained with the proposed method is moderate.

Table 1 The possibilities table

Method B
Positive Negative
Method A Positive a c s3
Negative b d s4
s1 s2 n
An Algorithm Oriented at Obtaining the Molecular Weight … 71

Table 2 Results of the calculation of the Kappa coefficient

Specialist Total
Correct reading Wrong reading
Algorithm Correct reading 48 1 49
Wrong reading 1 1 2
Total 49 2 51
P0 = 0.9608 Pe = 0.9245 k = 0.4808

Table 3 Levels of the

Kappa coefficient Level of agreement
agreement to the Kappa
coefficient <0 Without agreement
0–0.2 Poor agreement
0.2–0.4 Fair agreement
0.4–0.6 Moderate agreement
0.6–0.8 Good agreement
0.8–1.0 Very good agreement

4 Conclusions

The study contributes to obtain a better precision of the molecular concentration in

DNA samples, as well as to automate the process. The desired result of obtaining
a moderate “degree of agreement” as assessed by the kappa index was obtained,
the result was 0.4808. However, 51 samples were obtained, an observed agreement
ratio of 96.08%, i.e., a precision error in the results of less than 5%, compared to the
experts’ results.

References

1. Ken Academy: Electroforesis en gel. https://es.khanacademy.org/science/biology/biotech-dna-

technology/dna-sequencing-pcr-electrophoresis/a/gel-electrophoresis. Last accessed 27 Sept
2019 (2018)
2. Caridade, C.M.R., et al.: An automatic Method to identify and extract information of DNA bands
in Gel Electrophoresis Images. In: 31st Annual International Conference of the IEEE EMBS
(2009)
3. Lee, J.D., et al.: Automatic DNA sequencing for electrophoresis gels using image processing
algorithms. Int. J. Biomed. Sci. Eng. (2011)
4. Bajla, I., et al.: Improvemente of Electrophoretic Gel Image Analysis. Austrian Research Centers
Seibersdorf (2001)
5. Skutkova, H., Vitek, M., Krizkova, S., Kizek, R., Provaznik, I.: Preprocessing and classification
of electrophoresis gel images using dynamic time warping. Int. J. Electrochem. Sci. 8(2013),
1609–1622 (2013)
72 C. del Carpio et al.

6. Abeykoon, A., Dhanapala, M., Yapa, R., Sooriyapathirana, S.: An automated system for
analyzing agarose and polyacrylamide gel images. Ceylon J. Sci. (Bio. Sci.) 44(1), 45–54 (2015)
7. Intarapanich, A., Kaewkamnerd, S., Shaw, P., Ukosakit, K., Tragoonrung, S., Tongsima, S.:
Automatic DNA diagnosis for 1D gel electrophoresis images using bio-image processing tech-
nique. In: BMC Genomics Volume 16 Supplement 12, 2015: Joint 26th Genome Informatics
Workshop and 14th International Conference on Bioinformatics: Genomics (2015)
8. Salazar, C., Niño, C., Diaz, R.: Color bands detection on a gel electrophoresis image in one
dimension applying a location algorithm based on maximums and minimums. Iteckne 14(2),
122–130 (2017). https://dx.doi.org/10.15332/iteckne.v14i2.1766
9. Cohen, J.: A coefficient of agreement for nominal scales. Educ. Psychol. Measur. XX(1) (1960)
Classification of Dermoscopy
Skin Images with the Application of Deep
Learning Techniques

Pablo David Minango Negrete , Yuzo Iano ,

Ana Carolina Borges Monteiro , Reinaldo Padilha França ,
Gabriel Gomes de Oliveira , and Diego Pajuelo

Abstract The aid-computer assistant is important for the improvement in the diag-
nostic of skin lesions, principally to detect melanoma, which is the most dangerous
skin cancer type. The principal goal of the aid-computer assistant is to detect
melanoma in its initial stage because the probability of the five-year survival ratio
is over 95%. With the purpose of diagnosed early, in this paper we propose an algo-
rithm based on deep learning to classified dermoscopy images with two types of
skin lesions, which are melanoma or malignant cancer and benign cancer. We use
AlexNet architecture modified to attend our problem, which was trained twice times
with 2000 and 3000 images, divide into three sets (training, validation, and test set).
We analyze three types of optimizers in order to improve the process of learning,
where the best result was obtained by SGD with an accuracy of 99.79% for the
training set, 81.50% for the validation set, and 79.17% for the test set in the scenery
of 4000 images.

P. D. Minango Negrete (B) · Y. Iano · A. C. Borges Monteiro · R. Padilha França ·

G. G. de Oliveira · D. Pajuelo
School of Electrical and Computer Engineering (FEEC), University of Campinas – UNICAMP,
Av. Albert Einstein - 400, Barão Geraldo, Campinas, SP, Brazil
e-mail: [email protected]
Y. Iano
e-mail: [email protected]
A. C. Borges Monteiro
e-mail: [email protected]
R. Padilha França
e-mail: [email protected]
G. G. de Oliveira
e-mail: [email protected]
D. Pajuelo
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 73
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_7
74 P. D. Minango Negrete et al.

Keywords Skin cancer · Dermoscopy images · Melanoma · Non-melanoma ·

AlexNet · Deep learning · CNN

1 Introduction

There are several skin lesion types that can be classified in two ways, which are
melanoma and nonmelanoma [1], where melanoma is the most dangerous because
it is more likely to spread over other parts of the body if not detected, controlled and
treated early. Melanoma is developed by melanocytes (cells that give skin color),
which begins to grow quickly and out of control [2].
The ultraviolet rays that come from the sun are one of the main factors in the
generation of skin cancer because the radiation has enough energy to eliminate an
electron (ionize) from an atom or a molecule, this type of radiation can damage the
DNA in cells, which in turn may lead to cancer [3].
Melanoma can be developer anywhere of the skin but the most common areas to
find melanomas are on the face, neck, and arms, due to are more liable to be exposed
directly to sun rays. Likewise, the risk of having melanoma is almost more than 20
times for white people than people with darker skin color. Another risk factor is
the first-degree relatives (parents, brothers, sisters, or children) have had melanoma.
According to [4], around 10% of all people with melanoma have a family history of
the disease.
In Brazil, skin cancer is one of the most common to appear and correspond about
30% of all malignant tumors registered in the country, melanoma barely represents
about 3% of neoplastic malignancy, be the gravest because have a higher probability
of cause metastasis [5]. Early detection is considered when the melanoma is begin-
ning and in its initial stages when this happens there is an estimated 5-year survival
rate of 95% but if the melanoma is in its latest stages the survival rate drops to 14%
[2].
Melanoma has an elevated lethality, for this reason, its incidence is low. Brazil
has an estimative of 2920 new cases in men and 3340 new cases in women, and most
cases are found in the south region. The skin cancer nonmelanoma estimates 85,170
new cases for men and 80,140 for women.
Among men, the skin cancer nonmelanoma is the most incident in the south region
(160.08/100 thousand), southeast (89.80/100 thousand), and west-central (69.27/100
thousand), in the other regions, the estimated risk are: northeast (53.75/100 thousand)
and north (23.74/100 hundred). Among women, the skin cancer nonmelanoma is the
most incident in all regions of the country with an estimated risk of (97.64/100
thousand) in the south region, (95.16/100 thousand) in the southeast, (92.66/100
thousand) in the west-central, (45.59/100 thousand) in the northeast and in the north
(27.71/100 hundred) [1].
In the last years, the medical assistant by computer takes an important role over
decision making because it aids with a complementary diagnostic over the skin
lesions type that the patient has. Recent studies demonstrate acceptable performance
Classification of Dermoscopy Skin Images with the Application … 75

values of classification, segmentation, detection, and recognition over images with

different skin lesions.
This performance increases thanks to deep learning algorithms, which have excel-
lent results of accuracy at the moment of classification and recognized patterns in
images, several times without the necessity of a previously images processing.
In this paper, we employ AlexNet architecture with the task of classifying images.
We analyze three types of optimizers with the purpose to obtain higher values in the
accuracy. For training, the architecture uses 2000 and 3000 images divided into three
sets (training, validation, and test set).
The paper is organized as follows: Sect. 2, short bibliographic revision is
presented. Section 3, the proposed methodology is described. Section 4, experimental
results are shown, in Sect. 5 the conclusions are discussed.

2 Related Work

Due to the improvement of computers to have more processing power and several
services of GPU computing like Colaboratory from Google, Kaggle, AWS from
Amazon, etc. Nowadays, it is possible to have higher results of accuracy over different
tasks of classification and segmentation of medical images. Deep learning algorithms
that employ CNN architecture have shown excellent results, due to databases with a
large number of images. However, problems such as an imbalance of classes, wrong
labels affect system accuracy.
The principal goal to develop an aid-system to recognize and classify different
skin lesions is due to the diagnosis and recognition of the skin lesion are given
visually by a dermatologist, and it can be complicated and lead to subjective results.
Sabbaghi et al [6] applied a stacked sparse auto-encoders to discover the latent
information features from input image pixel intensities. The high-level learned of
features was employed into a classifier for classifying dermoscopy images. In [6],
this work did not implement CNN.
The idea of using a pre-trained ConvNet was implemented in [7], where it extracted
the features using the filters from a CNN pre-trained on natural images to classified
10 classes of non-dermoscopy skin images. Other techniques of deep learning were
successfully applied to skin lesion analysis like the works of Barata et. al [8], Abbas
et al [9], and among others.

3 Methodology

In this section, we describe the architecture type chosen and which parameters were
modified with the purpose to attend our problem of binary classification between
benign and malignant cancer. Furthermore, the dataset and implementation aspects
are described.
76 P. D. Minango Negrete et al.

3.1 Dataset Description

For our experiments, we use the open dataset of International Skin Imaging Collab-
oration (ISIC) [10]. This dataset contains about 23,906 images with different skin
lesions with their own histopathological diagnostic. Due to the imbalance of classes,
we focus on two types of lesions, which are benign and malignant skin cancer,
because it is one of the principal’s objectives of this organization try to reduce the
mortality caused by malignant skin cancer known as melanoma too. Thus, we use
and analyze 1000 and 2000 images for each class, which were chosen randomly.

3.2 Architecture Description

The proposed binary classification between benign and malignant cancer is based
in AlexNet architecture [11], which is a well-documented and usually used for the
convolutional neural networks in image classification problems.
The flow of the AlexNet architecture of the proposed system is shown in Fig. 1,
AlexNet consists of five convolutional layers with different sizes of filters or kernels,
which have the principal function of decrease the dimensionality of the image and
extract the information.
The first block of Fig. 1. is the input of the architecture where the images are
resized in 200 × 200 pixels RGB, and for the huge quantity of images is necessary
to divide into batches of images in order to not overload the system. The next section
of blocks is the convolutional layers where the size of kernels are: C1(11 × 11),
C2(5 × 5) and C3, C4 and C5 (3 × 3).
In each one of these layers, the Rectified Linear Unit (ReLU) is used as function
activation, likewise, each block integer the concept of MaxPooling [12] for reducing
the number of dimensionalities with the help of a filter which takes the maximum
number of the window. The size of these filters is (3 × 3). After each process in
convolutional layers, we normalize the output with the BatchNormalization [12], for
stabilizes the learning process and try to have a faster convergence.

Fig. 1 AlexNet architecture adapted to our problem

Classification of Dermoscopy Skin Images with the Application … 77

The last three blocks correspond to the dense layers, in the F1 and F2 we apply
Dropout [12], which permits the probability of turn off certain quantities of neurons
randomly, but maybe they activate in the next step with the goal of reducing the
overfitting. In the last block, we choose two neurons, which can try to predict the
true label of the class. As our problem is a binary classification, a better activation
function is a sigmoid because it is the most utilized to predict values between 0 and
1.
For compiling the architecture, we analyze the performance of three types of
optimizers, which are Adam, RMSprop [13], and Stochastic Gradient Descent (SGD)
[14]. The optimizers have an initial learning rate of 0.001. Binary cross entropy was
used as a loss function [15]. Early stopping was applied to interrupt the training;
thus, it was monitoring the validation accuracy (AUC) and if do not improve after
10 epochs, the training is stopped and save the best performance.

4 Results and Discussion

All experiments presented are trained with AlexNet architecture. They are performed
by using Keras package [16] as a deep learning framework in order to build and train
the model. For results visualization, we employed the Matplotlib package [17]. All
codes are implemented in the Kaggle environment, in its GPU mode, where it works
with 13 Gb of Ram and 2 cores.
The dataset employed was random split in three sets, where for the first training
of AlexNet, we employed 1400 images for training (700 images for each class), for
the validation set 400 images were employed (200 for each class), and for the test
set we used 200 images.
The same way for the second training, we increase the number of images in each
one set, leaving the experiment thereby with 2400 images of the training set (1200
for each class), 1000 images for validation set (500 for each class) and 600 images
for the test set. All sets have the same quantity of images to resolve the problem of
imbalanced classes with the purpose of not affect the accuracy of the model.
For the evaluation of our methodology, we use the following metrics in the three
sets of data.
• Accuracy represents the correct predictions in each one class, divided for the total
numbers of predictions.
• Loss measures the wrongly classified images with respect to the true label.
Table 1 shows the performance of the first training of the model, the training values
are tested in [18], where we compare with the validation values. The accuracy is not
high as the training, the SGD is the optimizer that achieves the better percentage of
learning so much for the two sets of data with 97.66% for the training set and 78%
for the validation set.
Figure 2 shows the efficiency of the three optimizers, where we note that the better
accuracies are attended by the training set, while the validation set attends until 78%
78 P. D. Minango Negrete et al.

Table 1 Accuracy and loss comparisons of the different types of optimizers into the training and
validation sets
Adam RMSprop SGD
Training Acc 96.66 87.64 97.66
Loss 0.0817 0.3141 0.0567
Validation Acc 75.50 70.25 78.00
Loss 0.8227 0.7583 0.4360
Bold value represents the better values of the learning

Fig. 2 Performance of AlexNet architecture, training, and validation accuracy

of learning. It is important to take attention with the loss values because here we
appreciate if the model is generalized or it is overfitting. In our case, the optimizer
Adam is a clear case of overfitting, however, SGD has an acceptable value of the
loss.
Table 2 shows the values obtained in the second training, as the first evaluation,
the better optimizer is SGD for both sets of data, we note that the values increased
with more images. SGD attends an accuracy of 81.50% in the validation set which
improves 3.5% over learning binary classification.
We appreciate that SGD shows a learn most permanent for training and validation
set. The other optimizers, the learning is more variable. This is appreciated in the
detailed process of training and validation shown in Fig. 3.
After analyzing the performance of these three optimizers with different quantities
of images, we proceeded to take SGD which attends better accuracy in the two training
for evaluated the learning over the test set. Table 3 shows that when we have more
Classification of Dermoscopy Skin Images with the Application … 79

Table 2 Accuracy and loss comparisons of the different types of optimizers into the training and
validation sets
Adam RMSprop SGD
Training Acc 97.85 89.72 99.79
Loss 0.0591 0.2434 0.0120
Validation Acc 77.70 79.60 81.50
Loss 0.8047 0.5528 0.4098
Bold value represents the better values of the learning

Fig. 3 Performance of AlexNet architecture, training and validation accuracy

Table 3 Evaluation of the two training over the test set

T1 T2
Test Acc 70.50 79.17
Loss 0.2950 0.2083

quantity of images the learning is better. In the case of the second training (T2) we
obtain an accuracy of learning of 79.19%, which represents an increase of 8.67% in
comparison to the first training.
80 P. D. Minango Negrete et al.

5 Conclusion

A modified AlexNet architecture was proposed for aid in the early detection of
melanoma skin cancer, which represents a binary classification to identify between
skin cancer non-melanoma and melanoma. We tested it with different quantities of
images divided into three sets (training, validation and test set), where we observed
that SDG is the better optimizer with an attendance of performance in the learning
of 99.79% in the training, 81.50% in the validation and 79.17% in the test set. These
results were obtained in the second training for 4000 images, stating the potential
of the proposal. It is important to take attention in the results on the loss evaluation
because in this metric, we can observe if the architecture of our model was generalized
the learning or it was fall in the overfitting.

References

1. INCA: Estimativa 2018 – síntese de resultados e comentários. Available http://www1.inca.gov.

br/estimativa/2018/sintese-de-resultados-comentarios.asp. Accessed 11 Sept 2019 (2016)
2. Society, A.C.: What is melanoma skin cancer? What is melanoma? Available https://www.can
cer.org/cancer/melanoma-skin-cancer/about/what-is-melanoma.html.. Accessed 11 Sept 2019
(2017)
3. Society, A.C.: Ultraviolet (uv) radiation. Available https://www.cancer.org/cancer/cancer-cau
ses/radiation-exposure/uv-radiation.html. Accessed 13 Sept 2019 (2017)
4. Society, A.C.: Melanoma skin cancer risk factors: Melanoma risk factors. Available https://
www.cancer.org/cancer/melanoma-skin-cancer/causes-risks-prevention/risk-factors.html.
Accessed 13 Sept 2019 (2017)
5. INCA: Câncer de pele melanoma. Available https://www.inca.gov.br/tipos-de-cancer/cancer-
de-pele-melanoma. Accessed 13 Sept 2019 (2018)
6. Sabbaghi, S., Aldeen, M., Garnavi, R.: A deep bag-of-features model for the classification of
melanomas in dermoscopy images. In: 2016 38th Annual International Conference of the IEEE
Engineering in Medicine and Biology Society (EMBC), Aug 2016, pp. 1369–1372
7. Kawahara, J., Ben Taieb, A., Hamarneh, G.: Deep features to classify skin lesions. In: 2016
IEEE 13th International Symposium on Biomedical Imaging (ISBI), April 2016, pp. 1397–1400
8. Barata, C., Ruela, M., Francisco, M., Mendonça, T., Marques, J.S.: Two systems for the detec-
tion of melanomas in dermoscopy images using texture and color features. IEEE Syst. J. 8(3),
965–979 (2014)
9. Abbas, Q., Celebi, M., Serrano, C., García, I.F., Ma, G.: Pattern classification of dermoscopy
images: A perceptually uniform model. Pattern Recogn. 46(1), 86–97. Available http://www.
sciencedirect.com/science/article/pii/S0031320312003457 (2013)
10. ISIC Project: Public skin lesion image archive. https://www.isic-archive.com/!/topWithHeader/
onlyHeaderTop/gallery, 2018
11. Krizhevsky, A., Sutskever, I., Hinton, G.E.: Imagenet classification with deep convolutional
neural networks. Neural Inf. Process. Syst. 25, 01 (2012)
12. Géron, A.: Hands-On Machine Learning with Scikit-Learn and TensorFlow: Concepts, Tools,
and Techniques to Build Intelligent Systems, 1st edn. O’Reilly Media, Inc. (2017)
13. Kingma, D.P., Ba, J.: Adam: a method for stochastic optimization. In: Cite
arxiv:1412.6980Comment: Published as a Conference Paper at the 3rd International Conference
for Learning Representations, San Diego (2015). Available http://arxiv.org/abs/1412.6980
14. Ruder, S.: An overview of gradient descent optimization algorithms. CoRR, vol.
abs/1609.04747 (2016). Available http://arxiv.org/abs/1609.04747
Classification of Dermoscopy Skin Images with the Application … 81

15. Janocha K., Czarnecki, W.M.: On loss functions for deep neural networks in classification.
CoRR, vol. abs/1702.05659, (2017). Available http://arxiv.org/abs/1702.05659
16. Chollet, et al.: Keras. https://keras.io (2015)
17. Hunter, J.D.: Matplotlib: a 2d graphics environment. Comput. Sci. Eng. 9(3), 90–95 (2007)
18. Minango, P., Iano, Y., Monteiro, A.C.B., França, R.P., de Oliveira, G.G.: Classification of
automatic skin lesions from dermoscopic utilizing deep learning. SET Int. J. Broadcast Eng.
ISSN Print 2446-9246 ISSN Online: 2446-9432. https://doi.org/10.18580/setijbe.2019.14. Web
Link http://dx.doi.org/10.18580/setijbe.2019.14 (2019)
Association Between Human
Papillomavirus Infection and Anus
Cancer Precursor Lesions: A Review

Ana Carolina Borges Monteiro , Yuzo Iano , Reinaldo Padilha França ,

Rangel Arthur , Pablo David Minango Negrete , Diego Pajuelo ,
Gabriel Gomes de Oliveira , and Lisber Arana Hisnostroza

Abstract Human papillomavirus, commonly referred to as HPV, is a virus with

tropism by differentiating tissues. Its pathogenesis is related to the disorder of genes
that inhibit cell apoptosis and cell suppression, this fact favors its action and spread by
an organism. Due to these characteristics, HPV is associated with cervical, anus, head,
and neck cancer. Based on this, the present work has as objective the accomplishment
of a bibliographical revision study of the main characteristics of this virus as well as
its effects on the human organism. Studies such as this can be seen as key elements
for understanding and preventing this sexually transmitted disease (STD), which has
its highest incidence rates in underdeveloped and developing countries, where health
and education policies are often scarce or nonexistent.

Keywords Human papillomavirus · Cytopathology · Molecular biology · Sexually

transmitted disease · Diagnostic methods

1 Introduction

Human papillomavirus (HPV) is a 72 capsomer non-enveloped virus belonging to

the Papovaviridae family, with a mean diameter of 55 nm. The genome of these

A. C. Borges Monteiro (B) · Y. Iano · R. Padilha França · P. D. M. Negrete · D. Pajuelo ·

G. G. de Oliveira · L. A. Hisnostroza
School of Electrical and Computer Engineering (FEEC), University of Campinas – UNICAMP,
Av. Albert Einstein - 400, Barão Geraldo, Campinas, SP, Brazil
e-mail: [email protected]
Y. Iano
e-mail: [email protected]
R. Padilha França
e-mail: [email protected]
R. Arthur
School of Technology (FT), University of Campinas – UNICAMP, R. Paschoal Marmo - 1888,
Jardim Nova Italia, Limeira, SP, Brazil

© The Editor(s) (if applicable) and The Author(s), under exclusive license 83
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_8
84 A. C. B. Monteiro et al.

viruses consists of eight to nine open reading frames and one regulatory region. This
virus has double circular deoxyribonucleic acid (DNA) containing about 7900 base
pairs and genes encoding eight proteins: E1, E2, E4, E5, E6, E7 (non-structural),
and L1 and L2 (structural) that form the viral capsid with icosahedral symmetry. L1
protein when produced in a heterologous expression system has the ability to self-
assemble. The genome of this virus is organized into three regions: early region (E),
late region (L), and regulatory region (URR). The L1 and L2 genes are responsible
for encoding viral capsid proteins and genes, as well as for the coding of proteins
involved in viral replication and cell transformation. E1 and E2 genes are involved
in viral replication, while E5, E6, and E7 genes encode proteins responsible for this
cellular transformation. HPV infections can be divided into two types: cutaneous or
mucosal [1, 2].
More than 200 types of HPV are currently described in the literature, which is
classified according to their oncogenic potential. Approximately 45 types of this
virus infect the male and female anogenital tract epithelium 3. HPV subtypes can be
classified as low risk (types 6, 11, 42, 43, and 44) and high risk (types 16, 18, 31, 33,
35, 39, 45, 46, 51, 52, 56, 58, 59, and 68) [3, 4].
Since the 1970 s, German scientist Harald zur Hausen of the Cancer Research
Center (Heidelberg—Germany) has devoted himself to the study of HPV. In her
research, there was evidence of the association of subtype 16 Human papillomavirus
with biopsies with positive reports of women with cervical cancer. Shortly after this
discovery, Hausen also associated subtype HPV 18 with the same pathology. Both
subtypes are found in approximately 70% of these biopsies. In the year 2008, Harald
zur Hausen won the 2008 Nobel Prize in Medicine for his research on HPV. Based
on these findings, further research on this virus has been boosted [3, 4].
This infection is due to unprotected sexual contact, which through micro-
abrasions, facilitates the penetration of the virus in the deep layer of epithelial tissue.
However, the infection can also occur through direct or indirect contact with lesions
on other parts of the body. There is also the possibility of vertical transmission during
pregnancy or at delivery. Low-risk oncogenic lesions are manifested as a common
wart, genital wart, or condyloma [5].
The classic risk factors for both oncogenic and non-oncogenic development are
early onset of sexual intercourse, multiplicity of lifelong sexual partners and high
parity (non-surgical births), being young, being a smoker, and being of low socioe-
conomic status, prolonged oral contraceptive use, nutritional deficiencies, human
immunodeficiency virus infection (HIV), and other genital infections caused by
sexually transmitted agents, such as Chlamydia trachomatis, Herpes simplex virus)
[6].
To understand the oncogenicity of HPV, it is necessary to consider that when the
daughter cells immediately enter the G1 phase and are capable of either restarting
the cell cycle or performing a temporary or permanent halt. The G1 phase has a
checkpoint, G1/S, controlled by the pRb (retinoblastoma protein) pathway. Once the
cell passes this point, it is compelled to replicate its DNA. If an incorrect copy of DNA
occurs during S or DNA damage, then the cell will not pass the G2/M checkpoint,
and gene-induced growth arrest and apoptosis will occur. In this process, several
Association Between Human Papillomavirus Infection … 85

inhibitor proteins can stop the advance of this cycle, such as p15 and p16, which
block fundamental components for cell cycle progressions, such as cyclin-dependent
kinases (CDK) and cyclins, preventing the advance of the phase cycle. G1 to S [7, 8].
Other inhibitors are p21, associated with the proto-oncogene ras, and p53, which
are responsible for monitoring the quality, integrity of their chromosomes, and the
correct execution of the different phases of the cycle. Human cells are equipped with
mechanisms for controlling cell division. Mutations in the genetic content of these
cells can overcome these defenses and contribute to cancer formation. One such
mechanism of action is apoptosis, which occurs when fundamental components are
damaged or production system control is unregulated. The development of tumor
cells implies an escape from this mechanism. Protein p53, among its many functions,
assists the onset of apoptosis, reducing the chance of genetically damaged cells being
eliminated by initiating a carcinogenic process. Another mechanism for controlling
cell division limits the number of times a given cell reproduces. In this mechanism,
the tips of the chromosomes (telomeres) mark the number of divisions, and at the
appropriate time initiate senescence and death at the time of telomerase. Activation of
this enzyme induces cell immortalization, an indispensable event for carcinogenesis
[7, 8].
Within this context in carcinogenesis two classes of genes are fundamental for
the development of cancer, because under normal conditions they are responsible
for cell cycle control. Proto-oncogenes stimulate, while suppressor genes inhibit the
processes of cell division. Collectively, these two classes of genes are responsible for
the uncontrolled proliferation found in cancers in humans. When mutations occur,
proto-oncogenes become oncogenes, which are carcinogenic and cause excessive cell
multiplication. These mutations cause the proto-oncogene to overexpress its growth-
stimulating protein or to produce a more active form. In contrast, tumor suppressor
genes when inactivated by mutations contribute positively to cancer when. The result
of these combinations is the loss of action of functional suppressor genes, which
deprives the cell of controls crucial for inhibiting inappropriate growth [7, 8].
Due to the dissemination of information about the incidence of STD, as well as
its oncogenicity capacity, there is an increase in condom use among young people
in some countries [9]. Therefore, the need to develop bibliographical review studies
is noteworthy. Thus, the present study aims to present to the scientific community
as well as the general population the main information about HPV that affects the
anal regions so that this information is transmitted in a clear, concise, and updated
way; after all, HPV in regions anal diseases is as risky as HPV housed in the uterine
cervix region, whereas it is less addressed by the general media.

2 Methodology

The present study was performed based on the analysis of scientific works related
to genital HPV, as well as diagnostic methods of this pathology. These studies were
analyzed focusing on scientific publications conducted between 2014 and 2019.
86 A. C. B. Monteiro et al.

3 Results and Discussion

The anal canal is histologically characterized by the presence of a multi-stratified

squamous lining. The anal canal is located from the anal border to the pectineal line,
which is characterized by being a transition zone between the squamous and non-
squamous epithelium. In turn, the area that delimits the anal canal and rectum is made
up of transition cells, which resemble the cells that make up the urogenital lining
and rectal glandular mucosa. Thus, the appearance of tumors in the distal region
of the pectineal line is called squamous cell carcinomas, or epidermoids, which are
classified as keratinized. However, those originating from tissues above the dentate
line are called epithelioid, which are non-keratinized. Women with cervical cancer
are considerably more susceptible to the development of anus or rectal canal cancer
[10].
Anal epithelial and anal canal neoplasms, besides presenting differences in embry-
onic origin, also present different segments for lymphatic drainage, according to
their location before the pectineal line. Just above the pectineal line, the lymph flows
toward the paravertebral lymph nodes and the perirectal spaces, which are indica-
tions for the diagnosis of rectal adenocarcinoma. Just below the pectineal line, lymph
flows toward the nodes of the inguinal region, being characteristic for the detection
of inguinal adenopathies [11].
Tumor severity and clinical presentation depend on anal size and location. Lesions
less than 2 cm are classified as small and have a probability of cure in 80% of cases.
However, tumors of 5 cm or more are usually invasive and compromise the lymph
nodes, thus presenting less than 50% chance of cure [12].
Just as HPV has tropism in the transitional zones of the cervix, HPV in the anal
canal has a preference for the anal transition zone (ZTA). Based on this feature,
the anal Pap smear test began. This exam consists of the removal of anal cells for
cytological studies by introducing anal swab or endocervical brush in the anal canal.
For this, the swab should be introduced between 3 and 4 cm inside the anal canal.
This occurs because the length of the anal canal varies from 2.52 to 2.93 cm in women
and from 3.27 to 3.4 cm in men. After the introduction, the swab or brush should be
rotated 360º around the axis. The material should be uniformly deposited on a glass
slide and subsequently fixed and stained properly [13].
The cytological analysis performed on anal smears aims to search for columnar
epithelium cells, squamocolumnar cells (coming from the transition zone) and squa-
mous cells. High-grade anal intraepithelial neoplasias (AIS) are considered precur-
sors of anal carcinomas, presenting a high relationship with infection with high
oncogenic HPV risk subtypes, mainly by subtypes 16 and 18. These subtypes have
their DNA identified between 35 and 90% of tumor cells [13, 14].
The risk of progression is associated with the severity of dysplasia, and HPV
virus classification is directly linked to premalignant and malignant lesions, without
correlation with the recurrence rate of lesions in the anal region. Thus, the treatment
of precursor lesions aims to prevent the evolution of squamous cell carcinoma [15].
Association Between Human Papillomavirus Infection … 87

AIS are classified as synonymous with severe dysplasia, severe atypia, and in situ
carcinoma. This pathology has a higher incidence in squamous cells, being respon-
sible for the cytological picture of binucleation or multinucleation, the presence of
koilocytes and atypical paraceratocyte cells in addition to moderate dysplasia. It is
noted that most anal carcinomas follow a cytological pattern similar to that found in
cervical carcinomas [18].
The biological effects caused by infections with other STDs are related to the pres-
ence of persistent inflammation accompanied by damage through oxidative metabo-
lites. Infected cells are prone to increased cytokine secretion. When this infection
becomes chronic there is the induction of tissue damage by indirect production
of reactive oxygen species, thus triggering the inflammatory cascade. As a conse-
quence of this process, there is a decrease in immunity. In addition, if the patient
is a smoker, immunosuppression will be even higher, as nicotine is a determining
factor for Langerhans cell depression. Thus, the somatization of immunosuppressor
cofactors further facilitates the installation of the HPV virus [18].
The anal canal is more affected than the anal margins, and the transition zone
between the anus and rectum is the most affected by tumors. The transition zone
has areas of normal rectal mucosa and the presence of squamous epithelium, with
the presence of cells of various sizes arranged in arrangements, also has microvilli
arranged in columns. This region is affected by high-grade squamous intraepithelial
lesions, which are associated with infections caused by high-risk oncogenic HPV
subtypes. When untreated, the lesions may progress to invasive carcinoma [17, 18].
About 20% of affected patients are asymptomatic, the others usually have
anorectal bleeding or pain usually accompanied by the sensation of mass occupying
the rectal region. This symptomatology, however, is often mistakenly mistaken for
hemorrhoidal diseases. Anal oncogenic low-risk HPV infection in the latent form
initially manifests itself through anal pruritus, with later onset of warts. However, the
presence of anal pruritus is not a determining factor for the closure of this diagnosis,
since pruritus has a multifactorial etiology such as inflammatory bowel disease,
uremia, parasitic diseases, dermatopathies, inadequate hygiene habits, inadequate
clothing, the presence of STDs, among others [19].
In turn, high-risk anal oncogenic HPV presents difficulties in its diagnosis, since
the lesions are subclinical and their detection is dependent on anoscopy exams. As
the lesions are inapparent and usually asymptomatic, the individual seeks medical
attention only when the lesions have progressed. Anal border tumors are classi-
fied as dermatological lesions and may develop as verrucous carcinoma, Kaposi’s
sarcoma, Bower’s disease, or squamous cell carcinoma. However, those that affect the
rectal transition zone or rectal canal are considered malignant and have four origins
histological: squamous cell carcinoma, malignant melanoma, adenocarcinoma, and
sarcomas [19, 20].
88 A. C. B. Monteiro et al.

4 Diagnostic Methods

The most suitable diagnostic methods for detecting the presence of HPV in the anal
region are anoscopy and anal Pap smear. Anuscopy consists of applying 5% acetic
acid to the anorectal mucosa for two minutes. Subsequently, the excess is removed
with a physiological solution, followed by toluidine blue. The examination should
be performed using a colposcope. The result is observed through the presence of
ketoacid reactions, which appear in coloration whitish when positive, indicating
HPV infection [20].
Polymerase chain reaction (PCR) is a high sensitivity test and is used to prove
the existence or the absence of HPV DNA. The PCR reaction is the amplification of
viral DNA (HPV) using conserved sequences from specific HPV regions as primers.
This test also enables the identification of the HPV genotype by amplifying specific
regions for each of the high- or low-grade viruses [21, 22].
Recent studies show that increased expression of miR-16, miR-25, miR-92a, and
miR-378 and reduced expression of miR-22, miR-27a, miR-29a, and miR-100 are
attributed to viral oncoprotein. E6 or E7. In the future, miRs are expected to be
powerful tools for early detection of HPV regardless of their subtype and site of
infection [23].
Researchers have developed a methodology for the detection of urine high-risk
oncogenic HPV. This technique is based on the detection of virus DNA present in the
samples. The results obtained were 90.9% accurate in identifying cervical lesions
with abnormal cells [24].
In recent years, there have been a growing number of researches related to
pathology detection through the digital processing of medical images [25–27]. In
this context, it is expected that in the coming decade’s methodologies will be devel-
oped that are capable of detecting precursor lesions and cancerous lesions of HPV,
similarly to blood cell detection and counting research. These studies have shown
high sensitivity, accuracy, specificity, low computational and runtime time consump-
tion, and low cost. Cost reduction is paramount for poor people’s access to quality
health [25, 26].
In addition, it is important to note that in the not too distant future, the results of
cytopathological examinations, patient anamnesis, and data from health institutions
can be transmitted and shared by real-time telecommunication channels without data
loss and low memory consumption of the devices involved regardless of the type of
operating system involved (Android, Windows, Linux, iOS, among others) [27, 28].
Such action is paramount, because the faster the doctor has access to the medical
reports, the faster the patient is referred for treatment, and consequently the chances
of cure increase.
Association Between Human Papillomavirus Infection … 89

5 Conclusions

The association between HPV and malignant neoplasms shows that this virus is
responsible for the precursor lesions that direct individuals to cancer development.
However, with recent advances in molecular biology coupled with cytopathology
practices, they have led to the possibility of early and more accurate diagnosis of
the subtype involved. The classification of the HPV subtype involved in lesions is
paramount for the correct management of the disease. Even with such advances,
prevention through education focused on adolescents, youth, and adults is the best
way to ensure the health of individuals. Only with massive awareness activities can
we avoid future HPV complications with other STDs. Importantly, not all STDs are
curable, some should only be followed and controlled throughout the patient’s life.
Such information is often not disseminated as it should. In addition to prevention, the
establishment of telemedicine-based data sharing networks is important to promote
better treatment and follow-up of patients who live in remote regions or where there
is a shortage of professionals specialized in this type of clinical management. In addi-
tion, it is necessary to create complex databases that, using artificial intelligence tech-
niques, are able to compile data on cytopathology, molecular biology, dermatoscopy,
cytology, and histology of patients affected by HPV. Thus, the medical diagnosis
may become increasingly accurate and less likely to fail to identify and classify the
HPV subtype; after all, an incorrect medical diagnosis may lead to the patient’s death
or irreversible damage.

References

1. Doorbar, J., Egawa, N., Griffin, H., Kranjec, C., Murakami, I.: Human papillomavirus molecular
biology and disease association. Rev. Med. Virol. 25, 2–23 (2015)
2. Maxwell, J.H., Khan, S., Ferris, R.L.: The molecular biology of HPV-related head and neck
cancer. In: Fakhry, C., D’Souza, G. (eds.) HPV and Head and Neck Cancers. Head and Neck
Cancer Clinics. Springer, New Delhi (2015)
3. Okami, K.: A new risk factor for head and neck squamous cell carcinoma: human papillo-
mavirus. Int. J. Clin. Oncol. 21(5), 817 (2016)
4. Mammas, I.N., Spandidos, D.A.: Paediatric virology as a new educational initiative: an inter-
view with Nobelist Professor of virology Harald zur Hausen. Exp. Therap. Med. 14(4),
3329–3331 (2017)
5. Jesus, S.P.D., et al.: A high prevalence of human papillomavirus 16 and 18 co-infections in
cervical biopsies from southern Brazil. Braz. J. Microbiol. 49, 220–223 (2018)
6. Ribeiro, A.A., Costa, M.C., Alves, R.R.F., Villa, L.L., Saddi, V.A., dos Santos Carneiro, M.A.,
Rabelo-Santos, S.H.: HPV infection and cervical neoplasia: associated risk factors. Infect.
Agents Cancer 10(1), 16 (2015)
7. Cseke, L.J., Kirakosyan, A., Kaufman, P.B., Westfall, M.V.: Hand-Book of Molecular and
Cellular Methods in Biology and Medicine. CRC Press (2016)
8. Erkekoglu, P.: Oncogenes and Carcinogenesis (2019)
9. Abreu, M.N.S., et al.: Conhecimento e percepção sobre o HPV na popu-lação com mais de 18
anos da cidade de Ipatinga, MG, Brasil. Ciência & Saúde Coletiva 23, 849–860 (2018)
10. Allen, D.C., Cameron, R.I. (eds.): Histopathology Specimens: Clinical, Pathological and
Laboratory Aspects. Springer (2017)
90 A. C. B. Monteiro et al.

11. Minsky, B.D., Guillem, J.G.: Neoplasms of the anus. In: Holland-Frei Cancer Medicine, pp. 1–
12 (2016)
12. Júnior, J.C.M.: S: Câncer Ano- reto-cólico - Aspectos atuais: I - Câncer Anal. Rev bras
Coloproct 27(2), 2109–2223 (2007)
13. Nadal, S.R., et al.: Quanto a escova deve ser introduzida no canal anal para avaliação citológica
mais eficaz? Rev. Assoc. Med. Bras. 55(6), 749–751 (2009)
14. Duarte, B.F., da Silva, M.A.B., Germano, S., Leonart, M.S.S.: Anal cancer diagnosis in patients
with human papilomavírus (HPV) and human immunodeficiency virus (HIV) coinfection. Rev.
Inst. Adolfo Lutz 75, 1710 (2016)
15. Monteiro, A.C.B., da Cruz Pires, D.V.D.: Characterization of the risk factors for anus cancer
and its relationship with Human Papillomavirus-es. Rev Saude em foco (2015)
16. Chaves, E.B.M., Capp, E., Corleta, H.V.E., Folgierini, H.J.: A citologia na prevenção do câncer
anal, vol. 39, no. 11, pp. 532–537. Rio de Janeiro, Femina (2011)
17. Cuevas, M.: Virus del papiloma humano y salud femenina. Ediciones i (2019)
18. Magalhães, M.N.: Carcinoma epidermóide do canal anal (2016)
19. Cutrim, P.T.: Papilomavírus humano (hpv) e sua associação entre as le-sões cervical e anal em
mulheres (2017)
20. Darragh, T.M., Palefsky, J.M.: Anal cytology. In: The Bethesda System for Reporting Cervical
Cytology, pp. 263–285. Springer, Cham (2015)
21. Bernardy, J.P., Bierhals, N.D., Possuelo, L.G., Renner, J.D.P.: Padro-nização da PCR em tempo
real para a genotipagem de HPV 6-11, HPV 16 e HPV 18 utilizando controle interno. Revista
Jovens Pesquisadores 8(1), 37–48 (2018)
22. Clifford, G.M., et al.: Comparison of two widely-used HPV detection and genotyping methods:
GP5+/6+ PCR followed by reverse line blot hybridization and multiplex type-specific E7 PCR.
J. Clin. Microbiol. JCM-0061 (2016)
23. Wang, X., et al.: microRNAs are biomarkers of oncogenic human papillomavirus infections.
Proc. Natl. Acad. Sci. U.S.A. 111(11), 4262–4267 (2014)
24. Allison, D.B., Olson, M.T., Maleki, Z., Ali, S.Z.: Metastatic urinary tract cancers in pap test:
cytomorphologic findings and differential diagnosis. Diagn. Cytopathol. 44(12), 1078–1081
(2016)
25. Monteiro, A.C.B., Iano, Y., França, R.P., Arthur, R.: Toxoplasmosis Gondii: from discovery
to advances in image processing. In Brazilian Technology Symposium, pp. 91–101. Springer,
Cham (2018)
26. Monteiro, A.C.B., Iano, Y., França, R.P., Arthur, R.: Methodology of high accuracy, sensitivity
and specificity in the counts of erythrocytes and leukocytes in blood smear images. In: Brazilian
Technology Symposium, pp. 79–90. Springer, Cham (2018)
27. Padilha, R., Iano, Y., Monteiro, A.C.B., Arthur, R., Estrela, V.V.: Betterment proposal to
multipath fading channels potential to MIMO systems. In: Brazilian Technology Symposium,
pp. 115–130. Springer, Cham (2018)
28. França, R.P., Peluso, M., Monteiro, A.C.B., Iano, Y., Arthur, R., Estrela, V.V.: Development
of a kernel: a deeper look at the architecture of an operating system. In: Brazilian Technology
Symposium, pp. 103–114. Springer, Cham (2018)
A New Approach for ECG Recording
in Rats: An Autonomic Nervous System
Analysis

Raphael Santos do Nascimento , Fernando da Silva Fiorin ,

Adair Roberto Soares Santos , Luiz Fernando Freire Royes ,
and Jefferson Luiz Brum Marques

Abstract Reliable electrocardiogram (ECG) recording systems in rats become tools

remarkably crucial for the drug safety evaluation in experimental physiological
methods. However, the currently existing techniques for ECG measurement in rats
still impair a better translational that occurs in the autonomic nervous system (ANS)
models. For this reason, we developed a new tool for the ECG measurement in rats
on the awake state, no surgical procedure, restraint, as well as anesthesia admin-
istration and compared to the isoflurane anesthesia administration. The heart rate
variability (HRV) method was used to evaluate the ANS status. The findings demon-
strated clearly that the isoflurane anesthesia influence the activities sympathetic and
parasympathetic of ANS. Therefore, our newly introduced tool showed to be adequate
to study cardiac electrical activity in rats.

Keywords Electrocardiogram · Autonomic nervous system · Isoflurane · Rat

R. S. do Nascimento (B) · J. L. B. Marques

Institute of Biomedical Engineering, Federal University of Santa Catarina, Florianopolis, SC,
Brazil
e-mail: [email protected]
J. L. B. Marques
e-mail: [email protected]
F. da Silva Fiorin · A. R. S. Santos
Laboratory of Neurobiology of Pain and Inflammation, Federal University of Santa Catarina,
Florianopolis, SC, Brazil
e-mail: [email protected]
A. R. S. Santos
e-mail: [email protected]
L. F. F. Royes
Laboratory of Exercise Biochemistry, Federal University of Santa Maria, Santa Maria, RS, Brazil
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 91
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_9
92 R. S. do Nascimento et al.

1 Introduction

Electrocardiogram (ECG) is widely used as a study tool for the understanding of

cardiovascular physiology, both clinically and in experimental research animals. In
this context, ECG measurement methods explored in rats have contributed actively
in the last decade to the pharmaceutical industries in regard to the drug safety eval-
uation. For this reason, rat models become remarkably useful for the assessment of
cardiac pathologies [1]. Importantly, the increasing development of the experimental
physiological models using rats is resulting in a more significant demand for less
invasive ECG measurement methods. However, currently, existing techniques for
recording ECG from rats still require the use of surgical procedures [2], anesthesia
administration [3], as well as restraint of animals [4].
In this perspective, recent studies have reported that the use of a telemetry system is
the most efficient standard for physiological measurement in vivo [5]. The telemetry
is a technique in which a transmitter is implanted on small animals, permitting the
recording of the signals automated and wireless, where the data are posteriorly
received by a receiver outside the home cage of the animals. This method allows
the recording to be done in rats over a long period, occurring in a state very close
to the animal’s natural cardiovascular physiology. Nevertheless, despite this method
provides data more optimized and reliable it still possesses many limitations such as
a very high cost of implementation, physical size (it requires a minimum bodyweight
of the animals for implantation), and external power requirements for its operation
[1, 5, 6]. In investigations of the cardiac autonomic nervous system (ANS), these
methods described here may affect the natural physiological features, impairing a
better translational approach. For example, the surgical procedure for telemetric
transmitter implantation in very diseased animals may result in the non-survival
of these animals. From the anesthesia administration standpoint, the use of several
drugs for cardiovascular evaluation is also sometimes questionable by influences in
the divisions sympathetic and parasympathetic of ANS [3, 7].
Indeed, all these conditions presented above reinforce the importance of the devel-
opment of new approaches to the recording of ECG in rats, which leads to better
control of confounders associated with the ANS experimental models. Based on
this, the heart rate variability (HRV) analysis has been suggested for examining
of the sympathetic and parasympathetic activities of ANS. This is attributed as an
acceptable tool for characterizing the influences of cardiac autonomic control on
the sinus node, which describes the fluctuations of the intervals among consecutive
heartbeats (NN intervals). Epidemiological studies have shown that decreased HRV
activity is correlated with a higher risk of mortality and morbidity in cardiac patients
[8].
Therefore, considering the use of HRV parameters analysis to evaluate the ANS
conditions on the natural animal physiology, we developed a new tool for ECG
recording in rats in the awake state, no permit the use of anesthetics, surgical proce-
dures, restraint or any other type of force exerted on the animals. In order to test
A New Approach for ECG Recording in Rats … 93

the efficacy of the system proposed, we also explored in set the use of isoflurane
anesthetics for comparative analysis.

2 Materials and Methods

2.1 Animals

Five male adult Wistar rats 150 days old (400–475 g) were used in this study. Animals
were maintained under controlled laboratory conditions (12 h light-dark cycle, in
constant temperature 22 ± 2 °C, 54% relative humidity), with free access to filtered
water and ad libitum access to a standard laboratory diet. Animals were randomly
housed (maximum 3) in boxes (27 cm × 36 cm × 31 cm), which were covered by
shavings and allocated on ventilated racks. All procedures performed on the animals
followed the Committee on Care and Use of Experimental Animal Resources of the
Federal University of Santa Catarina (protocol number 2712080317).

2.2 Data Acquisition Protocol

To data recording in the awake state, an electrocardiograph was developed a custom-

made jacket to the thoracic circumference of the rats, in which three electrodes
were allocated internally to record ECG signals. This was produced from brim with
elastane and three fabric velcro tapes. For adaptation in the jacket, the animals were
habituated during forty (40) days (60 min/day for each animal) with the jacket in a
glass box (30 cm × 20 cm × 23 cm) free to move. The animals were induced to enter
inside of the jacket using a feed pellet and with the hands. After the adaptation period,
ECG recording has been conducted in anesthetized and awake conditions. Before the
beginning of assessments, the animals were allowed 30 min dressed in the jacket.
The recording in the anesthetized state was done after the isoflurane administration
(2.5% mixed with 100% of oxygen) in the morning period. After the assessments in
the anesthetized state, the animals were carefully transferred to their home cage to
recovery consciousness (minimum rest of 3 h). In the afternoon period, the animals
were evaluated in the awake state on the new device developed.
94 R. S. do Nascimento et al.

Fig. 1 General block diagram of the ECG measurement system in rats

2.3 Device Architecture

ECG samples were acquired using lead I configuration in the awake and anesthetized
condition. The three electrodes included one for the driven right leg for reducing
noise. The ECGs were recorded during 8–10 min for each animal, from which
was selected a 5-min sample with the lowest noise (e.g., movement artifacts) for
processing. The recorded raw signals were amplified by an instrumentation ampli-
fier (INA129, Texas Instruments Inc., Texas, USA) and subsequently filtered by both
high-pass (first order, 0.1 Hz) and low-pass (Butterworth, fourth-order, 40 Hz) filters.
As this type of signal possesses amplitude in the range millivolts, it was adjusted
by a gain circuit before sending it to the digital-to-analog converter. We also used
a shifter level circuit allowing only the range accepted by the converter (0–5 V).
The signal was digitalized by a 12-bit analog-to-digital converter (ADS1015, Texas
Instruments Inc., Texas, USA) at 500 samples/s and sent through I2 C connection to
a microcontroller (Arduino Nano board V.3.0, San Jose, California, USA). The data
received by microcontroller were transmitted by USB connection to a computer for
storage and off-line HRV analysis. All the devices used in the ECG module were
battery-powered (as illustrated in the block diagram in Fig. 1).

2.4 Heart Rate Variability

To assess the HRV features in rats, the data were extracted by HRVtool application
(HRVtool, Greifswald, MV, Germany, version 1.03) custom-built software using
Matlab (MathWorks Inc., Natick, MA, USA, version 8.5). The data were manipu-
lated following the recommendation of the Task Force of the European Society of
Cardiology and the North American Society of Pacing and Electrophysiology [8].
The extraction of HRV parameters in the time domain was calculated three compo-
nents: mean HR (mean heart rate), SDNN (standard deviation of all NN intervals),
and RMSSD (square root of the sum of the squares of differences among NN inter-
vals). SDNN represents the contributing to all HRV components during the recorded
period, reflecting a measure of overall HRV and RMSSD provides information about
A New Approach for ECG Recording in Rats … 95

the parasympathetic tone in the short term recording. For the analysis in the frequency
domain, data were resampled using cubic spline interpolation to obtain time evenly
spaced. Two frequency bands were established: low frequency in the normalized unit
(LFnu, 0.2–0.75 Hz) and high frequency in the normalized unit (HFnu, 0.75–3.0 Hz)
using fast Fourier transformation (FFT) algorithm [9]. Although there are controver-
sies about the precise reflection of these parameters, both reflect the parasympathetic
(LF) and sympathetic (HF) activities, respectively, and the LF/HF ratio demonstrates
information about the global sympathetic-vagal balance.

2.5 Statistical Analysis

The data analysis was conducted using Paired Students t-tests to evaluate the effects
of isoflurane administration in comparison to the new tool developed in awake rats.
All data were expressed as the means ± standard error of the mean (S.E.M). p <
0.01 and p < 0.05 were considered statistically significant. Statistical analysis was
carried out by the GraphPad Prism (GraphPad Software, Inc., San Diego, CA, U.S.A.,
version 8.0) statistical program.

3 Results

The use of anesthesia administration and ECG recording methods in surgical settings
for evaluation of ANS has been questioned due to the influence on the sympathetic
and parasympathetic activities. In this present study, we tested a new model in the
awake and isoflurane-anesthetized state under the same animals and compared the
results.
HRV parameters varied in both time and frequency domains in the animals in
awake and anesthetized conditions (Fig. 2). The results obtained in the time domain
showed that the heart rate was statistically higher (p < 0.01; Fig. 2a) in the awake
state. On the other hand, the SDNN (p > 0.05; Fig. 2b) and RMSSD (p > 0.05; Fig. 2c)
components did not differ among both the animal states. However, we observed a
reduction in the component SDNN, although not statistically significant. Interest-
ingly, we also noted that the HRV parameters recorded in the anesthetized state
induced a more significant standard error of the mean in comparison to the awake
state, reflecting on all the metrics measured in the time domain. Data analysis of
linear HRV metrics in the frequency domain revealed that there were substantial
differences on all the evaluated parameters LFnu (p < 0.05; Fig. 2d), HFnu (p <
0.05; Fig. 2e) and LF/HF (p < 0.05; Fig. 2f), respectively, in comparison of animals
between both conditions.
96 R. S. do Nascimento et al.

Fig. 2 Heart rate variability metrics. Data are given as means ± standard error of the mean. BPM,
beats per minute; ms, milliseconds; nu, normalized units. Data analysis was conducted using the
student’s paired t-tests. *p < 0.05 and **p < 0.01 were considered statistically significant

4 Discussion

In this present work, we designed a new tool for non-invasive ECG recording in awake
rats, without the use of anesthetics, surgical procedures, restraint, or any other type of
force exerted on the animals, evaluating the possible influences on the sympathetic
and parasympathetic behaviors of ANS. The findings in our study demonstrated
that the proposed system for acquiring ECG signals in rats worked well, providing
HRV values more compatible with the natural physiology of rats. Furthermore, our
data showed that the use of isoflurane administration in experimental physiological
models triggers effects on the ANS activities (sympathetic and parasympathetic).
Due to the experimental physiological models in rats become essential tools to the
drug safety assessment, there is a growing interest in the search for models that may
record the ECG in rats non-invasively. Although the techniques developed until the
present moment have evolved, these still imply effects on the cardiac physiology of
animals. For example, the use of isoflurane anesthesia administration for evaluation
cardiac is widely used on several models, including physiological monitoring to
elicit greater control of prolonging and deepening anesthetic. Nonetheless, our results
demonstrated changes in the HRV indexes in both the time and frequency domain
measured in the awake and anesthetized conditions. In general, the measure of HRV
A New Approach for ECG Recording in Rats … 97

evidenced by HR and SDNN parameters, our findings showed low activity compared
to the anesthetized state, although SDNN was not statistically significant. These
results are consistent with previous studies in rats and mice, which showed low HR
after the isoflurane administration in comparison with the awake state [10, 11].
From a parasympathetic tone standpoint (indexed by RMSSD and HFnu), our
data indicated activation only in the frequency domain by HFnu, suggesting intense
activity on the vagal tone in the anesthetized state. On the other hand, some works
showed low activity after the use of this drug [12, 13]. Such discrepancy may be
explained by the different ECG models in rats used. Regarding the sympathetic tone
(LFnu) and balance sympathetic-vagal (LF/HF), our study showed strong activity
on both LFnu and LF/HF components under isoflurane. Similarly, studies reported
high activity in rats and dogs [12, 13]. Together, these results indicate that although
the global HRV is controlled by sympathetic activity, this may suffer the influence
of parasympathetic activity during isoflurane administration.
In this context, the tool presented here comprised a set of solutions for the conven-
tional methods used, which cause effects on the HRV parameters as well as other
physiological models. Unlike telemetry, our model may be applied to animals of
any size, with low cost for implementation. Although is necessary for an animal
adaptation time in the jacket, this may be performed for any experimental cardiac
model, even on very diseased animals. In this line thinking, this provides values with
smaller experimental errors which may occur by anesthesia administration, surgical
procedure, restraint as well as force exerted on the animals.

5 Conclusion

In summary, the results found herein regarding the HRV parameters in the time and
frequency domain, demonstrated clearly that the isoflurane anesthesia administra-
tion in physiological models interferes with the sympathetic and parasympathetic
activities of ANS. In this sense, our approach developed for ECG recording is more
suitable to explore the cardiac electrical activity features in rats.

Acknowledgements The authors gratefully acknowledge the student fellowships from the National
Council for Scientific and Technological Development (CNPq) (to R.S.N) and the Coordination of
Superior Level Staff Improvement (CAPES) (to F.S.F). A.R.S.S., L.F.F.R, and J.L.B.M are grantees
of CNPq research productivity fellowships.

References

1. Accardi, M.V., et al.: Rat cardiovascular telemetry: Marginal distribution applied to positive
control drugs. J. Pharmacol. Toxicol. Methods 81, 120–127 (2016)
2. Silva, L.E.V., et al.: Comparison between spectral analysis and symbolic dynamics for heart
rate variability analysis in the rat. Sci. Rep. 7, 8428 (2017)
98 R. S. do Nascimento et al.

3. Fenske, S., et al.: Comprehensive multilevel in vivo and in vitro analysis of heart rate fluctuations
in mice by ECG telemetry and electrophysiology. Nat. Protoc. 11, 61–86 (2016)
4. Mongue-Din, H., Salmon, A., Fiszman, M.Y., Fromes, Y.: Non-invasive restrained ECG
recording in conscious small rodents: a new tool for cardiac electrical activity investigation.
Pflugers Arch. 454, 165–171 (2007)
5. Niemeyer, J.E.: Telemetry for small animal physiology. Lab. Anim. 45, 255–257 (2016)
6. Deveney, A., Kjelltrom, A., Forsberg, T., Jackson, D.: A pharmacological validation of
radiotelemetry in conscious, freely moving rats. J. Pharmacol. Toxicol. Methods 40, 71–79
(1998)
7. Hildebrandt, I.J., Su, H., Weber, W.A.: Anesthesia and other considerations for in vivo imaging
of small animals. ILAR J. 49, 17–26 (2008)
8. Task Force of the European Society of Cardiology & The North American Society of
Pacing and Electrophysiology: Heart rate variability: standards of measurement, physiological
interpretation, and clinical use. Eur. Hear. J. 17, 354–381 (1996)
9. Cerutti, C., et al.: Autonomic nervous system and cardiovascular variability in rats: a spectral
analysis approach. Am. J. Physiol. 261, H1292–H1299 (1991)
10. Stein, A.B., et al.: Effects of anesthesia on echocardiographic assessment of left ventricular
structure and function in rats. Basic Res. Cardiol. 102, 28–41 (2007)
11. Pachon, R.E., Scharf, B.A., Vatner, D.E., Vatner, S.F.: Best anesthetics for assessing left ventric-
ular systolic function by echocardiography in mice. Am. J. Physiol. Hear. Circ. Physiol. 308,
H1525–H1529 (2015)
12. Kato, K., Wakai, J., Ozawa, K., Sekiguchi, M., Katahira, K.: Different sensitivity to the suppres-
sive effects of isoflurane anesthesia on cardiorespiratory function in SHR/Izm, WKY/Izm, and
Crl: CD (SD) rats.: Exp. Anim. 65, 393–402 (2016)
13. Kato, M., et al.: Spectral analysis of heart rate variability during isoflurane anesthesia.
Anesthesiology 77, 669–674 (1992)
Algorithm for Detection of Raising
Eyebrows and Jaw Clenching Artifacts
in EEG Signals Using Neurosky
Mindwave Headset

Luis Vélez and Guillermo Kemper

Abstract The present work proposes an algorithm to detect and identify the artifact
signals produced by the concrete gestural actions of jaw clench and eyebrows raising
in the electroencephalography (EEG) signal. Artifacts are signals that manifest in the
EEG signal but do not come from the brain but from other sources such as flickering,
electrical noise, muscle movements, breathing, and heartbeat. The proposed algo-
rithm makes use of concepts and knowledge in the field of signal processing, such as
signal energy, zero crossings, and block processing, to correctly classify the afore-
mentioned artifact signals. The algorithm showed a 90% detection accuracy when
evaluated in independent ten-second registers in which the gestural events of interest
were induced, then the samples were processed, and the detection was performed.
The detection and identification of these devices can be used as commands in a brain–
computer interface (BCI) of various applications, such as games, control systems of
some type of hardware of special benefit for disabled people, such as a chair wheel, a
robot or mechanical arm, a computer pointer control interface, an Internet of things
(IoT) control or some communication system.

Keywords EEG signals · Brain–computer interface · Neurosky mindwave

headset · Artifacts detection

1 Introduction

The electroencephalogram (EEG) signal is originated by neurons activities recorded

as fluctuations in electric differential potentials obtained from several areas and
specific points on the human brain. Commonly, EEG has been used for medical
research and medical applications, but in last years there has been a substantial

L. Vélez · G. Kemper (B)

School of Electronic Engineering, Universidad Peruana de Ciencias Aplicadas, Lima, Peru
e-mail: [email protected]
L. Vélez
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 99
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_10
100 L. Vélez and G. Kemper

perspective to use it for non-medical brain–computer interface applications like

device control, tools for education, and entertainment industry. The raw EEG signal is
usually contaminated by artifacts from different sources such as the heartbeats, eyes
blinks, muscle activities, and noise from electrical equipment [1]. Detection of these
artifacts is often an important subject in EEG research. Commonly, when electrodes
are positioned on central scalp zones containing mainly brain activity, temporopari-
etal sites that may contain muscle artifacts and frontal sites that may carry strong
blinks, eye movement, and other muscle artifacts [2]. Blind source separation (BSS)
methods, like independent component analysis (ICA) and canonical correlation anal-
ysis (CCA), have also been proposed for EMG artifacts detection and removal [3–5].
Such methods, however, require multi-channel data and long data epochs to produce
important results.
The computational complexity is another characteristic that difficults the choice
of ICA or CCA for artifact detection in applications that require real-time implemen-
tations [6].
Artificial intelligence and machine learning methods, such as neural networks and
vector support machines (SVM), are also widely used in the process of extraction
and classification artifacts [7, 8]. Although these methods have been giving very
good results, they require a long training process of the network and a large number
of standard signals, which must be collected from accredited sources, and which
will serve as a reference for their optimal operation, in addition to requiring greater
computational capacity.
The proposed algorithm correctly extracts and classifies two EMG artifacts of
interest (jaw clench and raising eyebrows), with an accuracy of 90%, allowing its
use in real-time processing, due to its light computational load, simplicity of imple-
mentation and its use with EEG sensors of low cost and a single active channel,
promoting the development of multiple applications for brain–computer interfaces
(BCI) such as games, manipulation of the computer pointer for people with disabil-
ities of upper limbs, or communication projects in people with disabilities of this
faculty. The aim is to detect these two particular artifacts generated by two muscular
events: the jaw clenching and eyebrows raising.
The gestures that were made to obtain the artifacts of interest are shown in Figs. 1
and 2, respectively.

2 Description of the Proposed Algorithm

It is important to indicate that the framework used in this study corresponds to the
development of a computational algorithm. Thus, ethical approval for the experiment
in this study was deemed unnecessary by the university, since the research was
designed to develop a computational framework with the help of a commercial non-
invasive sensor that does not represent any health risk. Likewise we consider that the
activities undertaken in the research do not pose risks greater than those ordinarily
encountered in daily life.
Algorithm for Detection of Raising Eyebrows … 101

Fig. 1 Jaw clench artifact gestural

Fig. 2 Eyebrows raising artifact gestural

Figure 3 shows a block diagram of the proposed algorithm. Details of each

processing stage are described in the following sections.

Fig. 3 Block diagram of the proposed algorithm

102 L. Vélez and G. Kemper

Fig. 4 Position of mindwave mobile headset electrodes on the 10–20 international system of
electrode placement. Fp1–A1 electrodes

2.1 Signal Acquisition

The signals were acquired using an economic BCI mindwave mobile headset from
Neurosky. This device is composed of a headset, an ear clip, and a sensor arm. The
dry electrode is on the sensor arm which is placed in frontage. According to [9] the
10–20 international system of EEG electrode placement on the human head, this
position is around corresponding to the Fp1 position shown in Fig. 4. The reference
and ground electrodes are on the ear clip at A1. The diadem also has a Bluetooth
module that sends the signals to the computer at a sampling frequency of 512 Hz.

2.2 Preprocessing Signal

The aim at this stage was to condition the signals to increase precision in the extraction
of characteristics.
Unwanted noise components (coming mainly from the 60 Hz power grid) are
removed for signal improvement. For this purpose, an infinite impulse response IIR
Butterworth second-order filter with a cutoff frequency of 59 and 61 Hz is used for
work like a 60 Hz band stop notch filter. The use of this filter provided the best
results in terms of lowering noise and unwanted distortion. The original signal and
the filtered signal are showed in Figs. 5 and 6, respectively.

2.2.1 Jaw Clench Detection

For the detection of the first artifact jaw clench, Cooper’s algorithm was used, which
extracts and classifies, combining the concepts of energy and zero crossings in a
single operation. Although this algorithm correctly detects the presence of the arti-
fact of interest, in some cases, it gives false detections, so it is combined with the
zero-crossing algorithm, additionally, to rule out the erroneous detections, which are
fragments confused by their similarity in features with jaw clench.
Algorithm for Detection of Raising Eyebrows … 103

Original Signal
2000

1500

1000

Signal Level 500

-500

-1000

-1500

-2000
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Time

Fig. 5 Original signal

Filtered Signal
2000

1500

1000
Signal Level

500

-500

-1000

-1500

-2000
0 200 400 600 800 1000 1200
Time

Fig. 6 Filtered signal with a 60 Hz notch filter

2.2.2 Cooper’s Algorithm

Step 1. Signal samples are recorded for times of 10 s, and this recording time was
chosen for practicality, considering that the signal sample is not so short so that it
is possible to perform the gestures, and not so extensive, that it produces a very
long processing time. Total samples obtained in 10 s are calculated by applying the
following expression:

N samples = Fs × Trecord
104 L. Vélez and G. Kemper

N samples = 512 Hz × 10 s
N samples = 5120 (1)

Stage 2. The signal is divided into a certain number of blocks, with a certain number
of samples per block. Finally, 15 samples per block were considered, since, and
according to experimental research, it provided a favorable total block density for
Cooper’s histogram analysis. Total epochs are calculated by applying the following
expression:

N samples
N epochs =
samples by Epoch
Samples by epoch = 15
5120
N epochs = ≈ 342 (2)
15

Stage 3. The entire signal is traversed, for each of the 342 blocks, and the Cooper
parameter of each block is calculated, applying the following expression with the 15
sample values of the evaluated block:


L
C= |y[m].|y[m]| − y[m − 1].|y[m − 1]|| (3)
m=0

where y[m] is the actual discrete sample analyzed and y[m − 1] is the precedent
sample, and C is the block Cooper value.
Stage 4. Start detector. A starting threshold range of u i min = 81,100 and u i max
= 90,000 is chosen. This threshold range was used because it was experimentally
possible to minimize the error in the detection of the event of interest and avoid
considering very high values of the Cooper parameter that are reached by noise or
other artifacts but never by the jaw clenching. Cooper value of the current block,
[blockcooper], is in this threshold range, the 15 samples of the block are saved in the
output vector [clench], which will contain the extracted artifact signal, and the final
sample of said block is also saved (15th sample) as the starting sample (“ki”) of the
extracted artifact signal.
Stage 5. End detector. An end threshold value of uf = 5000 is chosen. This threshold
value was used because it was experimentally possible to minimize the error in the
detection of the event of interest considering Cooper’s histogram. It is evaluated if
the Cooper value of the block is less than this threshold uf , if it is lower, it is saved
as the last block and the algorithm is completed. The 15th sample of the final block
is also saved as an end sample of the extracted artifact signal (“k f ”). Figure 7 shows
jaw clench detection using Cooper’s algorithm.
Algorithm for Detection of Raising Eyebrows … 105

Fig. 7 Jaw clench artifact classification and Cooper’s histogram

2.2.3 Zero-Crossing Density Algorithm

Stage 1. The previously extracted fragment is divided into a certain number of blocks,
with a certain number of samples per block. Finally, 15 samples per block were
considered, since, according to experience, it provided a favorable total block density
for the analysis of the calculation of zero crossings.
Stage 2. The entire fragment is traversed, for each of the samples of the signal, and
the sign of the present sample is evaluated with the sign of the previous sample, to
verify if there is a change of sign between two continuous samples. If this condition
exists, it is counted as zero crossing, causing the density to increase by a value of
two. All this is done according to the following expression:

N −1

Z= |sign[y(m)] − sign[y(m − 1)]| (4)
m=0

where y(m) is the actual discrete sample analyzed, y(m − 1) is the precedent sample,
N is the total samples, and Z is the zero-crossing density value.
Stage 3. Classification or discard. A decision threshold value of U z = 150 is chosen.
This threshold value was used because it was experimentally possible to minimize
the error in the detection of the event of interest. If the value of zero crossings of the
current fragment is greater than the “U z ” value, the start samples are saved (“K i ”)
and end (“k f ”) to delimit the artifact, otherwise the fragment is discarded.
106 L. Vélez and G. Kemper

Raising Eyebrows Artifacts Patterns

2500

2000

1500

1000
Signal Level

500

-500

-1000

-1500

-2000

-2500
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Time

Fig. 8 Original “eyebrows raising” artifacts signal recorded with Neurosky mindwave headset

2.2.4 Eyebrows Raising Detection

For the detection of the second “raising eyebrows” artifact, the algorithm of calcula-
tion of energy by blocks of the signal was used, which extracts the possible fragments
that contain the artifact of interest. In the case of this detection, the characteristic is
that it has very high energy values in the peaks, and a minimum value of zero cross-
ings, unlike the first artifact that has high values of zero crossings, and low energy
levels.
Figure 8 shows two original raising eyebrows artifacts registered by Neurosky
mindwave headset in two different times.

2.2.5 Energy Calculation Algorithm Per Signal Blocks

Stage 1. Steps 1 and 2 of the Cooper’s algorithm are performed.

Stage 2. The entire signal is traversed for each of the 342 blocks, and the energy
value of each block is calculated. This calculation is performed through the following
expression:

+N −1
i N
Ei = x 2 (n) (5)
n=i N
Algorithm for Detection of Raising Eyebrows … 107

where x(n) is the discrete sample analyzed, i is the block number, N is the block
total samples, and E i is the energy of the block.
Stage 3. Start detector. A start threshold value of ui = 1.5e8 was chosen. This
threshold value was finally used because it was experimentally possible to mini-
mize the error in the detection of the event of interest after 10 variations, and due to
the high energy value of the present artifact in the peaks. If the energy of the current
block exceeds this starting threshold, the 15 samples of the block are stored in the
output vector [eyebrow], which will contain the possible extracted artifact signal,
in addition, the final sample of said block is saved (15th sample) as a start sample
of the extracted artifact signal. The peak width counter is increased for subsequent
evaluation that will serve to validate the extracted segment.
Stage 4. End detector. An end threshold value of uf = 0.3e8 was chosen. This
threshold value was finally used because it was experimentally possible to minimize
the error in the detection of the event of interest after 10 variations. If the energy of
the block is less than this end threshold, it is saved as the last block of the extracted
signal.
Stage 5. The peak width is evaluated if it is less than a threshold width of 150.
This value was considered because the minimum peak width for a rapid brow lift is
approximately 160 samples. If it is smaller the extracted signal is discarded, otherwise
the algorithm culminates. The 15th sample of the final block is also saved as an end
sample of the extracted artifact signal if the discard is not performed. Figure 9 shows
eyebrows raising artifact detection.

Raising EyeBrows Artifact Detection

2000

1500

1000

500
Signal Level

-500

-1000

-1500

-2000

1000 2000 3000 4000 5000 6000 7000 8000 9000 1000011000
Time

Fig. 9 “Eyebrows raising” artifacts detection

108 L. Vélez and G. Kemper

3 Results

To evaluate the proposed algorithms, both the eyebrow lift detection and mandibular
pressure detection, independent tests were performed to validate the accuracy in
detecting these events. Placing and performing the Bluetooth pairing of the Neurosky
sensor with the desktop computer and the recording application, made in MATLAB,
we made recordings of 10 s, time that was considered adequate to be able to perform
one of the gestures with tranquility. After each recording, the recorded signal was
analyzed to detect the gesture made. With the results obtained, Table 1 was carried
out, and the percentage of error in the detection of each algorithm corresponding to
each gestural event was calculated (Figs. 10 and 11).
The goal of this recordings was a concept and preliminary study for later appli-
cations, where tests with individuals will be regulated by the respective ethical
committee.

Table 1 Results of the algorithms precision detection

Test number Jaw clench detection result Raise eyebrows detection result
1 Right Right
2 Right Wrong
3 Right Right
4 Right Right
5 Right Right
6 Right Right
7 Right Right
8 Right Right
9 Right Right
10 Right Right
11 Right Right
12 Right Right
13 Right Right
14 Right Right
15 Right Right
16 Right Wrong
17 Right Right
18 Wrong Right
19 Right Right
20 Right Right
Algorithm for Detection of Raising Eyebrows … 109

Fig. 10 Detection right rate in eyebrows raising artifact

Fig. 11 Detection right rate in jaw clench artifact

4 Conclusion

In conclusion, the present work provides a possibility to take advantage of jaw clench
and raising eyebrows artifacts, to be used as commands in any system that can be
controlled, using simple digital signal processing (DSP) algorithms and of easy
implementation. As a factor to take into account, it is that conductive gel should be
used in the electrode, since it is very helpful in the noise filter of different frequencies
that are introduced in the electrode, and that cannot be suppressed by digital filters.
As future work, these artifacts are planned to be used in a brain–interface system
to control the computer mouse pointer, very useful for people with disabilities in the
movement of their hands and arms to control it.
110 L. Vélez and G. Kemper

References

1. Matiko, J.W., Beeby, S.P., Tudor, J.: Real time eye blink noise removal from EEG signals using
morphological component analysis. In: 2013 35th Annual International Conference of the IEEE
Engineering in Medicine and Biology Society (EMBC), pp. 13–16. https://doi.org/10.1109/
embc.2013.6609425 (2013)
2. Delorme, A., Sejnowski, T.J., Makeig, S.: Enhanced detection of artifacts in EEG data using
higher-order statistics and independent component analysis. NeuroImage 34, 1443–1449 (2007).
https://doi.org/10.1016/j.neuroimage.2006.11.004
3. Jafari, A., Gandhi, S., Konuru, S.H., Hairston, W.D., Oates, T., Mohsenin, T.: An EEG artifact
identification embedded system using ICA and multi-instance learning. In: 2017 IEEE Inter-
national Symposium on Circuits and Systems (ISCAS), pp. 1–4. https://doi.org/10.1109/iscas.
2017.8050346 (2017)
4. Loring Z., Dornhege, J.R., Millan, T., Hinterberger, D.J., McFarland, K.M. (eds).: Toward Brain-
Computer Interfacing. The MIT Press, London (2007)
5. Xia Erkens, I.J., Molina, G.G.: Artifact detection and correction in neurofeedback and BCI
applications (2008)
6. Islam, M.K., Rastegarnia, A., Yang, Z.D.: Methods for artifact detection and removal from scalp
EEG: a review. Neurophysiologie Clinique/Clin. Neurophys. 46, 287–305. https://doi.org/10.
1016/j.neucli.2016.07.002 (2016)
7. Chambayil, B., Singla, R., Jha, R.: EEG eye blink classification using neural network. In:
Proceedings of the World Congress on Engineering, vol. 1, pp. 2–5 (2010)
8. Kousarrizi, M.R.N., Ghanbari, A.A., Teshnehlab, M., Shorehdeli, M.A., Gharaviri, A.: Feature
extraction and classification of EEG signals using wavelet transform, SVM and artificial neural
networks for brain computer interfaces. In: 2009 International Joint Conference on Bioinfor-
matics, Systems Biology and Intelligent Computing, pp. 352–355. IEEE (2009). https://doi.org/
10.1109/ijcbs.2009.100
9. Sanei S., Chambers J.: EEG Signal Processing. Wiley (2007)
Correspondence Between TOVA Test
Results and Characteristics of EEG
Signals Acquired Through the Muse
Sensor in Positions AF7–AF8
Ober Castillo , Simy Sotomayor , Guillermo Kemper ,
and Vincent Clement

Abstract This paper seeks to study the correspondence between the results of the
test of variable of attention (TOVA) and the signals acquired by the Muse electroen-
cephalogram (EEG) in the positions AF7 and AF8 of the cerebral cortex. There are a
variety of research papers that estimates an index of attention in which the different
characteristics in discrete signals of the brain activity were used. However, many of
these results were obtained without contrasting them with standardized tests. Due to
this fact, in the present work, the results will be compared with the score of the TOVA,
which aims to identify an attention disorder in a person. The indicators obtained from
the test are the response time variability, the average response time, and the d  prime
score. During the test, the characteristics of the EEG signals in the alpha, beta, theta,
and gamma subbands such as the energy, average power, and standard deviation
were extracted. For this purpose, the acquired signals are filtered to reduce the effect
of the movement of the muscles near the cerebral cortex and then went through a
subband decomposition process by applying transformed wavelet packets. The results
show a well-marked correspondence between the parameters of the EEG signal of
the indicated subbands and the visual attention indicators provided by TOVA. This
correspondence was measured through Pearson’s correlation coefficient which had
an average result of 0.8.

O. Castillo · S. Sotomayor · G. Kemper (B)

Universidad Peruana de Ciencias Aplicadas, Av. Prolongación Primavera 2390, Santiago de
Surco, Lima, Peru
e-mail: [email protected]
O. Castillo
e-mail: [email protected]
S. Sotomayor
e-mail: [email protected]
V. Clement
Martin-Luther University Halle-Wittenberg, Univesrtitätsplatz 010, 06108 Halle (Saale), Germany
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 111
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_11
112 O. Castillo et al.

Keywords Neurofeedback · Mindfulness · Visual attention · Muse · EEG signals ·

Signal processing · TOVA test

1 Introduction

Problems of attention and concentration are the most common in the learning
processes of children and adolescents. Fuster conducted a review of the different
definitions of attention in which he concludes that it is a mechanism directly related
to the activation and selection of cognitive processes. Attention has been divided
into different types among which sustained attention is found. Sustained attention is
defined as the ability to maintain the focus of attention on a stimulus or activity for
a long time [1].
In this context, a tool developed to support the diagnosis of attention deficit hyper-
activity disorder (ADHD) is the test of variables of attention—TOVA® . This test
consists of the presentation of two types of visual stimulation on a conventional
screen under certain technical specifications. These stimuli are divided into two, the
target stimuli and the nontarget stimuli. Figure 1 shows the TOVA target stimuli on
the left and the TOVA nontarget stimuli on the right. During the TOVA test, new
stimuli are presented every 2000 ms and are flashed for 100 ms. At the end of the
test, the results shown are the average response time of the correct answers (RT),
the variability of the correct response times (RTV), and the indicator that relates the
committed and the omitted errors (D ) [2].
On the other hand, studies on the monitoring of brain activity have associated
selective attention to the manifestations of brain activity in the gamma subband. [3].
Other studies provide strong results regarding the relationship of the alpha subband
with changes in the attention stimuli [4]. However, very few correspond their results
in standardized psychological tests.

Fig. 1 TOVA stimuli: target (left) and nontarget (right)

Correspondence Between TOVA Test Results … 113

Therefore, this research aims to find a correspondence between the electrical

activity of the brain and the results of the standardized TOVA attention variables
test. Research results reveal a correspondence according to Pearson’s correlation
coefficient of 0.84 at electrode AF7 and 0.76 at electrode AF8 between the band
Bβ−H standard deviation and the RT indicator of TOVA.

2 Description of the Proposed Algorithm

Figure 2 shows a block diagram of the proposed algorithm. Details of each processing
stage are described in the following sections.
Stage 1. A database was built from 10 sessions of the test TOVA.
Stage 2. The TOVA test was performed with one change in the conditions mentioned
above. Behind the monitor where the TOVA test was shown, a movie was projected
for distracting stimulus. Seven tests were performed.
These two different conditions in the performance of the TOVA test allowed us
to find the relationship between certain physiological changes in brain activity and
the parameters of the TOVA test.

2.1 EEG Signal

The signal used in this investigation was provided by the electroencephalogram

device Muse version 2016, developed by Interaxon Inc. This device has four elec-
trodes in the positions TP9, AF7, AF8, and TP10, according to the international
system for the electrode placement 10/10 [5, 6]. The device connects to the PC via
Bluetooth 4.0 and delivers 12 bits per sample at a sampling rate of 256 Hz. The
brain’s electrical activity is classified in different frequency bands. These frequency
signals are encountered from 0.1 Hz up to roughly 100 Hz. According to this, the
EEG device needs to enable at least twice the maximum frequency to be analyzed
(100 Hz) to match the sampling theorem [7].

Fig. 2 Block diagram of the proposed algorithm

114 O. Castillo et al.

2.2 Data Acquisition

The signal was acquired for 21.6 min, the duration of the TOVA test. Due to the
sampling frequency (256 Hz), 331,776 samples of raw data were obtained for each
session and for each electrode. Studies of ADHD reveal that in most cases the char-
acteristics of attention disorders are the product of damage in the prefrontal cortex.
This is why the two front electrodes of the Muse (AF7 y AF8) were used [1].
On the other hand, both for the generation of the database (Stage 1) and for
the subsequent tests (Stage 2), suitable space for the realization of these tests was
prepared and a TOVA test was coded following the specifications that are available in
the test manual. Some of these considerations are the size of the stimulus, that should
be found in between 15 and 30% of the monitor’s size (this is measured diagonally),
the displayed stimuli, that should be placed at eye level to the subject, and an eye to
monitor the distance of 75 cm [8].

2.3 Subband Decomposition Filtering and Reconstruction

In this stage, the acquisition signal of the EEG is separated into four main subbands
that were needed in the identification of a correspondence between the TOVA and
the EEG signals. These were obtained from the decomposition of the signal in 18
subbands 4 Hz wide as shown in Fig. 3 [9]. For this, a subband decomposition algo-
rithm was developed with the objective of eliminating the bands where the artifacts
signals were located and then proceed to the reconstruction of the bands of interest:
Bβ−L , Bβ−M , Bβ−H , Bγ−L , Bγ−M y Bγ−H whose description is in Table 1.
Due to the instability of the EEG signals, and their non-stationary characteris-
tics, it is recommended to use the wavelet packet transform. A wavelet transform
is a mathematical tool that allows the extraction of information both in time and
frequency. The families of this type of filters that are used for decomposition come
from a single function known as “mother wavelet” and its use has been shown to
effectively remove the different noises on non-stationary signals [10]. For this inves-
tigation, the Daubechies 9 (“db9”) family was used, which shows a great correlation
with the EEG signal according to Akkar and Hassim, 2017 [11].
To reduce the influence of blinks on the acquisition signal, the averaging filter,
the median filter, and the elimination of the band containing more energy during the
blink were tested. It was noted that the best method to eliminate intrusive signals in
brain activity is to eliminate the band that contains the most components of the signal
to be removed. The results of the filters are observed below. Figure 4a shows that
the average and median filter affect the entire signal; and, although they reduce the
peaks produced by the blinking, much information is lost from the part not affected by
these blinks, unlike Fig. 4b, where greater reduction is observed during the blinking
and conservation of the signal not affected by them. The bands where the highest
energy were produced by the artifacts was removed from the reconstruction process,
as shown in Fig. 4; these bands correspond to a bandwidth of de 0–12 Hz, and these
are components of the alpha and delta band of brain activity that, for purposes of this
paper, they are not of interest in the search of the state of attention.
Correspondence Between TOVA Test Results … 115

Fig. 3 Wavelet packet decomposition tree used in the proposed method

Table 1 Frequency bands association

Subband Hz i Subband Hz i
Bβ−L 12–20 1 Bγ−M 44–56 6
Bβ−M 20–24 2 Bγ−H 56–72 7
Bβ−H 24–32 3 Bβ 12–32 8
Bγ−L 32–44 4 Bγ 32–56 9

2.4 Signal Characteristics Extraction

In this stage, the signal characteristics of each subband indicated in Table 1 were
extracted. The characteristics extracted were:
Temporary mean power (P). To calculate, the vector x i was taken, where i describes
the corresponding subband and N i the number of elements contained in the vector of
the subband i; the absolute values added to the square of the N i elements contained
in the vector were summed to finally divide it by the size of the vector.
116 O. Castillo et al.

Fig. 4 a Original signal compared to the median and mean filtered signal. b Original signal
compared to the original signal without Bα

1  iN
Pi = × |xi (n)|2 (1)
Ni n=1

Temporary standard deviation (σ ). To calculate, the sum of the difference between

each element and its mean μi squared was made. Finally, it was divided by the number
of elements contained in xi .

Ni
n=1 (x i (n) − μi )2
σi = (2)
Ni
Correspondence Between TOVA Test Results … 117

Table 2 Mean TOVA results summary

Optimal TOVA test Non-optimal TOVA test
RTV 63.60 104.74
RT 344.13 426.18
d 4.55 4.07

The exposed characteristics were applied to all the rebuilt subbands during the
performance of the TOVA both for the obtaining of the database (Stage 1) and for
the tests TOVA under distracting conditions (Stage 2).
From these results, it was possible to identify the bands in which there was a
greater change between the two stages. Within this analysis, the band Bβ−H was
chosen and it was also decided to integrate a new band (BΣ ) defined in Eq. 3.

BΣ = Bβ−L + Bβ−M + Bβ−H + Bγ−L + Bγ−M (3)

2.5 Comparison

To compare the results of the EEG, the following TOVA parameters were extracted:
response time (RT), this parameter reflects the average time required for a correct
response to a target stimulus. Response time variability (RTV) corresponds to the
variability of the correct response times. D Prime score (d  ) indicates the accuracy to
discriminate a target stimulus and a nontarget stimulus. This score can be interpreted
as “perceptual sensitivity” [12]. Table 2 shows a summary of the results obtained
from the TOVA test. The second column corresponds to the average of the tests
performed under optimal conditions and the third column, to the average of the tests
performed under distracting stimuli.
Each of these parameters was contrasted with the characteristics of the bands
selected in the previous point to find some correspondence between both groups.

3 Results

Finally, Pearson’s correlation coefficient of the TOVA test with the selected EEG
signals was calculated. Ten sessions of Stage 1 and seven sessions of Stage 2 were
performed. From there, we needed to remove three sessions on electrode AF7 and one
session on electrode AF8 due to an accumulation of values that were far above from
the mean. A correlation coefficient was observed between the temporary standard
deviation of the Bβ−H band with the TOVA RT of −0.8824 for the AF7 electrode and
−0.8071 para the AF8 electrode, as shown in Fig. 5a.
118 O. Castillo et al.

Fig. 5 a Correlation between Bβ−H and TOVA response time. b Correlation between BΣ and
TOVA response time

A correlation coefficient between the temporary mean power of the band BΣ and
the TOVA RT of −0.7817 for electrode AF7 and −0.5154 for electrode AF8 is also
observed in Fig. 5b.
The x-axis shows the TOVA response time in which a higher value corresponds
to the slower response toward a stimulus, indicating a lower attention state [12]. The
y-axis shows the value of the characteristics extracted from the band obtained in the
TOVA tests performed, being the temporary standard deviation in Fig. 5a and the
temporary mean power in Fig. 5b. It was found that tests performed in Stage 1, with
optimal conditions, tend to have an EEG response of high mean power and standard
deviation, and tests performed in Stage 2, with the distracting stimulus, tend to have
lower mean power and lower standard deviation.

4 Discussion

The existence of a relationship is between the behavior of the physiological activity

of the brain and the TOVA results as evidenced by the good correlation between the
average response time of the TOVA and the standard deviation of the band Bβ−H ,
and this relationship shows that greater response times are manifesting, in inverse
proportion, into greater standard deviation of the power in the band Bβ−H . However,
it is important to continue taking data to reinforce this correlation and find other
parameters that could also meet this relationship.
For the continuation of this investigation, an analysis will be carried out in smaller
window times in the EEG signal to observe in more detail the behavior of the signals
throughout the test and, in this way, determine the existence of a greater correlation
between the EEG signals and other TOVA parameters.
Furthermore, the AF8 electrode signal shows more reliability in seeking the corre-
spondence between the parameters specified above. This asymmetrical performance
between AF7 and AF8 electrodes may be due to not so accurate EEG positioning or
degraded electrode. Due to this behavior, AF7 electrode presented samples that were
Correspondence Between TOVA Test Results … 119

far above from the mean in three sessions, so they were removed from the analysis:
two correspond to a stage 1 session and one to a stage 2 session.
This research allowed us to identify the brain activity band where sustained atten-
tion manifests in relation to standardized TOVA test. In the same way, we can build
now a characteristic vector in order to feed a neural network for classifying the atten-
tion state. The final goal of this research is to develop a neurofeedback therapy device
of the attention state.
It is important to indicate that the framework used in this study corresponds to
the unique case approach because we attempted to determine a correlated behavior
between two parameters. Thus, ethical approval for the experiment in this study was
deemed unnecessary by the university, since the research was designed to develop
a computational framework. For this, all the data presented were acquired from the
first author of this research and did not involve human participation. Additionally, for
this study, we use a commercial EEG non-invasive sensor with dry surface electrodes
that do not represent any health risk. The activities undertaken in this research do not
pose risks greater than those ordinarily encountered in daily life, as well as affecting
the preservation of the anonymity.
The goal of these recordings was a concept and preliminary study for later appli-
cations, where tests with several persons will be regulated by the respective ethical
committee.

5 Conclusions

The research carried out shows that sustained attention is manifested in the temporary
standard deviation of the Bβ−H band. This relationship has a behavior inversely
proportional to the response time of the TOVA. A lower level of sustained attention
corresponds to a longer average response time to the objective stimulation of the
TOVA test.
With the results obtained, a behavior of brain activity has been identified during the
state of sustained attention, so a K-means classifier will be implemented to develop
a neurofeedback therapy that will be used as a tool in psychology studies.

References

1. Fuster, M.: Human Neuropsychology. The Prefrontal Cortex, pp. 183–235 (2015)
2. Robert, L., Tammy, D., Lawrence, G., Carol, K., Steve, H.: TOVA Variables and Scoring,
pp. 4–7. T.O.V.A. Professional Manual (2018)
3. Müller, Matthias, M., Thomas, G, Andreas, K.: Modulation of induced gamma band activity in
the human EEG by attention and visual information processing. Int. J. Psychophysiol. 38(3),
283–299 (2000)
120 O. Castillo et al.

4. Sauseng, P., Klimesch, W., Stadler, W., Schabus, M., Doppelmayr, M., Hanslmayr, S., Gruber,
W.R., Birbaumer, N.: A shift of visual spatial attention is selectively associated with human
EEG alpha activity. Eur. J. Neurosci. 22(11), 2917–2926 (2005)
5. Trans Cranial Technologies. 10/20 System Positioning Manual (2012)
6. Mike, S., Mary Ann, W., Daniel, D.: Epilepsy and Seizures. The Little Black Book of
Neuropsychology: A Sindrome-Based Approach. Chapter 16, p. 429 (2011)
7. Luc, L.: Nyquist sampling theorem: understanding the illusion of a spinning wheel captured
with a video camera, pp. 49(6). Physics Department, Royal Military College of Canada, Physics
Education„ 697. (2014)
8. Robert, L., Tammy, D., Lawrence, G., Carol, K., Steve, H.: The T.O.V.A. Test, pp. 2–4. T.O.V.A.
Professional Manual (2018)
9. Kemper, G., Ponce, D., Telles, J., Carpio, C.: An algorithm to obtain boat engine RPM from
passive sonar signals based on DEMON processing and wavelets packets transform. J. Electr.
Eng. Technol (2019)
10. Krishnaveni, V., Jayaraman, S., Aravind, S., Hariharasudhan, V., Ramadoss, K.: Automatic
identification and removal ocular artifacts from EEG using wavelet transform. Measur. Sci.
Rev. 6, Section 2, Nr. 4 (2006)
11. Hanan, A., Faris, J.: Optimal mother wavelet function for EEG signal analyze based on packet
wavelet transform. Int. J. Sci. Eng. Res. 8(2) (2017)
12. Robert, L., Tammy, D., Lawrence, G., Carol, K., Steve, H.: TOVA Variables and Scoring.
T.O.V.A. Professional Manual, pp. 4–5 (2018)
Mobile Robot to Assist in Therapies
for Children with Autism

Madai Keila Taype Mateo and Grimaldo Wilfredo Quispe Santivañez

Abstract The increase in the autistic population and rapid technological advances
have led to further research on robot-assisted rehabilitation for autism therapy as a
practical system aimed at overcoming the distance between autistic patients and their
therapists. In this investigation, an interactive mobile robot was developed whose
main objective is to perform therapy for children with autism spectrum disorder
(ASD) that meets certain qualities that will be chosen by the therapist, so that the
robot will be able to give treatment in different approaches of language, speech, and
coordination. This mobile robot offers the ease of being manipulated by a geomag-
netic router (GRG) glove, RFID cards that allow the display of different expressions
(emotions), and SD cards that can save voice recordings or songs according to the
therapist.

Keywords Autism · Interactive mobile robot · GRG · RFID

1 Introduction

Currently, interactive mobile robots are still alien to the Peruvian market because
the delay and lack of importance of new implementations for the treatment of chil-
dren with autism spectrum disorder (ASD) in Peru, specifically in Huancayo, has
prevented our country from implementing modern and updated technology in this
field. Autism represents a challenge for medicine. It is an early onset of neuropsy-
chological development disorder. It is one of the most serious alterations that affect
behavior and communication with stereotyped patterns. Autism spectrum disorder
(ASD) is a neurological and developmental condition that begins in childhood and

M. K. Taype Mateo
Facultad de Ingenieria, Universidad Continental, Huancayo 15036, Peru
e-mail: [email protected]
G. W. Quispe Santivañez (B)
Universidad Nacional Autonoma Altoandina de Tarma, Tarma 12651, Peru
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 121
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_12
122 M. K. Taype Mateo and G. W. Quispe Santivañez

lasts a lifetime. People with ASD present problems that concern difficulties to talk
and a scattered vision when communicating with other people [1].
According to [2], the percentage of minors aged 3–11 years with a mental
disability, specifically autism, is 0.076% of the current population of Peru. In the
Junín region, out of 60 people diagnosed with ASD, 29 are minors aged 3–11 years
[3]. This is increasing owing to various circumstances. It is worth noting that within
this percentage, the majority of people are of low socioeconomic status, and conse-
quently the various therapies that can be applied are ignored as they can not afford
it. Although children with ASD need adequate care and education according to their
condition, there has been no progress in projects that provide adequate treatment
within the educational field. This research work aims to facilitate progress in the
scientific, technological, and educational field with respect to autism therapy in the
Junín region, specifically in Huancayo.
Robots can cooperate with humans to perform tasks. The image from science
fiction that comes to mind, of humanoid robots or service robots working alongside
humans, still invokes a certain sense of surprise and insecurity when performing a
joint task. This perception is perhaps influenced by the image of the industrial robot
working in manufacturing scenarios. The case of robots for medical applications
is especially interesting owing to the aspect of human relations in health issues [4].
Robotics brings together various disciplines such as mechanics, electronics, computer
science, artificial intelligence, control engineering, and physics. Other important
areas are algebra, programmable automatons, animatronics, and state machines [5].
Medical robots are robots that operate in any of the following scenarios: a surgical
environment, rehabilitation scenario, or domestic environment where assistance tasks
are performed. The common characteristic that we find in these robots is that apart
from collaborating to perform a certain task, these medical robots have a direct
physical relationship with human beings, i.e., there is physical contact between the
robot and a person. Obviously, this feature conditions the human-machine interface
of these robots and makes them different from industrial robots [4].
In recent years, the psychological approach and engineering and robotics advances
have been combined to offer an alternative in the treatment of children and adults with
ASD, robot-assisted therapy [6]. In this sense, there are therapies given to children
with autism from behavioral analysis when they interact with different toys [7]. Also,
through therapies driven by mistrust and deception games, children with ASD learned
complex social rules from a social robot [8]. The empirical evidence supports the
potential use of specific robotic and technological agents for increasing motivation
and socially relevant behaviors in some children with ASD [9]. The development
of this type of robot-assisted therapies aims at improving basic services to children
with ASD from the moment they are diagnosed and throughout their life cycle, with
support services for families, allowing parents to live with the affected child as long as
possible, without dismantling the family nucleus and thus achieving true integration
into society [10].
This work pretends to explore to what extent does the prototype development of
an interactive mobile robot for the treatment of children with moderate ASD at the
school stage may contribute. The main objective of the research is to design a mobile
Mobile Robot to Assist in Therapies for Children … 123

robot that will help therapists in their sessions, specifically with speech, language,
and music.

2 Function

The proposed prototype will fulfill the function of being the therapist’s assistant
because therapists will make improvements in the therapies for language, speech,
and music [11].
Next, the therapies where the robot will assist are explained:
• Speech: The therapist is available for therapy with RFID cards because thanks
to this the robot can change expressions in six opportunities. With the help of
speech, the therapist can interact with the patient when they identify themselves
and use expressions.
• Language: Here we propose a therapy that is simple for the patient because the
robot has an SD card in which words selected by the therapist can be recorded to
interact with the patient.
• Music: In this therapy, the robot has an SD card, where music selected by the
therapist can be recorded.
• Coodination: It has a geo-drive control system in the form of geomagnetic router
gloves (GRG) that can direct the robot to a specific point.

3 Materials and Methods

This project involves systematic creation and development, and the documentation
of the design activity can be used to manufacture the product. In terms of existing
work, the method standardized by the German association of engineers VDI (Verein
Deutscher Ingenieur), VDI2222, was developed [12]. This method can be easily used
by designers with or without experience, in addition to seeking optimization in each
of its stages, so that the solutions always aim to be the best. Next, a list of research
requirements is prepared where the design is analyzed, which will make it easy to
interact with the children with the objective of capturing the attention of children
with ASD, so that they can learn to socialize through the SD cards (provided in
the prototype) in which voice or songs will be transmitted according to the desired
treatment. A child will be able to see the changes in the expressions of the robot
using the RFID cards.
124 M. K. Taype Mateo and G. W. Quispe Santivañez

Fig. 1 Robot assembly–inventor

3.1 Design

It was achieved by separating the physical from the intangible, it was developed
through steps, through which the designs of both software and hardware were
obtained to finally assemble them together. The development was carried out using
Inventor [13] and Proteus software [14] (Figs. 1, 2, and 3).

3.2 Materials

The materials used in this work were Arduino mega, accelerometer, L293D Driver,
Bluetooth, RFID card reader, SD card reader, speakers, 8 × 8 matrixes, and MDF
frame.

3.3 Software Development

It was based on being able to control the robot by two methods: by an application
created for an application, by an accelerometer as well as the RFID cards or by
algorithms with the programming of microcontrollers.
Mobile Robot to Assist in Therapies for Children … 125

Fig. 2 A complete circuit in proteus

Fig. 3 Geomagnetic router glove (GRG)

126 M. K. Taype Mateo and G. W. Quispe Santivañez

Fig. 4 Laser cutting of the robot design

3.4 Hardware Development

Hardware was manufactured using easily accessible materials and efficient manip-
ulation. The operation was executed with the AutoCAD program and laser cutting
(Fig. 4).

3.5 Flowchart

This part explains the robot process and its working (Fig. 5).

4 Results

From the fabrication and testing process, an interactive mobile robot with two degrees
of freedom was obtained with dimensions of 23 cm × 23 cm and a height of 48 cm.
The total weight of the robot was 2 kg, and the rotation range of the head was from 9° to
180°. The human-machine interface was developed using open-source software and
hardware technology that allows cost reduction in the manufacturing of therapeutic
robots without degrading their quality. In this case, the software system was built in
Arduino. The hardware was implemented with technological components that were
compatible with everything that was used. The performance of the interactive robot
was evaluated at each stage of the child’s cognitive development. The obtained data
show that the signals coming from the controller and the commands given by the
Arduino do not have communication delay because they are working at the speed
of milliseconds. To improve the robot–patient interaction, audio (voice) and music
player were incorporated into the robot to listen to and observe the patient’s reactions
Mobile Robot to Assist in Therapies for Children … 127

Fig. 5 Flowchart of robot operation

in real-time. It also had an accelerometer installed, which was part of the robot and
was placed in the hand of the patient or therapist to observe the robot movement and
the obtained interaction. In addition, there was an option of the specialist starting a
conversation with the patient to evaluate the degree of attention and understanding
that the patient had through the robot. In conclusion, this development has allowed
the generation of knowledge about therapeutic social robotics. It has facilitated a
new field of application of interactive robots. In this way, the research in this area
can take advantage of it and served to implement new ideas in future projects where
may be implemented.

5 Development Cost

It is known that the robot mentioned as the background is in charge of examining

children with and without ASD who learned to socialize with a social robot through
128 M. K. Taype Mateo and G. W. Quispe Santivañez

games of mistrust and deception. The cost of this robot for academic use is EUR
5.79,000; within Peru, the cost is PEN 21.83,971. Therefore, it is very expensive,
especially in Huancayo. Thus, the robot proposed in this research, which works as
a therapist assistant with the objective of treating children with ASD, offers ease of
acquisition, as its components are very easy to find in the Peruvian market. Moreover,
its maintenance can easily be understood by users, and its cost for academic use in
Peru is PEN 2000 (Tables 1, 2, and 3).

Table 1 Cost of materials

Item Description Quantity Unit Unit cost Total cost
1 Accelerometer 1 PEN 15.00 15.00
2 Driver module 1 PEN 20.00 20.00
3 Arduino mega 1 PEN 60.00 60.00
4 Lithium battery 8 PEN 5.00 40.00
5 RFID card 6 PEN 5.00 30.00
6 Servomotor 1 PEN 25.00 25.00
7 RFID reader 1 PEN 15.00 15.00
8 SD card 1 PEN 4.00 4.00
9 Matrix 8 * 8 3 PEN 16.00 48.00
10 Gearmotor 4 PEN 6.00 24.00
Subtotal PEN 281.00

Table 2 Mechanical part cost

Item Description Quantity Unit Unit cost Total cost
1 Chassis 1 Und 30 30
2 Bolts 50 Und 15 15
3 Nuts 50 Und 15 15
4 Brackets 4 Und 5 20
Subtotal PEN 80.00

Table 3 Total cost

Item Description Quantity Unit Subtotal
1 Purchase of electronic part materials 1 Todo 281
2 Purchase of mechanical part materials 1 Todo 80
3 Services 1 Todo 700
Total PEN 1061.00
Mobile Robot to Assist in Therapies for Children … 129

6 Conclusion

In this study, we demonstrate the feasibility of manufacturing a mobile robot that

makes facial changes, music, and simple and interactive movement. Contributing
mainly to therapies as an assistant in the process of improvement in the patient. It
exhibits movements that can attract the attention of children being able to move easily
and without irregularities. The movement of the head gives realism to its expressions
with the design and construction of the interactive mobile robot for the treatment
of children with ASD were performed with two degrees of freedom where you can
easily interact with the user through the glove where you place the accelerometer,
and you can observe the movement of the robot, it is also able to interact with the
user by the voice recorded on the SD card, responding with sounds and gestures
images by the RFID card. Therefore, the interactive mobile robot is suitable for such
treatments.
As future work, we will improve the design of the mechanism to enlarge the arms
of the prototype and, also accompany and measure qualitatively the improvements
in therapy.

References

1. Medlineplus.gov.: Trastorno del espectro autista: MedlinePlus en español. Available at https://

medlineplus.gov/spanish/autismspectrumdisorder.html. Accessed 12 Feb 2020 (2020)
2. Zhang, Y., et al.: ¿Could social robots facilitate children with autism spectrum disorders in
learning distrust and deception? 2019. China: s.n. 7, S0747-5632-(19)30148-7 (2019)
3. Discapacidad, Observatorio Nacional de: Conadis. https://www.conadisperu.gob.pe/observato
rio/estadisticas/inscripciones-en-el-registro-nacional-de-la-persona-con-discapacidad-julio-
2019/. Accessed 2 July 2019 (2019)
4. Opensurg, Consorcio. Robotica Medica. Elche: Cyted. 978-84-15413-12-7 (2013)
5. Sánchez Martín, F.M., Milán Rodríguez, F., Salvador Bayarri, J., Palou Redota J., Rodriguez
Escovar, F.: Esquena Fernández S, Villavicencio Mavrich H. Historia de la Robótica: de
Arquitas de Tarento al Robot Da Vinci (Parte I). Actas Urol Esp. 31(2), 69–76 (2007)
6. Pennisi, P., et al.: Autism and social robotics: a systematic review. Autism Res. 9(2), 165–183
(2015). https://doi.org/10.1002/aur.1527
7. Yaser, A.A., et al.: Robotic Trains as an Educational and Therapeutic Tool for Autism Spectrum
Disorder Intervention. W. Lepuschitz, Italy (2019)
8. Provenzi L.: Interaction with social robots: Improving gaze toward face but not necessarily joint
attention in children with autism spectrum disorder. IRCCS Eugenio Medea, Italia. https://doi.
org/10.3389/fpsyg.2019.01503
9. Kumazaki, H., et al.: Brief Report: Evaluating the Utility of Varied Technological Agents to
Elicit Social Attention from Children with Autism Spectrum Disorders, p. 2018. Springer,
Japan (2018)
10. Anys, 40: Autismo. Autismo. https://www.autismo.com.es/autismo/tratamientos-del-autismo.
html. Accessed 20 May 2019 (1976–2019)
11. Asociación Americana del Habla, L. y. Autismo Diario. Retrieved from https://autismodiario.
org/2017/04/17/desarrollo-la-comunicacion-autismo-traves-la-musica-juego/. Accessed 17
Apr 2019
12. VDI: A systematic approach to the design of technical systems and products. Guideline 2222
(1986)
130 M. K. Taype Mateo and G. W. Quispe Santivañez

13. Autodesk: Inventor 2019 [Software]. Retrieved from https://latinoamerica.autodesk.com/pro

ducts/inventor/overview (2019)
14. LabCenter. (s.f.).: Proteus 8.7 [Software]. Retrieved from https://www.labcenter.com/
A Smart Home Control Prototype Using
a P300-Based Brain–Computer Interface
for Post-stroke Patients

Sergio A. Cortez , Christian Flores , and Javier Andreu-Perez

Abstract In this paper, we present and compare the accuracy of two types of classi-
fiers to be used in a Brain–Computer Interface (BCI) based on the P300 waveforms
of three post-stroke patients and six healthy subjects. Multilayer Perceptrons (MLPs)
and Support Vector Machines (SVMs) were used for single-trial P300 discrimina-
tion in EEG signals recorded from 16 electrodes. The performance of each classifier
was obtained using a five-fold cross-validation technique. The classification results
reported a maximum accuracy of 91.79% and 89.68% for healthy and disabled
subjects, respectively. This approach was compared with our previous work also
focused on the P300 waveform classification.

Keywords P300 · Multilayered perceptron · Support vector machines · Stroke

patients

1 Introduction

A Brain–Computer Interface (BCI) is a system that records, analyzes, and clas-

sifies brain signals to control an external device or a computer. Also, it allows
healthy subjects or patients to communicate through brain activity [1]. The elec-
troencephalographic signals (EEG) are a suitable and common way to register and
analyze brain signals while a subject develops different sorts of mental activities.
Some of these, such as resting states and cognitive tasks, are specially studied for

S. A. Cortez (B) · C. Flores

Department of Electrical and Computer Engineering, Universidad de Ingeniería y Tecnología,
Lima, Peru
e-mail: [email protected]
C. Flores
e-mail: [email protected]
J. Andreu-Perez
Department of Computer Science and Electronic Engineering, University of Essex, Colchester,
UK
e-mail: [email protected]
© The Editor(s) (if applicable) and The Author(s), under exclusive license 131
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_13
132 S. A. Cortez et al.

BCI design purposes. Different paradigms such as motor imagery (MI), oddball,
and steady-state visual evoked potential (SSVEP) [2], elicit specific potentials that
are then used to build BCI systems. Speller is one application of a BCI based on the
P300 waveform which is commonly used as an alternative communication way, espe-
cially for disabled people. This application uses the oddball paradigm which consists
of showing infrequent stimulus (targets) blended with irrelevant stimuli sequences
(non-targets). For each target, 300 ms after, the P300 waveform is elicited and can
be identified in the EEG [3]. This paradigm requires less training periods than others
[4], thereby making it a successful tool for BCI design [5].
In [5], a system for the detection of visual P300 in healthy and disabled subjects
was developed. Bayesian Linear Discriminant Analysis (BLDA) was tested for the
classification of visual P300. The best classification accuracy was on average 94.97%,
and the maximum bit rate reached for the disable subjects was 25 bits per minute.
EEG data in [6] was classified using a Backpropagation Neural Network (BPNN).
The authors achieved a score classification of 96.3% and 96.8% for disabled and
healthy subjects. The best bit rate was 21.4 and 35.9 bits per minute for disabled and
healthy subjects, respectively. A support vector machine (SVM) algorithm and one
modification of the speller (6 × 6 to 4 × 10) were proposed in [7] to classify the
P300 visual evoked potential. The data from six healthy subjects were analyzed, and
the results reported an average accuracy of 75%. In [8], it was developed a P300 BCI
system based on ordinal pattern features and Bayesian Linear Discriminant Analysis
(BLDA). Their best bit rate was 16.715 bits per minute for disabled subjects and
26.595 bits per minute for healthy subjects. A stepwise linear discriminant analysis
(SWLDA) was proposed in [9] to classify P300 waves from a 6 × 6 speller. The
results from eight healthy subjects reported an average accuracy of 92%. In [10], a
Hidden Markov Model (HMM) was used to classify EEG data of a 6 × 6 speller
reaching a classification result of 85% offline and 92.3% online.
About Smart Homes using BCI, in [11], the authors proposed a random forest
classifier for P300 classify from three subjects. The best classification accuracy for
the subjects was on average 87.5%. A 6 × 6 character speller and icon speller were
compared and their performance analyzed in [12] for nine subjects. The authors
achieved a score classification of 80% for character speller and 50% for icon speller.
A Stepwise Linear Discriminant Analysis (SWLDA) was used to classify EEG data
from one subject with Amyotrophic Lateral Sclerosis (ALS) using a 6 × 6 row–
column speller to control television in [13]. The classification accuracy was 83%.
Regarding Virtual Reality Environments (VRE), [14] proposed to control a smart
home application using different speller masks (i.e., music, tv channels, etc.) testing
with twelve subjects. Average classification accuracy of 75% was reported.
This work presents a P300-based BCI for post-stroke people and healthy people.
The output of this BCI can be used to control home appliances. The aim of this work
is to discriminate the presence of the P300 waveform in EEG from 16 channels using
Multilayer Perceptron (MLP) and Support Vector Machine (SVM) classifiers. Results
outperformed our previous work [15]. This work is presented as follows: In Sect. 2
is presented the experimental procedure including participants, experimental setup,
and EEG acquisition. The methods are presented in Sect. 3 which describes prepro-
cessing, feature vectors, and classifiers. The results and discussion are presented in
Sect. 4. Finally, we present the conclusions and future work in Sect. 5.
A Smart Home Control Prototype Using a P300-Based … 133

Table 1 Subjects information

Subject Age Gender Diagnosis
S01 33 Male Healthy
S02 21 Male Healthy
S03 20 Male Healthy
S04 21 Male Healthy
S05 24 Male Healthy
S06 29 Male Healthy
S07 20 Male Hemorrhagic post-stroke
S08 52 Female Ischemic post-stroke
S09 55 Male Ischemic post-stroke

2 Experimental Procedure

2.1 Participants

The volunteers are grouped by healthy and disabled subjects, aged between 20 and
55. Table 1 shows the age, gender, and medical diagnosis of each volunteer. The
six healthy participants (S01 to S06) acted as the control group for the three post-
stroke patients (S07 to S09). Subject S08 exhibited limited spoken communication
and was able to perform restricted movements of his legs. Subject S07 exhibited
reduced spoken communication skills and was able to perform movements with his
extremities. Finally, subject S09 was able to perform restricted movements with his
hands and arms. The Ethics Committee from the Universidad Peruana Cayetano
Heredia issued the ethical approval for the experiment and informed written consent.
All participants that opted to participate were also informed about the academic
objective of the research, as well as ensured preservation of their anonymity.

2.2 Experimental Setup

Six images that are placed in two rows and three columns, above a white background
as shown in Fig. 1, were displayed on the screen of a computer. Each image represents
an option that provides the subject the capability to interact with another individual
or with its surroundings without speaking or moving his arms [5].
The protocol of the experiment, based on Hoffmann’s [5], is displayed in Fig. 2.
As can be seen, one image (out of six) is flashed in a random sequence during the
first 100 ms, and during the next 300 ms, as shown in Fig. 2, a white background is
displayed. This process is repeated until the completion of the six images sequence.
The group of six images randomly flashed is called one block, and one run represents
134 S. A. Cortez et al.

Fig. 1 Display of six images on a computer screen

Fig. 2 Stimuli timing diagram used in each experiment

the interval between 20 and 25 blocks, which are also chosen randomly in each exper-
iment. The database contains four recording sessions per subject, and each session
includes six runs. All subjects who participate in this experiment were forbidden to
talk and were told to count how many times the image they were told to pay attention
to appear on the screen. Each session was separated by one break of 10 min, and two
sessions were performed per day.

2.3 EEG Acquisition

The EEG signals were recorded using an electroencephalograph of the brand g.tec.
Sixteen bipolar electrodes (Fz, FC1, FC2, C3, Cz, C4, CP1, CP2, P7, P3, Pz, P4, P8,
O1, O2, and Oz) were placed in accordance with the international 10/20 system in
order to record the brain signals. The ground and reference electrode were placed
at the right mastoid and at the left earlobe, respectively. Also, the EEG signals were
sampled at 2400 Hz. Figure 3 displays the experiment setup. Furthermore, the C++
programming language and MATLAB R2016a were used to analyze the EEG data.
A Smart Home Control Prototype Using a P300-Based … 135

Fig. 3 Subject in P300 speller experiments

3 Methods

3.1 Preprocessing

A six-order bandpass Butterworth filter with cut-off frequencies 1 and 15 Hz was

used to filter the EEG data after it was downsampled from 2400 to 120 Hz. For each
electrode, all data samples between 0 and 1000 ms posterior to the beginning of the
stimulus were extracted obtaining 120 samples per trial. Artifacts were eliminated
by winsorization: each channel signal under the 10th percentile and above 90th
percentile was replaced by the 10th percentile or the 90th percentile, respectively
[5]. Finally, the data was standardized.

3.2 P300 Classification

Feature Vectors and Training Matrices The feature vector construction begins
rearranging the data from a single trial. Specifically, the samples of each channel are
concatenated with the others in the following way:
 
V = S11 S12 S13 . . . S116 S21 S22 . . . S120
16

where for a single sample Sik , i indicates the channel to which it belongs and k its
position with respect of time. The vectors obtained from all the trials present in the
four sessions of a subject are stacked shaping the training and testing matrices. Each
136 S. A. Cortez et al.

Fig. 4 EEG signal processing block diagram

trial, and therefore each V , has a label associated which points out the presence of a
P300 waveform in it. When that is the case, the trial is considered as target and when
not, non-target.
Since there is an uneven amount of trials target and non-target, due to the experi-
mental procedure for EEG recording, a balanced process is required before training
to avoid biasing in classification. The same number of target trials is chosen randomly
from a non-target trial pool, resulting in an even-class matrix. Two types of classi-
fiers will use these matrices for their training process: Multilayered Perceptron and
Support Vector Machine. The whole process is illustrated in Fig. 4.

Multilayered Perceptron Architecture Authors, like [16], had used neural

networks to discriminate the presence of ERPs in EEG signals. Multilayer Percep-
trons (MLPs) are neural nets that can be described as function approximators [17].
Multiple parameters, like weights and biases, are adjusted in the training stage trying
to match known outputs (target or non-target) with specific inputs (trials data). This
process results in a model able to predict the class or label of future new inputs with
similar characteristics.
The number of inputs, hidden layers, outputs, and total neurons defines MLP
architecture. Modifying these can completely change its behavior and could also
affect its capacity to find patterns and learn from them [18]. The architecture proposed
here for P300 discrimination consists of a net with four hidden layers of 60, 40, 30,
and 20 neurons on each, respectively. Each neuron has a hyperbolic tangent sigmoid
as its activation function, except for the output one, which is a logistic sigmoid.

Support Vector Machine It is a machine learning algorithm for binary classifica-

tion. The idea behind its classification criteria is the construction of a decision hyper-
plane in a high dimensional feature space were input vectors are mapped [19]. This
decision boundary can be linear or nonlinear depending on the kernel employed, and
it has a direct impact on the model generalization or learning capability. An SVM with
a Gaussian or radial basis function (RBF) kernel is trained for P300 discrimination
in EEG signals.

Classifiers Training and Validation The classification accuracy of each classifier is

obtained after fivefold cross-validation. For the MLPs, the inputs were renormalized
from −1 to 1 before these enter the net. In each subject training stage, the number
of epochs used was defined trying to achieve the net full convergence. All subjects
A Smart Home Control Prototype Using a P300-Based … 137

required only 150 epochs except for subjects S03, S04, and S07, which required
300, 2000 and 2500 epochs, respectively. The algorithm used in the MLPs training
process was the scaled conjugate gradient backpropagation. The training stage for
each MLP took approximately three minutes using an Nvidia GTX 1050 Graphics
Processing Unit (GPU) for the calculations; and for the SVM classifiers, it took less
than one minute in an Intel Core i7 CPU.

4 Results and Discussions

In this section, we compare the performance of the classifiers for all the subjects.
Table 2 presents the classification accuracy of the two classifiers in the training and
testing stage. It also highlights which of the two classifiers obtained higher results
in testing for a single subject. Overall, the performance of each healthy subject was
similar and above 80% with the exception of S04. Clearly, the MLP model achieved
a higher precision in classification over the SVM one. Nonetheless, both classifiers
can be used for the design of a P300 based BCI.
The performance of patients S08 and S09 was similar and even better than most
healthy subjects. The P300 waveform in subjects S01 and S08 was apparently easily
recognized by the MLP yielding the highest accuracies in classification. However,
patient S07 performance was the lowest in comparison with the rest. A possible
explanation for this is the patient critical condition (hemorrhagic stroke). Our results
show that patients who had suffered ischemic strokes are able to use a BCI based on
the P300 paradigm as their average classification accuracy is greater than 80%. In
comparison with our previous work [15], both classifiers presented here outperformed
the Adaptive Neuro-Fuzzy Inference Systems (ANFIS) ensemble for the healthy
subjects (S01, S02, S03, and S06) and post-stroke patients (S08 and S09).

Table 2 Cross-validated classifiers accuracy

Subject MLP RBF SVM
Training (%) Testing (%) Training (%) Testing (%)
S01 100 91.79 99.06 91.51
S02 100 80.28 97.27 79.07
S03 100 85.32 97.05 83.90
S04 100 75.70 95.18 78.89
S05 100 84.94 97.60 83.40
S06 100 83.00 97.56 81.65
S07 100 68.73 92.82 69.21
S08 100 89.68 97.68 85.46
S09 100 86.98 98.16 87.43
Bold indicates the two classifiers obtained a higher classification accuracy for a specific subject
138 S. A. Cortez et al.

5 Conclusions and Future Work

A comparison between the SVM and MLP classifiers for P300 discrimination in
EEG signals was presented. Both obtained significant results for the two types of
subjects present in this study. Patients who had suffered an ischemic stroke seem to
be suitable users of a P300-based BCI since their performances were similar to the
healthy ones. However, patients with more critical conditions, like the one who had
experienced a hemorrhagic stroke, may require another type of classifier with more
learning capabilities or a longer training period.
As future work, we intend to validate these classifiers with more post-stroke
patients and include amyotrophic lateral sclerosis (ALS) ones. We will also extend the
methodology presented in this work by employing deep learning tools and extreme
learning machine algorithms.

Acknowledgements This work used the database made in a collaborative project between the
Universidad Cayetano Heredia and the Universidad de Ingeniería y Tecnología supported by the
Programa Nacional de Innovación para la Competitividad y Productividad of Peru, under the grant
PIAP-3-P-483-14.

References

1. Wang, Y.T., Nakanishi, M., Wang, Y., Wei, C.S., Cheng, C.K., Jung, T.P.: An online brain-
computer interface based on SSVEPs measured from non-hairbearing areas. IEEE Trans. Neural
Syst. Rehabil. Eng. 25, 14–21 (2017)
2. Lotte, F., Bougrain, L., Cichocki, A., Clerc, M., Congedo, M., Rakotomamonjy, A., Yger, F.: A
review of classification algorithms for EEG-based brain–computer interfaces: a 10 year update.
J. Neural Eng. 15, 031005 (2018)
3. Farwell, L.A., Donchin, E.: Talking off the top of your head: toward a mental prosthesis utilizing
event-related brain potentials. Electroencephalogr. Clin. Neurophys. (1988)
4. Han-Jeong Hwang, S.C., Kim, S., Im, C.-H.: EEG-based brain–computer interfaces: a thorough
literature survey. J. Human–Comput. Interact. 29, 814–826 (2013)
5. Hoffmann, U., Vesin, J.-M., Ebrahimi, T., Diserens, K.: An efficient P300-based brain–
computer interface for disabled subjects. J. Neurosci. Methods 167(1), 115–125 (2008)
6. Turnip, A., Soetraprawata, D.: The performance of EEG-P300 classification using backpropa-
gation neural networks. Mechatron. Electr. Power, Veh. Technol. 4(2), 81–88 (2013)
7. He, S., Huang, Q., Li, Y.: Toward improved p 300 speller performance in outdoor environment
using polarizer. In: 2016 12th World Congress on Intelligent Control and Automation (WCICA),
pp. 3172–3175, June 2016
8. Alhaddad, M.J., Kamel, M.I., Bakheet, D.M.: P300-brain computer interface based on ordinal
analysis of time series. In: 2014 International Conference on Computational Science and
Computational Intelligence (2014)
9. Krusienski, D.J., Sellers, E.W., Cabestaing, F., Bayoudh, S., McFarland, D.J., Vaughan, T.M.,
Wolpaw, J.R.: A comparison of classification techniques for the p300 speller. J. Neural Eng.
3(4), 299 (2006)
10. Speier, W., Arnold, C., Lu, J., Deshpande, A., Pouratian, N.: Integrating language information
with a hidden markov model to improve communication rate in the p300 speller. IEEE Trans.
Neural Syst. Rehabil. Eng. 22, 678–684 (2014)
A Smart Home Control Prototype Using a P300-Based … 139

11. Masud, U., Baig, M.I., Akram, F., Kim, T.: A p 300 brain computer interface based intelligent
home control system using a random forest classifier. In: 2017 IEEE Symposium Series on
Computational Intelligence (SSCI), pp. 1–5, Nov 2017
12. Carabalona, R., Grossi, F., Tessadri, A., Castiglioni, P., Caracciolo, A., de Munari, I.: Light on!
real world evaluation of a p300-based brain–computer interface (BCI) for environment control
in a smart home. Ergonomics 55(5), 552–563 (2012)
13. Sellers, E.W., Vaughan, T.M., Wolpaw, J.R.: A brain-computer interface for long-term
independent home use. Amyotrophic Lateral Sclerosis 11(5), 449–455 (2010)
14. Holzner, C., Guger, C., Edlinger, G., Gronegress, C., Slater, M.: Virtual smart home
controlled by thoughts. In: 2009 18th IEEE International Workshops on Enabling Technologies:
Infrastructures for Collaborative Enterprises, pp. 236–239, June 2009
15. Achanccaray, D., Flores, C., Fonseca, C., Andreu-Perez, J.: A p 300-based brain computer
interface for smart home interaction through an ANFIS ensemble. In: 2017 IEEE International
Conference on Fuzzy Systems (FUZZ-IEEE), pp. 1–5, July 2017
16. Gupta, L., Molfese, D.L., Tammana, R.: An artificial neural-network approach to ERP
classification. Brain Cogn. 27(3), 311–330 (1995)
17. Winston, P.H.: Artificial Intelligence, 3rd edn. Addison-Wesley Longman Publishing Co., Inc.,
Boston, MA, USA (1992)
18. Wu, F.Y., Slater, J.D., Honig, L.S., Ramsay, R.E.: A neural network design for event-related
potential diagnosis. Comput. Biol. Med. 23(3), 251–264 (1993)
19. Cortes, C., Vapnik, V.: Support-vector networks. Mach. Learn. 20(3), 273–297 (1995)
Electronic Ecomap as an Instrument
to Improve Physician–Family Interaction
in Preventive Medicine

Néstor Mamani-Macedo , Julio Raime , Luzmila Pro , Jaime Pariona ,

and Olga Solano

Abstract According to the World Health Organization (WHO), the main causes of
mortality in low-income countries are communicable diseases, maternal and peri-
natal conditions, and nutritional conditions. The issue arises because early detection
requires the implementation of a primary healthcare system. As part of the family
record, we have the Ecomap, an instrument for recording family relationships with
their environment in their day-to-day activities graphically, whether this comprises
people or institutions such as the church and health centers. In Peru, the use of Ecomap
has traditionally been in a hard-copy format, which has problems updating because it
is handwritten. Given this limitation, a computer application “electronic Ecomap” (e-
Ecomap) has been developed, a resource with which healthcare personnel can interact
in a user-friendly manner. In an aggregate mode, it will be possible to visualize the
families of a community as a whole, highlighting the behavior of the institutions
surrounding them. In this manner, it will be possible to prevent the effects of these
risks in a generation of psychosocial problems that can later lead to psychosomatic
illnesses. Finally, the e-Ecomap will be able to infer the connections of the family
with its environment in conjunction with the data registered in the family record.

Keywords Electronic ecomap · Family record · Preventive medicine

N. Mamani-Macedo (B) · J. Raime · O. Solano

Facultad de Ciencias Matemáticas, Universidad Nacional Mayor de San Marcos, Lima 1, Peru
e-mail: [email protected]
J. Raime
e-mail: [email protected]
O. Solano
e-mail: [email protected]
L. Pro · J. Pariona
Facultad de Ingeniería de Sistemas e Informática, Universidad Nacional Mayor de San Marcos,
Lima 1, Peru
e-mail: [email protected]
J. Pariona
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 141
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_14
142 N. Mamani-Macedo et al.

1 Introduction

During the last twenty years, Peru has shown a steady growth of between 4 and 9%,
excluding 2009, given the problems faced in the global economy, 2014, and 2017
because of the decrease in the price of minerals and weather effects such as “El Niño”
[1, 2]. Nevertheless, according to Miguel Malo, advisor to the Pan American Health
Organization, there is still a need in primary healthcare to close gaps in services; for
instance, in the area of chronic diseases, improving the resolution capacity of frequent
diseases such as hypertension and diabetes is one such concern [3]. According to
the WHO [4], the main causes of death in low-income countries are communicable
diseases, maternal and perinatal conditions, and nutritional conditions. Chronic and
communicable diseases are multicausal, and the company of family and other external
entities, such as the church, health centers, neighborhood associations, and public
and private social support institutions is of utmost importance in their treatment.
The Ecomap (ecological map) is a graphic representation that shows all the enti-
ties that are involved in the life of an individual or family (ecological system), and its
results can be observed in a simple manner. Ann Hartman developed the instrument
in 1975 [5], and its use soon spread from social workers to nursing and subsequently
to human medicine; at present, it is widely used by primary care physicians. In Peru,
an Ecomap is a part of the family record [6], used in a hard-copy format. The purpose
of the family record is to facilitate a biopsychosocial approach to a family and its
members, assessing their health problems within the economic, social, and environ-
mental context in which the family operates. A genogram provides us with a historical
image of the family through and between generations, but the Ecomap locates the
family in its current social context, providing a visual map of the connections of
the family and its members to the external world. These connections can include,
for example, the relationship between family members and their neighborhood or
with the public entities of social, cultural, sports, educational, and religious support,
among others [7]. In this context, our proposal is to develop an electronic ecomap
linked to an electronic family record application.

2 Theoretical Framework

2.1 Ecomap

An Ecomap is an instrument that allows the physician and health team to identify the
interrelationships of a family with the environmental and socio-cultural context in
which it operates quickly, thereby providing information on the extra-family network
of resources. Therefore, it constitutes a complement to the family diagram and visual
aid for an understanding of the environment in which the life of families takes place.
Its routine use allows for representing the family and its contacts with their supra
systems, in other words, with the environment that surrounds the family, namely,
Electronic Ecomap as an Instrument to Improve Physician–Family … 143

extended family, health systems, work, educational institutions, recreation, religious

institutions, friends, neighbors, etc. On one hand, from these contacts or relationships,
resources can be obtained that the family can use to meet its objectives in an adequate
manner. On the other hand, these contacts can become considerable energy drains,
thereby making it difficult for the family to achieve their goals. Hartman [5] described
the Ecomap as a picture of the family and its situation, highlighting the connections
extra-family positive and negative nature, mapping points of conflict to be mediated.
The solution implies finding the resources necessary, internal and external, to mitigate
them. Ecomap enables the family doctor to accomplish the following [8–11]:
• Identify extra-family social support resources that are absent and/or present and
that may be useful in situations of family crises.
• Detect shortcomings that the family is going through with the goal of working to
address them.
• Enhance the doctor–patient-family relationship as well as communication.
• Suspect problems both inside and outside the family, which can be corroborated
with the application of other instruments.
In Peru, the Ecomap is currently used in a hard-copy format [6], and healthcare
personnel in charge of the family visit must draw the existing relationships between
the family and its environment. In this process, the risk of making errors in the
drawing is present in addition to the possibility of corrections affecting the ease of
reading that an Ecomap must have; in the worst-case scenario, a new format should
be used. Consequently, there is an increase in the family visit time, thereby increasing
the dissatisfaction of families with the family doctor visit. In addition, when using a
printed format, the family’s ecomap has to be redrawn at each visit, which may fall
back on the aforementioned errors.

2.2 State-of-the-Art

Ecomap is a tool used in several countries in the field of family medicine and its
electronic implementations are:
(1) Design templates on generic online platforms such as Canva,1 Smartdraw,2
and Scribblar3 that allow for the creation of Ecomaps from scratch or for the
customization of any of the predesigned templates.

1 https://www.canva.com/graphs/ecomap/.
2 https://cloud.smartdraw.com/.
3 https://scribblar.com/.
144 N. Mamani-Macedo et al.

(2) Standalone applications, which are tools that allow for the creation of
Ecomaps and storing the data in a database; these also use or customize prede-
fined templates. They include Genogram Analytics4 and EcotivityTM5 from
WonderWare Inc.

2.3 Computer Technology

The following technologies were used for the development of our proposal:
• SVG (Scalable Vector Graphics)6 is an XML language used to draw vector
graphics. It can be used to create an image either by specifying all the necessary
lines and shapes, by modifying existing images or through a combination of both.
• MVC (model–view–controller)7 is a type of software architecture that separates
data from the model, the user interface, and the control logic.
• JAVA8 is an object-oriented programming language that allows for code porta-
bility because it runs on a virtual machine (Java virtual machine or JVM). There
are currently JVMs for all known platforms.
• MySQL9 is a database management system (DBMS) system for managing data
tables.

3 Developed Solution

3.1 Methodology

The prototyping methodology was used for the development of the software, which
is also known as the evolutionary development model [11], beginning with the defi-
nition of the global objectives for the software, followed by identifying the known
requirements and the areas of the schema for which more detail is required. The
construction of the prototype ensures better software quality as well as an appealing
interface for users.
To develop the electronic ecomap, the format to be used was first proposed, which
should be in accordance with the programming language of the family file application;
SVG was selected for its versatility. Subsequently, the Ecomap was integrated with
the electronic family record to collect information about family members.

4 http://www.genogramanalytics.com/index.html.
5 https://www.interpersonaluniverse.net/.
6 https://www.w3.org/Graphics/SVG/.
7 https://www.codecademy.com/articles/mvc.
8 https://go.java/index.html.
9 https://www.mysql.com/.
Electronic Ecomap as an Instrument to Improve Physician–Family … 145

Fig. 1 Main classes that interact with the ecomap

The family record has the family visit as one of its components, with the Ecomap
being one of its instruments. Therefore, its implementation could significantly
improve the interaction between the family doctor and the family, enhancing confi-
dence with respect to home visits and thus creating a pleasant environment for both
of the parties involved.
Figure 1 shows the main system classes that interact with electronic ecomap.
With information captured from the characteristics of the family and home and
the member registry, given the risks that each may present, the aim is to infer the
connections that the family may maintain with its environment and, by doing so,
recommend these connections to the healthcare personnel. To begin with, this infor-
mation is based on statistics from previous visits or from other families, thereby
allowing the system to learn. Figure 2 shows the flow diagram of the process carried
out when working with the electronic ecomap. In addition, the developed solution
does not require an internet connection, which also facilitates its use in rural or
remote areas, such as in the forest or Peruvian highlands, where it is difficult to
reach families and help them with respect to the prevention of diseases. The solution
developed is in accordance with the needs of the field of preventive medicine and
meets the requirements of the Ministry of Health from Perú.
146 N. Mamani-Macedo et al.

Fig. 2 Process diagram of the electronic Ecomap

3.2 Case Study

To validate our proposal, we use information from the INEI10 (National Institute
of Statistics and Informatics) obtained in the national household survey on living
conditions and poverty and choose to analyze the anonymized information of a lower-
middle-class family. Family members attend social programs that provide economic
assistance. Their relationship with community institutions is stressful; they attend
a church where they are required to tithe and shop at their uncles’ store, which is
close to the neighborhood. The father wants to leave work soon, and the children
have missed school four times in the last two months. The family prefers to self-
medicate rather than attending a health center. All of the aforementioned generates
the following information concerning the connections with their environment:
• Stress with friends and neighbors
• The flow of inputs with social programs
• Stress with community institutions
• The flow of inputs with the church
• The flow of inputs with relatives outside the household
• Weak regarding work
• Weak regarding school
• Weak regarding their interaction with health centers.
Having this information will help build their ecomap as shown in Fig. 3a, with
the help of the electronic ecomap.

10 https://www.inei.gob.pe/.
Electronic Ecomap as an Instrument to Improve Physician–Family … 147

Fig. 3 a Electronic Ecomap and b the menu to select connections

To start creating connections with external agents, one must select the agent with
the mouse and click on it; the program will display a dropdown menu of four available
options: weak, strong, flow, and stressful as shown in Fig. 3b. By selecting the desired
connection(s) and clicking on either “Add” or “Delete,” the Ecomap will update the
drawing associated with the selected connection.

4 Discussion

The main problem with the family record in the printed format lies in its management,
transportation, reading, reliability, and integrity, which are characteristics that must
be improved and updated according to state-of-the-art technology. Our research group
previously developed the electronic family record and, as an essential part of this,
now, we have implemented the electronic ecomap, a virtual version of the ecomap
currently used in the hard-copy family record. To safeguard the integrity of the
database, using a database that is not visible to anyone but the personnel working
on it has been proposed; consequently, readings would only be possible through the
electronic family record software.
The electronic family record and the electronic ecomap positively influence the
manner in which the family accepts the visit of a family doctor because the visit
does not take more time than necessary, which can occur if there are errors in the
family record or in the ecomap in the printed format. In addition, the family will be
more motivated in the next visit because of the clarification that the main objective
of the visit is to help them in preventing diseases and illnesses. There are some
other aspects of improvement for the ecomap. One of its main shortcomings is the
treatment of the electronic ecomap an object and not as a mere drawing. Addressing
this shortcoming is possible with the use of the object-oriented programming (OOP),
through which an ecomap object can be connected with other objects to facilitate
148 N. Mamani-Macedo et al.

comprehension or help improve the reading of other problems that the family might
present. Being able to link the Ecomap with the genogram, for instance, will allow
for a greater understanding of the family with respect to external and internal entities
that influence the health of the family.

5 Conclusion

In the electronic family record, we focus on solving the problems of biopsychosocial

problems in the family with its environment. In the electronic ecomap, the positive
and negative interactions between the family and all the entities that are involved
around it were displayed. In this work, we demonstrate its importance in the area of
preventive medicine.
As future work, an electronic signature used by each healthcare staff member
will be developed in order to member to verify authenticity, to identify responsible
individuals/users for each visiting user, and to determine whether the healthcare
personnel signed the electronic ecomap as established. This could help to strengthen
the family´s trust in the healthcare crew.

References

1. BCRP: Banco Central de Reserva del Perú. Memoria (2014)

2. BCRP: Banco Central de Reserva del Perú. Memoria (2017)
3. Arroyo, J., Arroyo, J., Velásquez, A., Céspedes, S., Malo, M., Pedroza, J., Saco, A., Zamora,
V.: La Salud Hoy: Problemas y Soluciones. CENTRUM. PUCP (2015)
4. World Health Organization: The top 10 causes of death. https://www.who.int/es/news-room/
fact-sheets/detail/the-top-10-causes-of-death. Accessed 30 Nov 2019
5. Hartman, A.: Diagrammatic assessment of family relationships. Soc. Casework 59, 465–476
(1978)
6. Ministerio de Salud del Perú: NTS N°139-MINSA/2018/DGAIN: Norma Técnica de Salud
para la Gestión de la Historia Clínica. Resolución Ministerial N° 214-2018/MINSA
7. Jesuit Social Services: Strong Bonds. http://www.strongbonds.jss.org.au/workers/cultures/eco
maps.pdf. Last accessed 30 Nov 2019
8. Cox, Ruth P.: Health-Related Counseling with Families of Diverse Cultures: Family, Health
and Cultural Competencies. Greenwood Press, Westport, CT (2003)
9. Modelo de atención integral de salud basado en familia y comunidad. https://determinantes.
dge.gob.pe/archivos/1880.pdf
10. El ECOMAPA: medicina familiar, educación continuada para el médico general. https://enc
olombia.com/libreria-digital/lmedicina/ecmg/fasciculo-1/ecmg1-recursos2/
11. Gowtham, V., Manoj, Y., Pooventhiran, G., Praveen, A., Shivaram, R., Kathiresan, A.: Evolu-
tionary models in software engineering. Int. J. New Technol. Res. (IJNTR). 3(5) (2017). ISSN
2454-4116
Intelligent Management Model in an IoT
Environment Applied for the Healthcare
Sector

Jessie Bravo and Roger Alarcón

Abstract This work pretends to solve the inadequate management of devices and
their integration in an Internet of Things (IoT) environment which limits the perfor-
mance of the network in a hospital. The objective of the work is to develop an
integration plan, based on an intelligent management model of devices in an IoT
environment. The model was validated through the digital transformation maturity
model and is applied to the cardiology department at the hospital as a part of the
integration plan in the healthcare sector.

Keywords Internet of Things · Digital transformation maturity model · Healthcare

sector

1 Introduction

According to the International Telecommunications Union [1], IoT can be defined as

a dynamic global infrastructure for the information society with self-configuration
capability based on standard communication protocol interconnecting physical and
virtual things, which have identities, physical attributes, and virtual personalities;
they use smart interfaces and are perfectly integrated inside the information network.
It is evident that many problems to overcome exist in the Peruvian Healthcare System.
Peruvian Government budget allocates only 2.2% of the Gross Domestic Product
(GDP) in healthcare sector; this is much lower only compared to other Latin American
countries [2], problems like the scarce infrastructure, which generates a saturation
of patients in hospitals and healthcare centers, the poor management of these care
centers, which cause an inadequate quality of the service for the patients, the lack

J. Bravo (B) · R. Alarcón

School of Computer and Informatics Engineering, Pedro Ruiz Gallo Public University,
Lambayeque, Peru
e-mail: [email protected]
R. Alarcón
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 149
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_15
150 J. Bravo and R. Alarcón

of medical staff which doesn’t allow a timely patient care, the lack of automation of
monitoring and patient control, which creates a lack of immediate action in the face
of a patient critical situation.
This work pretends to apply computational intelligence to a device management
model in an IoT environment, which allows the improvement of the network perfor-
mance in the healthcare sector in order to overcome the inefficient management of
medical devices and processes.

2 State of the Art

Regarding the management of an IoT system, in [3], they clearly indicate that IoT
systems and pure network solutions differ one from the other because the latter offer
low-level services and support for business administration, and an IoT system is
vastly more complex than a communication system.
An IoT proposed architecture model by Kosmatos et al. [4] explains the architec-
ture based on three layers: the RFID’s perspective, the smart object’s perspective,
and the social perspective, but does not consider the response time in real-time and
the diversity of technologies of existing connectivity.
In addition, there is a model proposed by the ITU-T [1] consisting of four layers:
application, support services, and applications, network, and device with manage-
ment and security capabilities related to them, approved in 2012. This model does
not focus on device management.
The different investigations carried out in this field corroborate these advances.
In [5], the IoT benefits the healthcare sector, such as improving availability and
accessibility, the ability to personalize content and a cost-effective delivery, being
energy consumption one of the critical points, all of this will be achieved through
new IoT technologies such as smart portable sensors for health care, body area
sensors, advanced generalized health care systems, and big data analysis. In [6] is
presented a proposal to improve the Quality of Service (QoS) in terms of throughput
of cognitive body area networks by performing a mathematical channel model for
off-body communication. In [7], the results of an exhaustive survey of IoT technolo-
gies, methods, statistics, and success stories applied to healthcare were presented.
Furthermore, [8] carry out an analysis of current solutions in the healthcare sector,
proposing a healthcare model where benefits such as a lower labor intensity and
a lower operational cost are entailed. In [9] is proposed a generalized monitoring
system capable of sending patients’ physical signs to remote medical applications
in real-time. The system consists of two parts: the data acquisition part and the
data transmission part. An IoT implementation in the healthcare sector is proposed
by Sokac [10], who designed a Holter monitor that has electrodes and other elec-
tronic components integrated into an elastic shirt, and it is able to communicate to a
smartphone via Bluetooth, allowing a permanent patient monitoring.
Intelligent Management Model in an IoT Environment … 151

3 Proposal

In this paper, an intelligent device management model was developed in an IoT

environment, and it was applied to the healthcare sector specifically, taking into
account, mainly, computational intelligence, the ITU-T device management archi-
tecture, the IoT environment architecture and having the knowledge that achieving
network quality of service implies, at least, complying with the following aspects:
reliable and consistent connectivity, a device real-time operation, efficient online
monitoring, and error detection. For the proposed model elaboration, five dimen-
sions have been taken into account: the device dimension, the network dimension,
the manager dimension, the provider dimension, and the user dimension, as shown
in Fig. 1.
Furthermore, the security is presented in a transversal way to the model and all
of this, from a holistic standpoint, will allow integration into an IoT environment
applied to the healthcare sector.
The supplier dimension will allow managing the real-time data generated from
several IoT and non-IoT devices, and it will facilitate its access anywhere, at any
time.

Fig. 1 Smart device management model in the IoT environment

152 J. Bravo and R. Alarcón

In this dimension, there are three possible ways of storing the information gener-
ated by the devices that were considered: Cloud Computing, Fog Computing, and
Edge Computing. A Cloud Computing solution is based on data centers with the
capacity to process and store large volumes of data, having as a disadvantage that
they are generally far away and are slow to respond, and it is also subject to fluctu-
ation that is defined as a variation in the delay of the received packages [11]. Some
IoT devices require real-time processing to make decisions and execute immediate
actions, also greater mobility and a large number of devices, consequently, in those
cases, consider solutions such as Fog Computing and Edge Computing is a neces-
sity [12, 13]. These two architectures can allow the immediate use of the data of
the equipment and/or medical sensor, which will deliver better results. For example,
glucose values and insulin doses will probably require an immediate response and,
therefore, require an automatic analysis and an immediate load in the cloud with Fog
Computing or Edge Computing.
As for the device dimension, it is necessary to, first, explain what is understood
as a device: physical equipment that can be connected to a network and has some
degree of computing and storage capacity. Identification technology will allow the
unique identification of the resource.
For this work, a device that interacts directly with users is called a terminal device,
such as personal computers, laptops, smartphones, and others. Instead, a device that
can help restricted resources to connect to the Internet is called the agent device,
such as the wireless sensor networks (WSN) gateway, the RFID reader. The devices
that connect directly to the Internet, such as sensors, can also be identified. In this
specific case, there are also medical devices, such as tomography, X-ray equipment,
ECG, which use the Internet for medical control or self-support. The four types of
devices, mainly identified, are:
• Terminal device: a PC, a laptop, a tablet, or a Smartphone.
• Agent device: such as a WSN Gateway, an RFID reader.
• Sensor device: as a temperature, blood pressure, or glucose sensor.
• Medical device: for example, tomography, ECG, X-ray.
The processing capacity is another form of classification, and the devices were
classified as basic, regular, high, and the classification is based on their storage
memory, the response time, and the energy duration of the device. To achieve the
quality of service of the network in this, dimension is necessary to take into account,
mainly, the quality of the device: Here, the number of users connected simultaneously
to the device must be taken into account, also the device capacity to support a high
volume of traffic data and their capacity to work in critical situations.
In the network dimension, the interconnection of the different devices identified
in the previous dimension is contemplated, mainly through: Ethernet, WLAN, WSN,
RFID, WPAN, WBAN, and WWAN.
In this dimension, in order to achieve an improved quality of service of the
network, the following aspects need to be taken into account:
Intelligent Management Model in an IoT Environment … 153

• Quality of the network: If the network is capable of supporting applications in

real-time if the traffic is prioritized, and the type of data transmitted over the
network.
• Transmission quality: Aspects such as bandwidth, throughput, latency, jitter, and
error rate are measured.
Furthermore, in the device and network dimension, computer intelligence is
applied, specifically, machine learning and inaccurate computing that allows the
device to make a smart decision and selects how to respond to failure and how
to access critical information in real-time that allows an immediate response. The
devices must be able to learn from the generated data stored on the Internet in such
a way that it will allow immediate access from anywhere. And through inaccu-
rate computing, it will allow prioritizing according to three types of levels: critical,
acceptable, and normal and have an adequate response.
The decision trees are the machine learning classification technique that was
applied, and the following variables are used in the learning:
• Device type (Medical, Sensor, or Terminal)
• Processing Capacity (Basic, Regular, and High)
• Device priority (Critical, Acceptable, Normal)
• Health condition (Type 0, 1, 2, and 3)
• Network Platform (Ethernet, WLAN, WSN, RFID, WPAN, WBAN, WWAN).
Where the objective is to determine the supplier that best suits each need. The
supplier can be Cloud Computing, Fog Computing, and Edge Computing.
In the decision tree, a series of observations are mapped based on the variables
mentioned above, and it is determined which would be the best provider, for this the
algorithm C5.0 has been used.
As for the user dimension, there are three different roles: patient, doctor, and
network administrator. According to the severity of the health condition of the patient,
it was classified as:
• Type 0: highest priority requires a real-time monitoring.
• Type 1: requires almost real-time monitoring in every few hours.
• Type 2: requires periodic monitoring, twice a day possibly.
• Type 3: requires monitoring from time to time.
This dimension includes two key aspects to achieve an effective quality of service
of the network:
• The density of users: The quality of the service will be influenced by the number of
users who will use of patient monitoring and follow-up services. The classification
is as follows: high-density, medium-density, and low-density.
• Access level: There is a necessity to define what rights and privileges those users
will have to safeguard the security and privacy of the stored data. The device is
important, but the knowledge of who is accessing it, from where the access is
coming from, and the time of the access is equally important. Regarding the level
of access, the following have been considered: high, medium, and low.
154 J. Bravo and R. Alarcón

The management dimension is the graphical interface that will allow controlling
and monitoring the resources. First, the security management that has been framed as
a transversal task immersed in each dimension of the presented model, and second,
the alert and sensor management in real-time since this aspect is critical for the
medical devices in the context of the proposed model.
The five dimensions defined in a transversal manner include security management,
having the need to safeguard the privacy of stored data. To achieve this, aspects
of reliability, availability, and integrity in each of the dimensions must be checked.
Finally, another considered aspect is, to achieve the integration of existing IoT devices
in the healthcare sector to a data network, which will allow improving the service of
the hospital that is being studied, mainly in two aspects: real-time patient monitoring
and online access to the patient’s clinical information.

4 Validation and Discussion

The project was carried out focusing on one of the most relevant departments of
the hospital—case study, the cardiology department. The reason is the increase in
cardiovascular diseases in the last years. Figure 2 shows the proposed modeling
whose validation is further discussed after.

Fig. 2 Model applied to the case study

Intelligent Management Model in an IoT Environment … 155

To corroborate the intelligent management model in an IoT environment in the

healthcare sector, to assess whether it is feasible to adopt such a proposal in the
organization under study and determine its level of maturity, the application of the
maturity model tool was also carried out.
The maturity model suitable for its application is the digital transformation matu-
rity model of an organization and for this purpose [14]. The maturity levels derivate
from the model are as follows:
1. Beginner: The digital channels are incipient, the web interfaces are not used in an
efficient way, and the products or services are not offered through it. Management
is still reluctant to change and it is immature in terms of digital culture, and the
benefits it generates in the company.
2. Medium level: The customer requirements are taken into account but it is still
lacking a customer-centric approach, their digital channels are not completely
developed, and the mobile channel is inefficient. These companies recognize the
necessity of change, and the digital culture is present in some small groups or
departments, but it’s necessary to be extended to the entire company.
3. Advanced: Many innovative initiatives are carried out, which allows the devel-
opment of digital culture. All channels are fully developed, and the products and
services are offered on all channels. The omnichannel experience is complete.
The user experience is taking into account, an advanced analysis of the data is
applied, and the customer is the priority.
4. Expert: The digital transformation is fully applied, and the change is welcome.
The customer-centric and omnichannel is approached from the design. Business
intelligence is applied at all levels of the company. The digital culture is part of
the strategic plan.
The maturity model takes into account three dimensions: The culture and orga-
nization, the technology and the business. The following formulas were applied to
determine the level of maturity:
For the score calculation in each dimension:
n
Scorecriteria
ScoreDimension = i=1 (1)
n
For the score calculation of the model:
nd
j=1 ScoreDimension
ScoreModel = (2)
nd

Table 1 Score by dimension

Dimension Dimension score
Culture and organization 14.67
Technology 7.33
Business 8.67
156 J. Bravo and R. Alarcón

Table 1 summarizes the scores regards each dimension after applying the
validation instrument to the IT staff of our study case.
For the culture and organization level, an advanced maturity level is reached,
given that the organization has the disposition and the digital culture that is required
to adopt the change that is proposed. Regarding the technology dimension, a medium
maturity level is reached, since the digital channels are not completely implemented.
For the business dimension, the medium maturity level is reached, where employees
are involved in the change and the need for the use of analytics for decision making
is recognized. The average score of the model is 10.22, this means that the company
is in the medium maturity level, so it is concluded that it is feasible to adopt the IoT
in our case study.

5 Conclusions

The digital transformation maturity model was applied, obtaining an average level
of maturity, so it is concluded that the application of the proposal in the place under
study is feasible. It was partially exemplified in a private hospital specifically in
the area of cardiology demonstrating its contribution and the achievement of the
objectives set.
As future work, it is intended to develop a methodology based on each of the
dimensions of the proposed model.

References

1. U.-T. Y.2060: Unión Internacional de las Telecomunicaciones. https://www.itu.int/rec/T-REC-

Y.2060-201206-I/es. Last accessed 18 Jan 2019
2. Cabani, L.: Diario El Peruano. https://www.elperuano.pe/noticia-la-gestion-de-recursos-salud-
77727.aspx. Last accessed 10 July 2019
3. Jurado Perez, L., Velásquez Vargas W., Vinueza Escobar, N.: Estado del Arte de las Arquitec-
turas de Internet de las Cosas (IoT). In: Academia.edu, pp. 1–48. Springer, Berlin, Heidelberg
(2014)
4. Kosmatos, E., Tselikas, N., Boucouvalas, A.: Integrating RFIDs and smart objects into a unified
Internet of Things architecture. Adv. Internet of Things 1, 5–12 (2011)
5. Firouzi, F., Farahani, B., Ibrahim, M., Chakrabarty, K.: From EDA to IoT eHealth: promise,
challenges. IEEE Trans. Comput. Aided Des. Integr. Circ. Syst. 37(12), 2965–2978 (2018)
6. Ahmed, T., Le Moullec, Y.: A QoS optimization approach in cognitive body. Sensors (Basel)
17(4), 780 (2017)
7. Castro, D., Coral, W., Cabra, J., Colorado, J., Méndez, D., Trujillo, L.: Survey on IoT solutions
applied to healthcare. DYNA 84, 192–200 (2017)
8. Sanmartín Mendoza, P., Ávila Hernández, K., Vilora Núñez, C., Jabba Molinares, D.: Internet
de las cosas y la salud centrada en el hogar, Revista Salud Uninorte 32(2), 337–351 (2016)
9. Li, C., Hu, X., Zhang, L.: The IoT-based heart disease monitoring system for pervasive. Procedia
Comput. Sci. 112, 2328–2334 (2017)
10. Sokac, M.: A new design for a Holter monitor based on Internet of Things technology.
ResearchGate 1–5 (2017)
Intelligent Management Model in an IoT Environment … 157

11. Kumar Sahoo, P., Kumar Dehury, C.: Design and implementation of a novel service
management framework for IoT devices in cloud. J. Syst. Softw. 119, 149–161 (2016)
12. OpenFog: Reference Architecture for Fog Computing. http://www.OpenFogConsortium.org,
http://www.springer.com/lncs. Last accessed 10 Jan 2019
13. Progress Software Corporation: Progress Open Edge. https://documentation.progress.com/
output/ua/OpenEdge_latest/index.html#page/gsdev%2Fgetting-help-using-the-openedge-kno
wledge-servic.html%23. Last accessed 6 Feb 2018
14. Paradigma: paradigmadigital.com. https://dtma.paradigmadigital.com/. Last accessed 25 Aug
2019
Evaluation Method of Variables
and Indicators for Surgery Block Process
Using Process Mining and Data
Visualization

Piero Rojas-Candio , Arturo Villantoy-Pasapera ,

Jimmy Armas-Aguirre , and Santiago Aguirre-Mayorga

Abstract In this paper, we proposed a method that allows us to formulate and eval-
uate process mining indicators through questions related to the process traceability,
and to bring about a clear understanding of the process variables through data visu-
alization techniques. This proposal identifies bottlenecks and violations of policies
that arise due to the difficulty of carrying out measurements and analysis for the
improvement of process quality assurance and process transformation. The proposal
validation was carried out in a health clinic in Lima (Peru) with data obtained from
an information system that supports the surgery block process. Finally, the results
contribute to the optimization of decision-making by the medical staff involved in
the surgery block process.

Keywords Process mining · Data visualization · Healthcare · Surgical block ·

Medical variables

1 Introduction

The Social Health Insurance, ‘EsSalud’, is a decentralized public organization whose

purpose is to provide coverage to insured people and their right-holders, through the
provision of different types of insurance to the population in the face of human’s risks

P. Rojas-Candio · A. Villantoy-Pasapera (B) · J. Armas-Aguirre

Universidad Peruana de Ciencias Aplicadas (UPC), Lima 15023, Peru
e-mail: [email protected]
P. Rojas-Candio
e-mail: [email protected]
J. Armas-Aguirre
e-mail: [email protected]
S. Aguirre-Mayorga
Pontificia Universidad Javeriana, Bogotá, Colombia
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 159
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_16
160 P. Rojas-Candio et al.

[1]. This institution provides care to approximately 11,493,000 insured Peruvians,

who represent 35.7% of the total population [2]. These insured people are concen-
trated in a higher proportion in ages from 0 to 14 years, from 25 to 44 years and
65 to more [3]. According to the annual report drawn up in 2017 by this institution,
it shows that there is a total of 28,149 complaints registered in the Insured Service
information system, which represents 18.8% of the total requests for that system with
a 25-day completion time [4]. Also, according to the newspaper El Comercio, one of
the main reasons for the complaints is due to the lack of access for health services due
to the waiting time of care for the insured [5]. EsSalud seeks to propose strategies to
reduce these complaints and waiting times, but such investigations involve a higher
work effort and use of human resources. The results of the tests do not show to be
very effective since the insured may still be dissatisfied because the attention time is
still high [6].
In recent years, the development of technology has been enormous, and this is
because support for different industries in their organizational processes is vital [7],
as in the case in the health sector, which has access to large volumes of data in the
information systems they manage to support their business processes. At this point,
process mining allows extracting process execution information from the informa-
tion systems available in an organization to build process models, compare existing
models with those built automatically to identify bottlenecks and deviations and to
improve business processes [8].

2 Literature Review

Given the importance of the proper structuring of this life-saving process, previous
studies that applied the discipline of process mining in various industries were
reviewed to visualize the results obtained in the case studies and the contribution
that the authors presented.
Therefore, as can be seen in Table 1, a comparison of process mining tools made in
research was taken as a Ref. [7], since this is aligned with our purpose of identifying
an easy-to-use tool that incorporates the disciplines of Process Mining and Data
Visualization. The tool chosen in this research is Celonis, for its ease of use and the
excellent user experience it provides.

3 Proposed Method

3.1 Proposal Description

The method to be presented uses the seven activities of [15] because indicators are
planned to provide quantitative and real answers about the surgery process based on
the specific needs of stakeholders [7], which is mainly designed according to the PM2
methodology. Our method proposes a definition of variables within data visualization
Evaluation Method of Variables and Indicators … 161

Table 1 Comparison of process mining tools. [7]

Tool ProM Disco Celonis QPR Process
Analyzer
References [9–11] [10, 12] [10, 13, 14] [10, 11]
Import type MXML, XES CSV, XLS, CSV, XLS XES
support MXMl, XES y
FXL
Amount of event Unlimited 5 million events Unlimited 10 million
logs events
Output model BPMN, Work flow, Fuzzy model Fuzzy model Work flow
notation Petri nets, EPCs,
Transition systems
Monthly cost Free $1618 $598–1929 $1126
Number of users Unlimited 1 1 1
per license
Support time – Unlimited Unlimited 1
(Months)
License – X X X
Filtering data X X X –

techniques to understand in depth what is going to be represented graphically and,

at the same time, be of interest to those responsible for the surgery block process at
a business level. Figure 1 shows the evaluation method.

3.2 Stages of the Method

Objectives and Question Definition As a start, for this process mining and data
visualization project, exploitation of the domain knowledge from the process experts,
end-users, clients or administrative staff will be carried out to make a definition
of the project scope, goals, and questions related to these goals. The goals of a
project are related to improving KPIs related to cost, time, quality, etc. Associated
with these goals, performance-driven questions are used to determine how to assess
and achieve these goals. Each question can be answered with operational scenarios
including one or more process mining features measuring indicators of the process.
These indicators are defined to provide quantitative answers based on the specific
needs of stakeholders. The output of this stage is a document that contains a GQFI
(Goal-Question-Process Mining Feature-Performance Indicator) table.

Data Extraction In this stage, event data is extracted related to the patient for
process mining. Healthcare data attributes are extracted according to the questions
and objectives set. There are three minimum requirements for the control-flow anal-
ysis of healthcare date; activity id, information about activities, and timestamps of
these activities.
162 P. Rojas-Candio et al.

Fig. 1 Evaluation method of variables and indicators for the surgery block process using process
mining and data visualization

Data Preprocessing In order to solve data quality issues, such as incorrect, impre-
cise, missing, and irrelevant data, the extracted event data, according to the GQFI
table, must be formatted, filtered, abstracted, or clustered at different levels so that
the results are meaningful and valuable. To apply process mining techniques and
connect the event log to the model, event data obtained as a query must be expressed
in the form of event logs.
Registration and Pattern Inspection This stage refers to gaining a first impression
of the log and to gathering statistics to create a log and a pattern summary. This
inspection consists of the number of cases, events, duration of events, traces of
patterns, resources, frequencies and relative frequencies of events. The aim of this
stage is to fully understand and visualize the process quickly.
Process Mining Analysis Given a discovered or documented process model and
an event log related to that process, conformance checking techniques are used to
detect and verify inconsistencies related to business process compliance in aspects
of quality, time, resources, and costs.
Data Visualization Techniques Based on the event logs uploaded into a tool, it is
important to consider what is the most important information and how long is the
magnitude of your data to represent it graphically. Therefore, it is necessary to decide
the most important aspects and characteristics of the uploaded data.
Evaluation Method of Variables and Indicators … 163

Outcome Evaluation At this stage, results related to process mining indicators

and medical variables displayed in data visualization graphs are evaluated. After
each process mining activity, healthcare data can be divided into smaller sets. New
datasets can be investigated in terms of patient attributes, service, department, or
doctor information.

Process Improvement Approaches In process mining, the improvement is associ-

ated with the identification and elimination of bottlenecks, loop reprocessing, preven-
tion of unnecessary states, and understanding of activities that consume more than
average time.

4 Case Study: Validation

The method presented was validated in a Peruvian network of health centers. The
institution allowed us the implementation of pilot planning for the application and
validation of the proposed method. The data provided refers to the period from May
to August 2018 of a clinic located in Lima, Peru. For the validation of our proposal,
we first ran a run through the entire flow of activities of the method that compose it,
thanks to the support of the IT area of the health center, since they were responsible
for providing information about their process. Likewise, together with a specialist
from the surgery block, the indicators and variables related to process mining and
data visualization, respectively, were validated.

4.1 Data Extraction

Due to the confidentiality of data that the network of clinics is affected to fulfill, this
stage of the method was realized by the IT area of the company. For this, said area
extracted the information considering the minimum structure of an event log. They
consider a unique identification of each patient, the associated activities associated
with each patient from their reception to their departure, and the times for each
registered activity.

4.2 Data Preprocessing

In this activity, the information provided by the IT staff began to be refined because
some records were not with the complete data due to possible operational errors
generated by the medical team of surgery block. The records associated with this
incomplete data were deleted so as not to compromise the project result. The
164 P. Rojas-Candio et al.

minimum requirements of an event log were considered so that it can be loaded

and used in the Celonis tool.

4.3 Registration and Pattern Inspection

For this activity, the information processed in the Celonis tool is loaded. After the
information is loaded, this tool allows generating an overview of the process. For this
view, graphics are generated automatically to have the first impression of it. It was
observed that there was an average of 15 cases per day, with a total of 135 events per
day. Likewise, it was shown that there was an average of 11 h of execution between
the beginning and end of the process without considering the extreme values.
On the other hand, the bottlenecks were obtained that increased the execution time
considerably. It is recorded that between reception and preparation of the patient it
took 15 h on average. All this information obtained will be useful for review in the
following analysis activity.

4.4 Process Mining Analysis

Figure 2 shows the entire process by using the Celonis tool. This process begins from
the patient’s entrance to the health center and ends until his or her exit after passing
through a recovery due to the surgical operation received.
After having an overview of the process, we perform the discovery of the model
by executing the fuzzy miner algorithm in the system. In the first view, under 100%
of the connections, the spaghetti effect of the process was shown. Therefore, it was
decided, with the process specialist, to work with 99.1% of activities because of these
covered variants greater than 1%, while the others were below this percentage.
After the discovery of the model, the process mining feature that refers to the
conformity check was continued. For this task, a documented model provided by
the process specialist was loaded, and the comparison was made with all traces of
the event records. This is aimed at knowing the percentage of affinity between the
documented process and the loaded event logs. The result shown by the tool was that
there are 75% compliance and a total of 21 violations of the process.

4.5 Data Visualization Techniques

Figure 3 shows the representations that were used according to each defined variable.
It should be noted that four of the nine variables were represented quantitatively
because three of them refer to average values, while the remaining is simply the sum
of surgeries performed in the process. Regarding the other five variables that were
represented graphically, in the variable of surgical care per month, the bar graph was
chosen, because the dimension “month of activity” and the measurement “number
Evaluation Method of Variables and Indicators … 165

Fig. 2 Model of the surgery block process discovered in 99.1% of Celonis connections

of surgeries” are necessary without using a legend in that type of graph. Secondly, to
evaluate trends by operating room, it was decided to use the donut diagram because
it consists of seven operating rooms, and each one is represented on a different scale
due to the number of surgeries. Thirdly, to represent the amount of surgical care by
type of encounter, a ring diagram was used for the same reason presented previously.
Finally, to know the average time in which a surgical operation is performed by type
of encounter and by operating room, the bar chart was used as it helps us to easily
identify what type of encounter or operating room takes more or less time to perform
a surgery.
166 P. Rojas-Candio et al.

Fig. 3 Dynamic graphs made with the information of the surgery Block and using the defined
medical variables

4.6 Outcome Evaluation

In the inspection phase of records and patterns, a first impression was made of the
surgery block process with some indicators automatically generated by the Celonis
tool. This allowed us to have a first filter of the cases that we want to review. The
important thing to highlight is the following information: the process, from start to
finish, has an average duration of approximately 11 h without considering the extreme
values. Likewise, a bottleneck was highlighted between the activities of the patient
reception and patient preparation. That bottleneck had a value of 15 h, considering
the total of the event records.

4.7 Process Improvement Approaches

The improvement of adapting the documented model was identified in order to corre-
spond to the findings of the event records in the information systems. This will allow
the elimination of future observations in the following process mining projects. Like-
wise, the need to be able to count on a technology that facilitates the recording of
the start or end times of activity was identified. As a proposed solution in the face of
incomplete or regularized records of patient-related activities, it is recommended to
adopt Bluetooth-based beacon technology.
Evaluation Method of Variables and Indicators … 167

5 Final Remarks

A method of evaluating indicators and variables was used for the surgery block
process using process mining and data visualization. The method was executed
through operative scenarios in the surgical process in a network of clinics to answer
typical and frequent questions for this process. To perform this, 1710 cases were
reviewed with a total of 15,390 surgical encounters.
It is validated that the application of the method allowed us to identify bottlenecks,
variants, violations, and variances of the process. Also, this model improves the
documented process model regarding the records of events of the health center’s
information systems.
As future work, the development of a tool that supports all the process mining
features listed in the GQFI table to respond to all indicators related to those features
so that they are evaluated, and new improvements in the health process are proposed.

References

1. EsSalud Institution: http://www.essalud.gob.pe/nuestra-institucion/. Last accessed 01 Sept

2019
2. EsSalud’s Institutional Statistics: http://www.essalud.gob.pe/estadistica-institucional/. Last
accessed 01 Sept 2019
3. EsSalud’s Presentation to the Labor and Social Security Commission of the Congress of the
Republic. http://www.congreso.gob.pe/Docs/comisiones2018/Trabajo/files/04_09_2018_e
xposicion_presidenta_essalud_fiorela_molineli_situacion_financiera_de_essalud.pdf. Last
accessed 01 Sept 2019
4. EsSalud’s 2017 Annual Memorandum. http://www.essalud.gob.pe/downloads/memorias/mem
oria_2017.pdf. Last accessed 01 Sept 2019
5. EsSalud loses more than S/2,900 millions in a year. https://elcomercio.pe/economia/peru/ess
alud-pierde-s-2-900-millones-ano-noticia-616248. Last accessed 01 Sept 2019
6. Interventions to reduce waiting times for access to health services: synthesis of evidence.
http://www.essalud.gob.pe/ietsi/pdfs/direcc_invest_salud/6_intervenc_para_reduc_tiemp_e
spera_acceso.pdf. Last accessed 01 Sept 2019
7. Armas, J., Aguirre, S., Coronado, A., Evangelista, M.: Evaluation of operational process vari-
ables in healthcare using process mining and data visualization techniques. In: Proceedings of
the 17th LACCEI International Multi-conference for Engineering, Education and Technology.
Latin American and Caribbean Consortium of Engineering Institutions, Montego Bay (2019)
8. Gurgen Erdogan, T., Tarhan, A.: A goal-driven evaluation method based on process mining for
healthcare processes. Appl. Sci. 8(6), 894 (2018)
9. ProM 6.7: http://www.promtools.org/doku.php?id=prom67. Last accessed 01 Sept 2019
10. van der Aalst, W.: Process Mining: Data Science in Action, 2nd edn. Springer, Heidelberg
(2016)
11. Ailenei, I.: Process mining tools: a comparative analysis. Student thesis: Master. Eindhoven
University of Technology, Eindhoven, The Netherlands (2011)
12. Disco: https://fluxicon.com/disco/. Last accessed 01 Sept 2019
13. PwC’s Finance Digital Day. https://fdocuments.us/document/finance-digital-day-pwcdk-the-
pwcs-process-intelligence-is-an-approach.html. Last accessed 26 Jan 2020
14. Celonis Operation Guide. https://help.sap.com/rc/6163e388c6ce431a893758775e8dca41/4.3/
en-US/SAP-Process-Mining-by-Celonis-4.3_Operation-Guide-1.6.pdf. Last accessed 01 Sept
2019
168 P. Rojas-Candio et al.

15. R’Bigui, H., Cho, C.: The state-of-the-art of business process mining challenges. Int. J. Bus.
Process Integr. Manage. 8(4), 285–303 (2017)
A Technological Solution to Identify
the Level of Risk to Be Diagnosed
with Type 2 Diabetes Mellitus Using
Wearables

Daniela Nuñovero , Ernesto Rodríguez , Jimmy Armas ,

and Paola Gonzalez

Abstract This paper proposes a technological solution using a predictive analysis

model to identify and reduce the level of risk for type 2 diabetes mellitus (T2DM)
through a wearable device. Our proposal is based on previous models that use the
auto-classification algorithm together with the addition of new risk factors, which
provide a greater contribution to the results of the presumptive diagnosis of the user
who wants to check his level of risk. The purpose is the primary prevention of type
2 diabetes mellitus by a non-invasive method composed of the phases: (1) Capture
and storage of risk factors; (2) Predictive analysis model; (3) Presumptive results
and recommendations; and (4) Preventive treatment. The main contribution is in the
development of the proposed application.

Keywords Wearable · Primary prevention · Type 2 diabetes mellitus

1 Introduction

According to the International Diabetes Federation [1], is estimated that around 425
million people of the world, that is, 8.8% of adults between 20 and 79 years old, suffer
diabetes. It is important to mention that more than 50% of these people are not aware
that they suffer from this disease, avoiding the possibility of preventing harmful and

D. Nuñovero · E. Rodríguez (B) · J. Armas

Universidad Peruana de Ciencias Aplicadas (UPC), Av. Primavera 2399 Surco, Lima, Peru
e-mail: [email protected]
D. Nuñovero
e-mail: [email protected]
J. Armas
e-mail: [email protected]
P. Gonzalez
Dalhousie University, 6299 South Street, Halifax, Canada
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 169
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_17
170 D. Nuñovero et al.

expensive complications. In 2045, if there is not detection and early prevention of

this chronic disease, 629 million people could have diabetes [1]. In addition, in 2017,
it was estimated that 10.7% of deaths in the world were due to diabetes, exceeding
the percentage of other diseases such as tuberculosis and HIV/AIDS [1].
Diabetes is a chronic disease that occurs when a person has elevated blood glucose
levels, because their pancreas stops producing insulin or because the body does
not produce it effectively [1]. The predominant type of this disease in the world
population is type 2, which is characterized by the inadequate production of insulin
and the inability of the body to respond to this hormone, in other words, resistance
to it [2]. One of the particularities of type 2 is that it a chronic disease of slow
progression and difficult to determine the exact moment in which it appears [1].
In Peru, 3.3% of the population between 15 and over were diagnosed with diabetes
mellitus, where the percentage of people is higher in Metropolitan Lima and the rest
of the coast [3]. In addition, according to the National Institute of Health (INS)
indicated that 70% of Peruvian adults are suffering from obesity and overweight,
being young 42.4% [4]. This information means that the risk of people suffering
from this chronic disease is increasing.
For this reason, this paper proposes a technological solution that evaluates risk
factors, without the need for laboratory tests or other invasive procedures. From this
evaluation, through a predictive analysis model based on the inference of the data, it
is possible to be prone as a person can be diagnosed with type 2 diabetes mellitus.
From this information, the result is obtained with a percentage of certainty of the risk
that the person is suffering from. It is there, where an objective is defined with respect
to physical activity, which will be measured through a wearable device. Similarly,
the mobile application provides recommendations on how to maintain an adequate
diet to reduce the risk of contracting the disease.

2 Literature Review

2.1 Diabetes Prediction Techniques

In the previous study [5], a prediction model for the diagnosis of T2DM is presented,
where three types of algorithms to SAP predictive analytics tool have been compared,
in order to validate and determine which one has greater precision.
The first category, decision trees, includes the R-CNR tree algorithm. According to
SAP [6], the algorithm allows classifying observations into groups and predicting one
or more discrete variables based on other variables. Among the output options, a trend
or fill field can be obtained based on two types of analysis (classification and regres-
sion). On the other hand, the category of neural networks algorithm includes the R-
Nnet neural network. The most representative characteristic of this type is that it does
not follow a linear path, because is processed collectively, in parallel through nodes
(or neurons) [5]. According to the description provided by SAP [7], the algorithm
allows predicting, classifying, and recognizing statistical patterns through library
functions. Finally, the classification category includes the native auto-classification
algorithm. According to SAP [8], the algorithm is used for binary or categorical
classification. In addition, the algorithm detects the type of model and algorithm
A Technological Solution to Identify the Level … 171

used and that best suits based on the target variable that was selected. It also decides
whether the input should be continuous or categorical and determines the appropriate
classification for these variables. As a result, this allows reducing data preparation
and modeling activities.
Performed the tests [9], it is obtained that the “Auto Classification” algorithm has
91.7% accuracy and obtains greater positive results compared to the other evaluated
algorithms.
For this reason, it was decided to use it for the solution, in order to improve
precision and accuracy. The variables proposed in the previous study will continue
to be used and add new ones found in the literature review.

2.2 Technology Platforms

To choose the platform that best suits the solution, an analysis was carried out under
certain criteria defined by Forrester, who is a market researcher, in charge of advising
leading business and technology companies. Hence, the Microsoft Azure platform
suits very well the requirements and demands of the solution.

2.3 Wearable Device

For the solution, the wearable device “Mi band 4” was used, since it has the charac-
teristics that are required. In addition, this device is focused on capturing information
about the user’s physical activity.

3 Proposed Technology Solution

3.1 Description of the Solution

Using the elements described in the literature review, a technology solution is

proposed to identify the presumptive level of user predisposition to be diagnosed
with type 2 diabetes mellitus. This proposal solution focuses on two key users: The
user that will use in the future, whom we will call the patient; and the doctor.
The solution is governed by two stages: The first is related to the diagnosis of
the disease and the second focused on preventive treatment. First, the diagnostic
stage focuses on the data collection and storage phases, the execution of a predictive
analysis model to identify the presumptive level of patient risk, plus results and
recommendations delivery. Secondly, the preventive treatment stage, formed by the
monitoring and control phases followed by some feedback from the doctor. Figure 1
depicts the technological solution of the work.
172 D. Nuñovero et al.

Fig. 1 Proposed technology solution

3.2 Stages and Phases of the Solution

3.2.1 Diagnosis

• Risk Factors Collection and Storage. For this phase, in the future applications, the
patient will fill out a form in the application. The form contains Yes/No questions
and questions with predefined answers to identify the presence and values of
risk factors of the patient, which will later be sent via Web service to an Oracle
database.
• Predictive Analysis Model. Responses from the form will be automatically sent
to the SAP predictive analytics tool and analyzed using the auto-classification
algorithm, which will result in a presumptive diagnosis of type 2 diabetes mellitus
and stored automatically by the tool on an Oracle database.
• Results and Recommendations Delivery. The results of the predictive analysis
model will be displayed in the mobile application. After this, a suggestion will be
displayed about following a preventive treatment regularly monitored by a doctor
in order to reduce the level of predisposition in the medium term.

3.2.2 Preventive Treatment

• Monitoring and control. In this stage, the wearable device is connected to the
mobile application. The wearable, in the future applications, will send information
about the physical activity of the user to the mobile application using Bluetooth
Low Energy (BLE) technology. This information will be recorded in the database
so that at the end of the day, the risk factors information of the patient is up to date.
The patient will be able to see a historical evolution of their risk level through a
A Technological Solution to Identify the Level … 173

Fig. 2 Architecture of the proposed technology solution

dashboard in the mobile application, as well as a view with healthy food recipes
in order to reduce the patient’s risk.
• Feedback. The application is also developed so that a doctor can periodically
monitor the patient’s risk level evolution in order to provide constant feedback to
the patient without the need to schedule an appointment.

3.3 Proposed Architecture

The following architecture will support the proposed stages of the technology
solution. This architecture enables the interaction of server components hosted in
Microsoft Azure with the mobile application and the wearable device. The archi-
tecture shown in Fig. 2 shows how equipment can interact with the mobile app and
wearable device, which communicates with the database and predictive analytics
model using Web services, both hosted in Azure. It is important to mention that any
cloud platform allows deploying Web services that interact with any database that
can be used to support this solution.

4 Case Study

4.1 Implementation

It is essential that the wearable device is associated with the phone and that the mobile
is connected to the Internet. The dashboard of the solution is shown in Fig. 3.
174 D. Nuñovero et al.

Fig. 3 Case study

Risk Factors Collection and Storage For the implementation of the solution, it
was provided with the mobile application so that it could register and enter his data,
and then consider the data for the presumptive diagnosis.

Data Reception in Mobile Application The information collected will be stored

in a database.

Predictive Analysis Model After obtaining the respective information with the help
of the “SAP Predictive Analytics tool,” it will be analyzed with the auto-classification
algorithm, from which it yielded the results composed of a Boolean (1 or 0) value indi-
cating whether or not the user is prone to suffer from the disease, and the percentage
of certainty of the result.

Results and Recommendations Delivery The results of the predictive analysis

model will be shown in the mobile application, along with a recommendation
indicating to approach a doctor or specialist for the corresponding tests.

Preventive Treatment From the result, it will be possible to decrease the risk
shown at baseline. So, the physical activity was measured through the wearable
and encouraged to have a much healthier diet.

The goal of this recordings was a concept and preliminary study for later applica-
tions, since the research was designed to develop a computational framework, where
tests with several persons will be regulated by the respective ethical committee.
A Technological Solution to Identify the Level … 175

5 Conclusions

The solution presented serves its purpose by reducing the risk of developing T2DM.
We expect that users when using our solution allowed them to feel responsible for
their health, letting them notice that a change in their physical activity habits has a
great representative impact on their level of risk from T2DM.
As future work, having a historical database with information from people who
have previously performed the basal glucose test would be ideal for improving the
accuracy and effectiveness of the model results used by the proposed solution. Like-
wise, we agree that higher samples and time periods are needed on monitored users
in order to improve the quality of preventive treatment.

References

1. International Diabetes Federation (FID): IDF Diabetes Atlas, 8th edn. Bruselas (2017)
2. Organización Mundial de la Salud (OMS): Informe Mundial Sobre La Diabetes. Ginebra (2016)
3. Instituto Nacional de Estadística e Informática: Perú: Enfermedades No Transmisibles y
Transmisibles. Lima (2018)
4. Instituto Nacional de Salud: Cerca del 70% de adultos peruanos padecen de obesidad y
sobrepeso. https://web.ins.gob.pe/es/prensa/noticia/cerca-del-70-de-adultos-peruanos-padecen-
de-obesidad-y-sobrepeso. Last accessed 07 Oct 2019 (2019)
5. Barrios, O., Alberto, D., Infantes, V., Raphael, E., Aguirre, A., Alexander, J.: Predictive modeling
for presumptive diagnosis of type 2 diabetes mellitus based on symptomatic analysis. In: 2017
IEEE XXIV International Conference on Electronics, Electrical Engineering and Computing
(INTERCON), pp. 1–4. IEEE (2017)
6. SAP: SAP Help Portal: R-CNR Tree. Retrieved from https://help.sap.com/viewer/94dbf2
ba9d4047618880187451c3b253/3.2/es-ES/e790ec0a80b7101487253286b0e91070.html. Last
accessed 06 Oct 2019
7. SAP: SAP Help Portal: R-Nnet Neural Network. Retrieved from https://help.sap.com/viewer/
94dbf2ba9d4047618880187451c3b253/3.2/es-ES/e7911c7b80b7101487253286b0e91070.
html. Last accessed 06 Oct 2019
8. SAP: SAP Help Portal: Auto Classification. https://help.sap.com/viewer/94dbf2ba9d40476188
80187451c3b253/3.3/en-US/38ad5a8b54e84f37a545a197f7af3adf.html. Last accessed 06 Oct
2019
9. Ollila, M.-M. E., West, S., Keinänen-Kiukaanniemi, S., Jokelainen, J., Auvinen, J., Puukka, K.,
Järvelin, M.R., Tapanainen, J.S, Franks, S., Piltonen, T.T., Morin-Papunen, L.C.: Overweight
and obese but not normal weight women with PCOS are at increased risk of Type 2 diabetes
mellitus—a prospective, population-based cohort study. Hum. Reprod. 32(2), 423–431 (2016)
Mamey Apple Peel for Cr3+ Removal
from Contaminated Waters

Nathali Cotrina and Ricardo Vejarano

Abstract The aim of this study was to evaluate the capacity of an adsorbent based
on mamey apple (Mammea americana L.) peel for Cr3+ removal. Aqueous mediums
based on distilled water and different initial concentrations of Cr3+ , with a dose of
10 g of adsorbent per liter, were used. The residual concentrations of Cr3+ were
determined by atomic absorption spectrophotometry, obtaining optimum values of
pH of 2.9, the adsorbent particle size of 300 µm, and stirring rate of 300 rpm, for
Cr3+ removal. Kinetic studies indicate a chemical adsorption process, with the best
adjustment to pseudo-second-order models (R2 ≥ 0.9973), a contact time of 120 min
to reach the adsorbent-Cr3+ equilibrium, and a good fit to the Freundlich model (R2
= 0.9291), with an adsorption intensity n of 1.5876. So, our results suggest that Cr3+
ions can be efficiently removed by using mamey apple peel-adsorbent.

Keywords Chromium-contaminated waters · Adsorption · Mamey apple peel

1 Introduction

The leather tanning is an important industrial activity in Peru [1]. However, this
industry generates considerable volumes of effluents with high contents of chromium
(Cr), in many cases surpassing the maximum limits permitted by Peruvian legislation
[2]. Chromium cannot be degraded or removed easily, affecting the biodiversity of
soils, rivers, and lakes, killing living organisms present in those ecosystems [3]. Like
most metals, it can be transported into bodies of water, filtered through the soil, reach
drinking water sources or food crops, and ultimately be incorporated into the food

N. Cotrina
Ingeniería Ambiental, Universidad Privada del Norte, Trujillo, Peru
e-mail: [email protected]
R. Vejarano (B)
Dirección de Investigación y Desarrollo, Universidad Privada del Norte, Trujillo, Peru
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 177
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_18
178 N. Cotrina and R. Vejarano

chain and affect human beings with severe health consequences when is ingested at
high levels [3, 4].
The leather tanning industry is considered the major source of Cr, due most of
the tanneries adopt the Cr tanning process, where the leather takes up only 80% of
added Cr, and the rest is usually discharged into the sewage system [5], representing
a great environmental risk.
Chromium exists in several oxidation states. The tannery effluents mainly contain
Cr3+ , which under aerobic conditions oxidizes to Cr6+ [6]. The Cr6+ specie is the most
toxic, due to their higher solubility and mobility [5, 6]. In fact, previous studies report
that high concentrations of Cr6+ cause skin rashes, nose bleeds, respiratory problems,
suppressed immune system, liver and kidney diseases, diarrhea, cancer, among other
diseases [3, 4]. Besides, Cr6+ can cross the cell membrane, and be converted to Cr3+
into the cells, where may lead to subsequent damage of lipids, proteins, DNA, and
RNA molecules [7]. Therefore, it is needed to develop procedures focused on an
efficient Cr3+ removal from these effluents, and thus mitigate its impact on affected
ecosystems.
Treatments of industrial effluents with high levels of heavy metals based on chem-
ical precipitation and coagulation, ion exchange chemical treatments, sorption using
activated carbon, use of zeolites and organic resins, solvent extraction, or physical
processes as filtration, reverse osmosis, among others, have been applied in the last
years, which is evidenced by the studies published by different authors [3, 5, 6, 8].
However, these methods have shown some disadvantages, for example, the produc-
tion of huge amounts of sludge that must also be treated before final disposal. Thus,
these processes require a technological investment that, in many cases, make them
economically unfeasible [3]. So, there has been increased interest in the last years in
the use of low-cost technological alternatives.
The use of residual biomass as adsorbents, for example, the peel of some fruits
[3, 9–11], is shown as a promising alternative due to its availability, low cost, and
composition. Generally, these residues are underutilized and discarded. Previously
we reported the feasibility of fruit peels as adsorbents to remove heavy metals [9–11].
The main reason for their high metal affinity is the presence of polymers, mainly
pectin, in their cell walls.
Mamey Apple (Mammea americana L.) is a tropical fruit produced in Peru.
Mostly, this fruit is consumed as fresh fruit, and its main residue is the peel, which
contains pectin [12]. This polymer due to its content of hydroxyl (OH), carbonyl
(CO), carboxyl (COOH) is capable of fixing metal ions through the chemical adsorp-
tion mechanism [11, 13]. The literature reports the use of mamey apple husks (chem-
ically modified with formaldehyde) for Cr6+ removal [14]. Respect to Cr3+ removal,
adsorbents based on algae biomass [4], and activated carbon from sugar cane waste,
passion-fruit shell, and orange waste, have been used [5, 15, 16]. Likewise, we
reported the feasibility of the passion-fruit shell as an efficient adsorbent for Cr3+
removal [9]. Nevertheless, the use of mamey apple peel-based adsorbents for Cr3+
removal has not been reported yet.
In this sense, the aim of the study was to evaluate the feasibility of mamey apple
(Mammea americana L.) peel, as an adsorbent for Cr3+ removal from contaminated
aqueous mediums through the mechanism of chemical adsorption, as a potential
alternative to the treatments currently applied at industrial level.
Mamey Apple Peel for Cr3+ Removal … 179

2 Materials and Methods

2.1 Adsorbent Based on Mamey Apple Peel

Mamey Apple (Mammea americana L.) was purchased from the local market in the
city of Trujillo, Peru. Peels were washed with distilled water and cut into pieces of
1.0 cm × 1.0 cm. Then, they were dried at 80 °C for 24 h in a UN55 PLUS oven
(Memmert GmbH Co. KG, Germany) [17], and ground in a knife mill and sieved
through different sieves (Tyler series, 75–1180 µm) to standardize the size of the
particles. This material was stored and used directly without any later treatment.
The characteristics of this adsorbent were evaluated by measuring the point of
zero charges (pHPZC ), i.e., the pH in which the resultant net charge of the adsorbent
surface is zero [18]. Distilled water was taken in 100 ml beakers, adjusting the pH
between 2.0 and 6.0. These samples were added 0.5 g of adsorbent and after 12 h
under stirring the final pH value was measured. The PZC corresponds to the value
where the final net charge (pH) curve cuts the X-axis (pH = 0).

2.2 Adsorption Experiments

Optimum Removal Conditions Aqueous mediums based on distilled water and

Cr(OH)SO4 (commercial CUIREXTAN B33) were prepared. Three different sequen-
tial tests were performed in order to determine the optimum values of the parameters
showed in Table 1. Parameters as Cr3+ concentration 200 mg l−1 , adsorbent dose
10 g l−1 , temperature 25 °C, and stirring time 300 minutes, were fixed at constant
values.
In all cases, after the stirring period, the samples were filtered with Whatman
No. 42 filter paper, and the supernatant was mixed with 3 ml of HNO3 and 1 ml of
HCl. The mixture was heated at 140 °C in a DigiPREP Jr digestor (SCP Science,
Canada) [9, 10], and then, each digested sample was analyzed by Atomic Absorption
Spectophotometry (Agilent Technologies 200 Series AA, USA) [9]. All experiments
were conducted in triplicate.

Table 1 Assays performed in order to determine the optimum removal parameters

Test Parameter pH Adsorbent particle size Stirring rate (rpm)
(µm)
2.0–6.0 75–1180 200–500
1 pH – 300 300
2 Adsorbent particle size Optimum – 300
(µm)
3 Stirring rate (rpm) Optimum Optimum –
180 N. Cotrina and R. Vejarano

Adsorption Isotherms and Kinetics Study Once the optimum values of pH, adsor-
bent particle size, and stirring rate were defined, the adsorption process was carried
out using solutions of Cr3+ at 20, 50, 100, 250, and 400 ppm, with a fixed dose of 10 g
of adsorbent per liter in 500 ml beaker at 25 °C. Aliquots of 15 ml were extracted at
5, 10, 15, 30, 60, 90, and 120 min, and filtered to separate the supernatant, following
the same digestion procedure mentioned in the previous section. The amount of Cr3+
adsorbed per gram of adsorbent (q, mg/g) was calculated according to Eq. 1 [3]:

(C0 − C)
q=V (1)
m
where V is the solution volume (l), m is the adsorbent dose (g), and C 0 and C is the
initial and final Cr3+ concentrations, respectively (mg l−1 ).
In order to study the system behavior, the isotherm models of Langmuir and
Freundlich were fitted, judging the coefficients of determination (R2 ) in order to find
the applicability of the developed models. The Langmuir model suggests monolayer
adsorption on a homogenous surface. For its part, the Freundlich model explains
the interaction adsorbate-adsorbent with multilayer adsorption on heterogeneous
surfaces [19]. Additionally, in the case of the Langmuir model, the dimensionless
parameter R was calculated, which describes the type of isotherm, that according to
Weber and Chakraborti [20] is defined by the equation:

1
R = (2)
1+b.C0

Where b is a Langmuir constant, which indicates the nature of adsorption and

the shape of the isotherm. R -values <1 indicate that the adsorbent-adsorbate system
shows a favorable adsorption process [5].
Finally, to evaluate the adsorption kinetics, the same results were evaluated by the
mathematical models of pseudo-first and pseudo-second-order [21].

2.3 Statistical Analysis

Analysis of variance (ANOVA) and Fisher’s LSD (Least Significant Difference) test
at a 5% significance level (p < 0.05) were performed using Statistica 7.0 software
[9], in order to identify significant differences.
Mamey Apple Peel for Cr3+ Removal … 181

3 Results and Discussion

3.1 Point of Zero Charge (pHPZC ) and Effect of pH on Cr3+

Removal

From the chemical point of view, when the aqueous medium has a higher pH than
pHPCZ , negative charges predominate on the adsorbent surface [22]. Figure 1 shows
a pHPZC value of around 3.8 for adsorbent based on mamey apple peel.
In aqueous mediums, Cr3+ exists in several states as a function of pH. The form
Cr predominates at pH < 3.0, the forms Cr3+ and Cr(OH)2+ at pH 3.0–6.0, and
3+

the species Cr(OH)3 at pH 6.5–10.0 [5]. Figure 1 indicates a good Cr3+ removal
process at pH values greater than 3.8 since no statistically significant differences
(LSD Fisher, p < 0.05) were evidenced at higher pH values. In this region, Cr3+ is
adsorbed principally as Cr(OH)2+ , since the adsorbent surface becomes negatively
charged (carboxylic acids COOH of pectins in mamey apple peel) [12].
Previous studies reported a good removal process at pH values of 4.0, 5.0, 5.0, and
4.0, by using adsorbents based on algae biomass [4], and activated carbon from sugar
industrial waste [5], passion-fruit shell [15], and orange waste [16], respectively.
Taking in account that most tannery effluents in Peru, and in other countries, are
discharged at pH values between 3.0 and 4.0 into the sewage system [23–25], the
obtained pHPCZ value could be considered as an appropriate condition for the Cr3+
removal treatment.
However, at an industrial level, chemical precipitation is the usual way for Cr3+
removal, by using NaOH or Ca(OH)2 as precipitating agents (Eq. 3), with greater
efficiency as the pH value increases [26]. This phenomenon was verified in the present
study by using different values of pH (NaOH addition), in mediums with an initial
concentration of Cr3+ of 275 mg l−1 , in the absence of adsorbent (Fig. 2).
−
(aq) + 3OH(aq) → Cr(OH)3(s)
Cr3+ (3)

According to Fig. 2, at pH values greater than 2.9, Cr removal would be more

related to the presence of NaOH. Since at industrial level, chemical precipitation is
one of the most used methods, which has among its main disadvantages the high
sludge generation [3], in the present study the pH value of 2.9 was chosen for the
subsequent treatments, in order to evaluate the efficiency of the mamey apple peel-
based adsorbent, without the interference of the basifying/precipitating agent.
182 N. Cotrina and R. Vejarano

Point of zero charge (pHPZC)

1
ΔpH

0
0 1 2 3 4 5 6 7

-1

-2

(a)
-3

pHinitial

q (mg Cr3+/g adsorbent)

20 0.8

16
0.6

12

0.4

0.2
4

0 0.0
1 2 3 4 5 6 7

pH
(b)
200 ppm 10 ppm

Fig. 1 a pHPCZ for the apple mamey peel-based adsorbent and b the effect of different pH values
on Cr3+ removal (agitation speed 300 rpm)

3.2 Effect of Adsorbent Particle Size and Stirring Rate

on Cr3+ Removal

Since the adsorption process occurs inside the particles (on the pore walls), the
decrease in adsorbent particle size increases the internal surface area, and therefore,
the number of available binding sites for the contact adsorbent-adsorbate [19].
Mamey Apple Peel for Cr3+ Removal … 183

(a)
2.01 2.72 2.86 3.96 4.75 6.11

40

35

30
Cr removal (%)

25

20

15

10

5
1 2 3 4 5 6 7
(b)
pH

Fig. 2 a Presence of Cr(OH)3 and b the effect of different pH values on Cr3+ solubility in aqueous
mediums at 275 mg l−1 of Cr3+ , in absence of adsorbent

According to Fig. 3, higher Cr3+ removal rates (q) were evidenced at smaller
particle sizes [27], with significant differences (LSD Fisher) between 75 and 425 µm.
However, the addition of adsorbent at particle sizes smaller than 300 µm caused
sludges-like suspensions, making difficult the stirring and the supernatant separation.
Therefore, the 300 µm adsorbent was chosen for further Cr3+ removal treatments.
For its part, the stirring rate helps the adsorbate overcome the boundary layer
mass transfer resistance [28]. Figure 3 shows that agitation speeds of up to 300 rpm
favor high Cr3+ adsorption, in accordance with Pathak et al. [28], who report efficient
adsorption processes with agitation speeds between 200 and 300 rpm. Nevertheless,
higher rpm values (>300), the observed decrease in adsorption could be due to the
high mobility of Cr3+ ions in the solution, generating a weakening of the adsorptive
forces, and therefore, greater Cr3+ desorption in relation to the adsorbed Cr3+ .
184 N. Cotrina and R. Vejarano

14

q (mg Cr/g adsorbent) 12

10

0
0 200 400 600 800 1000 1200 (a)
µm (micras)

14

12
q (mg Cr/g adsorbent)

10

0
200 300 400 500 (b)
rpm

Fig. 3 a Effect of different adsorbent particle sizes and b stirring rate, on Cr3+ removal from
aqueous mediums (Cr3+ : 200 mg l−1 ) at 25 °C, pH 2.9, and contact time 120 min

3.3 Kinetics Study and Adsorption Isotherms

Table 2 shows kinetic parameters for the removal of Cr3+ . The best values of R2 were
obtained with the pseudo-second-order models. These results indicate a chemical
adsorption process between mamey apple peel and the Cr3+ ions [21]. Besides, in
this model, k 2 is the pseudo-second-order rate constant, and h is the initial sorption
rate.
The low values of k 2 at high concentrations of Cr3+ indicate a low number of
available sites for adsorption of Cr3+ ions, i.e., the surface of the mamey apple
peel-based adsorbent begins to saturate [19].
Table 2 Kinetics (pseudo-second-order models) for Cr3+ removal by mamey apple peel
Solution (ppm) Equation R2 k 2 (g mg−1 min−1 ) h (mg g−1 min−1 ) qe_exp. ( mg g−1 ) qe_calc. ( mg g−1 )
Mamey Apple Peel for Cr3+ Removal …

20 t/q = 0.6207t + 2.2002 0.9973 0.1898 0.4545 1.6111 1.5476

50 t/q = 0.2950t + 0.3919 0.9999 0.1278 1.4453 3.3898 3.3633
100 t/q = 0.1286t + 0.5115 0.9987 0.0343 1.9550 7.7760 7.5530
250 t/q = 0.0998t + 0.2779 0.9998 0.0349 3.5984 10.0200 10.1546
400 t/q = 0.0533t + 0.2139 0.9994 0.0137 4.6751 18.7617 18.4720
qe _exp.: qe _experimental ; qe _calc.: qe _calculated
185
186 N. Cotrina and R. Vejarano

Respect to equilibrium studies (Fig. 4), the capacity of adsorption of Cr3+ increases
while the initial concentration of Cr3+ increases from 20 up to 400 mg l−1 , a
phenomenon attributed to the competition for the available binding sites, in concor-
dance with kinetic parameter h (Table 2), as a consequence of the necessary strength
to overcome the resistance in the mass transfer of the chromium ions between the
medium and the adsorbent, provided by high metal concentrations [29].
Besides, a contact time of 120 min was necessary to reach the equilibrium
adsorbent-adsorbate, i.e., the mamey apple peel-based adsorbent is no longer able
to adsorb further Cr3+ from that moment, with qe values of 1.5476, 3.3633, 7.5530,
10.1546, and 18.4720 for solutions of 20, 50, 100, 250, and 400 mg l−1 of Cr3+
(Fig. 4).

Langmuir
R2 0.7978
qmax 21.7865
b 0.0117
20 ppm 0.7942
50 ppm 0.6429
R’ 100 ppm 0.4419
250 ppm 0.2599
400 ppm 0.1754
Freundlich
R2 0.9291 (a)
k 0.5866
n 1.5876

q (mg Cr/g adsorbent)

20

16

12

0
0 60 120 180 240 300

Time (min)

20 ppm 50 ppm 100 ppm 250 ppm 400 ppm (b)

Fig. 4 a Models for adsorption process of Cr3+ from aqueous mediums. Langmuir model: C e /qe
= 0.0459 C e + 3.9326; Freundlich model: Ln qe = 0.6299 Ln C e —0.5334. b Contact time to reach
the equilibrium adsorbent-adsorbate at different concentrations of Cr3+
Mamey Apple Peel for Cr3+ Removal … 187

Based on the coefficient of determination (R2 = 0.9291), the Freundlich model was
suitably fitted for Cr3+ adsorption (Fig. 4). In this model k and n constants indicate
the capacity and the intensity of adsorption, respectively [19]. The mamey apple
peel-based adsorbent showed a capacity of adsorption k of 0.5866. A close value
was obtained in a previous study for Cr3+ removal by adsorbent based on passion-
fruit shell [9]. Likewise, a good adsorption intensity (n = 1.5876) was obtained, so
it can be considered as a favorable adsorption process (n > 1, [30]).
Additionally, in the case of Langmuir model, the obtained values of the dimen-
sionless parameter R of 0.7942, 0.6429, 0.4419, 0.2599, and 0.1754 for solutions
of 20, 50, 100, 250, and 400 mg l−1 of Cr3+ indicate a favorable adsorption process
(R < 1 [5]).
These results indicate a high affinity between Cr3+ ions and the polymers present in
the mamey apple peel-based adsorbent, mainly pectin [12]. Pectin contains functional
groups as COOH, OH, and CO, which would be the sites to Cr3+ ions fixation [11, 13].

4 Conclusions

The adsorption capacity exhibited by the mamey apple peel-based adsorbent was
very well obtaining optimum values of pH 2.9, adsorbent particle size 300 µm, and
stirring rate of 300 rpm, for Cr3+ removal. Likewise, the results of kinetics studies
showed the best adjustment to pseudo-second-order models (R2 ≥ 0.9973), indicating
a chemical adsorption process, and a contact time of 120 min would be enough to
reach the adsorbent-Cr3+ equilibrium, showing the best adjustment to the Freundlich
model (R2 = 0.9291), with a favorable adsorption process, since the obtained value
of adsorption intensity (n), was 1.5876. So, our results suggest that Cr3+ ions can
be efficiently removed from contaminated waters by using mamey apple peel-based
adsorbent, which with a minimum treatment could be used as a low-cost adsorbent, as
a potential alternative to traditional methods nowadays used. Finally, further studies
are needed to improve the adsorption capacity of the mamey apple peel, such as
the evaluation of chemically modified mamey apple peel, in order to elucidate the
efficacy of an activated adsorbent.

Acknowledgements This study was funded by “Project 20171002: Removal of heavy metals from
contaminated waters by agroindustrial wastes” (Universidad Privada del Norte, Trujillo, Peru). The
authors thank Yuliana Pairazaman and Julio Gurreonero (Instrumental Analysis Laboratory, UPN,
Trujillo, Peru) for their technical assistance.

References

1. Vilar, V.J.P., Botelho, C.M.S., Boaventura, R.A.R.: Influence of pH, ionic strength and temper-
ature on lead biosorption by Gelidium and agar extraction algal waste. Process. Biochem.
40(10), 3267–3275 (2005). https://doi.org/10.1016/j.procbio.2005.03.023
2. Ministerio del Ambiente: http://www.minam.gob.pe/disposiciones/decreto-supremo-n-004-
2017-minam/. Last accessed 05 Jan 2020
188 N. Cotrina and R. Vejarano

3. Ali, A., Saeed, K., Mabood, F.: Removal of chromium (VI) from aqueous medium using
chemically modified banana peels as efficient low-cost adsorbent. Alex. Eng. J. 55, 2933–2942
(2016). https://doi.org/10.1016/j.aej.2016.05.011
4. Cazón, J., Benítez, L., Donati, E., Viera, M.: Biosorption of chromium(III) by two brown algae
Macrocystis pyrifera and Undaria pinnatifida: Equilibrium and kinetic study. Eng. Life Sci.
12(1), 95–103 (2012). https://doi.org/10.1002/elsc.201100098
5. Fahim, N., Barsoum, B., Eid, A., Khalil, M.: Removal of chromium(III) from tannery wastew-
ater using activated carbon from sugar industrial waste. J. Hazard. Mater. 136(2), 303–309
(2006). https://doi.org/10.1016/j.jhazmat.2005.12.014
6. Mohan, D., Pittman Jr., C.U.: Activated carbons and low cost adsorbents for remediation of
tri- and hexavalent chromium from water. J. Hazard. Mater. 137, 762–811 (2006). https://doi.
org/10.1016/j.jhazmat.2006.06.060
7. Shrivastava, R., Upreti, R.K., Seth, P.K., Chaturvedi, U.C.: The effects of chromium on the
immune system. FEMS Immunol. Med. Microbiol. 34(1), 1–7 (2002). https://doi.org/10.1111/
j.1574-695X.2002.tb00596.x
8. Gupta, V.K., Kumar, R., Nayak, A., Saleh, T.A., Barakat, M.A.: Adsorptive removal of dyes
from aqueous solution onto carbon nanotubes: a review. Adv. Colloid Interface Sci. 193–194,
24–34 (2013). https://doi.org/10.1016/j.cis.2013.03.003
9. Campos-Flores, G., Castillo-Herrera, A., Gurreonero-Fernández, J., Obeso-Obando, A., Díaz-
Silva, V., Vejarano, R.: Adsorbent material based on passion-fruit wastes to remove lead (Pb),
chromium (Cr) and copper (Cu) from metal-contaminated waters. In: AIP Conference Proceed-
ings, pp. 020079-1–020079-4. AIP Publishing, Melville NY (2018). https://doi.org/10.1063/
1.5032041
10. Vejarano, R., Gurreonero-Fernández, J., Castillo-Herrera, A.: Adsorption of lead (Pb) from
contaminated aqueous mediums using banana (Musa paradisiaca) peel. In: Proceedings of the
LACCEI International Multi-conference for Engineering, Education and Technology, p. 67.
LACCEI Inc., Boca Raton (2018). https://doi.org/10.18687/laccei2018.1.1.67
11. Chao, H.P., Chang, C.C., Nieva, A.: Biosorption of heavy metals on Citrus maxima peel, passion
fruit shell, and sugarcane bagasse in a fixed-bed column. J. Ind. Eng. Chem. 20(5), 3408–3414
(2014). https://doi.org/10.1016/j.jiec.2013.12.027
12. Yahia, E., Guttierrez-Orozco, F.: Mamey apple (Mammea americana L.). In: Postharvest
Biology and Technology of Tropical and Subtropical Fruits. Woodhead Publishing Limited,
Cambridge, pp 474–482 (2011). https://doi.org/10.1533/9780857092885.474
13. Demirbas, A.: Heavy metal adsorption onto agro-based waste materials: a review. J. Hazard.
Mater. 157(2–3), 220–229 (2008). https://doi.org/10.1016/j.jhazmat.2008.01.024
14. Serrano-Gómez, J., Beltrán, C., Bonifacio-Martínez, J., Olguín, M.: Mamey (Mammea amer-
icana L.) husks for the removal of Cr(VI) from aqueous media. Desalin. Water Treat. 74,
207–215 (2017). https://doi.org/10.5004/dwt.2017.20574
15. Jacques, R., Lima, E., Dias, S., Mazzocato, A., Pavan, F.: Yellow passion-fruit shell as biosor-
bent to remove Cr(III) and Pb(II) from aqueous solution. Sep. Purif. Technol. 57(1), 193–198
(2007). https://doi.org/10.1016/j.seppur.2007.01.018
16. Marín, A., Ortuño, J., Aguilar, M., Meseguer, V., Sáez, J., Lloréns, M.: Use of chemical modi-
fication to determine the binding of Cd(II), Zn(II) and Cr(III) ions by orange waste. Biochem.
Eng. J. 53(1), 2–6 (2010). https://doi.org/10.1016/j.bej.2008.12.010
17. Cruz-Tirado, J.P., Vejarano, R., Tapia-Blácido, D., Barraza-Jáuregui, G., Siche, R.: Biodegrad-
able foam tray based on starches isolated from different Peruvian species. Int. J. Biol. Macromol.
125, 800–807 (2019). https://doi.org/10.1016/j.ijbiomac.2018.12.111
18. Mimura, A.M.S., Vieira, T.V.A., Martelli, P.B., Gorgulho, H.F.: Utilization of rice husk to
remove Cu2+ , Al3+ , Ni2+ and Zn2+ from wastewater. Quím. Nova 33(6), 1279–1284 (2010).
https://doi.org/10.1590/S0100-40422010000600012
19. Tejada-Tovar, C., Villabona-Ortiz, A., Garcés-Jaraba, L.: Adsorption of heavy metals in waste
water using biological materials. TecnoL 18(34), 109–123 (2015)
20. Weber, T., Chakraborti, R.: Pore and solid diffusion models for fixed bed adsorbers. AIChE J.
20(2), 228–238 (1974). https://doi.org/10.1002/aic.690200204
Mamey Apple Peel for Cr3+ Removal … 189

21. Ho, Y., Mckay, G.: Pseudo-second order model for sorption processes. Process. Biochem.
34(5), 451–465 (1999). https://doi.org/10.1016/S0032-9592(98)00112-5
22. Mohd Salleh, M., Mahmoud, D., Abdul Karim, W., Idris, A.: Cationic and anionic dye adsorp-
tion by agricultural solid wastes: a comprehensive review. Desalination 280(1–3), 1–13 (2011).
https://doi.org/10.1016/j.desal.2011.07.019
23. Cooman, K., Gajardo, M., Nieto, J., Bornhardt, C., Vida, G.: Tannery wastewater characteri-
zation and toxicity effects on Daphnia spp. Environ. Toxicol. 18(1), 45–51 (2003). https://doi.
org/10.1002/tox.10094
24. Lofrano, G., Meriç, S., Zengin, G., Orhon, D.: Chemical and biological treatment technologies
for leather tannery chemicals and wastewaters: a review. Sci. Total Environ. 461–462, 265–281
(2013). https://doi.org/10.1016/j.scitotenv.2013.05.004
25. Córdova, H., Vargas, R., Cesare, M., Flores, L., Visitación, L.: Tratamiento de las aguas resid-
uales del proceso de curtido tradicional y alternativo que utiliza acomplejantes de cromo. Rev.
Soc. Quím. Perú 80(3), 183–191 (2014)
26. Esmaeili, A., Mesdaghi nia, A., Vazirinejad, R.: Chromium (III) removal and recovery from
tannery wastewater by precipitation process. Am. J. Appl. Sci. 2(10), 1471–1473 (2005). https://
doi.org/10.3844/ajassp.2005.1471.1473
27. Mishra, V., Balomajumder, C., Agarwal, V.: Biosorption of Zn (II) onto the surface of non-
living biomasses: a comparative study of adsorbent particle size and removal capacity of three
different biomasses. Water Air Soil Pollut. 211(1–4), 489–500 (2010). https://doi.org/10.1007/
s11270-009-0317-0
28. Pathak, D., Sachin, A., Bhaskar, D.: Fruit peel waste as a novel low-cost bio adsorbent. Rev.
Chem. Eng. 31(4), 361–381 (2015). https://doi.org/10.1515/revce-2014-0041
29. Llhan, M., Nourbakhsh, S., Kilicarslan, S., Ozdag, H.: Removal of chromium, lead and copper
from industrial waste by Staphylococcus saprophyticus. Turkish J. Biotechnol. 2, 50–57 (2004)
30. Masel, R.: Principles of Adsorption and Reaction on Solid Surfaces. Wiley, Hoboken, NY
(1996)
Method of Estimating River Levels
with Reflective Tapes Using Artificial
Vision Techniques

Lidia E. López Huamán , Marco Paul E. Apolinario Lainez ,

and Samuel G. Huamán Bustamante

Abstract Water level measurement in a river is a necessary step for flood prevention.
Early recognition of this factor could reduce the vulnerability of the population in the
surroundings. In this work, we use frames from videos as images to obtain water level
measurement indirectly. We applied digital image processing techniques over these
images to perform segmentation, edge detection, and we also applied multiple view
geometry techniques as projective transformation in a plane. The proposed method
estimates water level in specific locations where it is possible to install a reflective
tape with a two-color pattern, used as an indicator of water level. We obtained the
segmentation of the section out of the water of the reflective tape and so we get the
estimation of water level. Through the relation of the real distances (in centimeters)
that have four collinear points, seen from a perpendicular view, with the distances
(in pixel unit) of those same points contained in an image, which has undergone a
projective transformation. We made testing in a water tank built to this work and the
tests produced a percentage error at the range of 0.01–2.06% to a distance of 1.5 m
from a wall and 2 m high (location of camera).

Keywords Water level measurement · Image segmentation · Projective geometry

1 Introduction

Rivers are among the most important hydrographic reliefs present on the coast of
Perú. In recent years, floods and river overflow, especially occurring in seasons

L. E. López Huamán (B) · M. P. E. Apolinario Lainez · S. G. Huamán Bustamante

Instituto Nacional de Investigación y Capacitación de Telecomunicaciones, Universidad Nacional
de Ingeniería, Lima 15021, Peru
e-mail: [email protected]
M. P. E. Apolinario Lainez
e-mail: [email protected]
S. G. Huamán Bustamante
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 191
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_19
192 L. E. López Huamán et al.

summer when the oceanic and climatic phenomenon “El Niño” manifest more
strongly, have been increasing and causing more damage than expected in the
surrounding areas, such as those caused by the Rímac River (see Fig. 1c). For this
reason, the continuous monitoring of water level variation is essential.
Several techniques for measuring the level of water exist, but some are more
accurate than others, in proportion to the investment of costs of equipment. Another
important consideration is the selection of measurement points of a river basin.
Normally, this task is in charge of the National Meteorological and Hydrological
Service of Peru (SENAMHI) and the National Authority of Water (ANA). The auto-
matic hydrological stations are used for measuring the level of rivers and rainfall,
using pressure, bubble and radar meters (see Fig. 1a, b); the latter being the most
popular [1, 2].
In this work, we focus on water level monitoring using artificial vision techniques,
which are based on digital image processing for the acquisition of characteristics and
interpretation of them, on the images acquired by an RGB video surveillance camera.
The visual stability of the camera is essential for level measurement, so it is necessary
to install it in a high area of the riverbank.
The objective of this work is to propose a method for the indirect measurement of
the river level, using simple elements, such as a camera and a measurement indicator
(see Fig. 1d) that has a color pattern for the differentiation of the variation of the water
level. Then, we used computer vision techniques to perform a segmentation indicator
as an object for the projection of its geometric characteristics and thus estimate the
value of the water level.

Fig. 1 Area for which the study method is proposed. a Risk map of the sector in Chosica [11].
b Radar meter nearest and c limiter, in the area of the Rimac River in Chosica, San Juan de
Lurigancho, Lima—Perú. d Proposal for the river level indicator based on reflective tape
Method of Estimating River Levels with Reflective Tapes … 193

2 Methodology

2.1 Water Level Estimation System Requirements

As the main point, we have to consider the use of a water level indicator. In this
study, we have chosen a reflective tape with a red and white color pattern as the
indicator. This tape is commonly used to make objects or vehicles visible to reflec-
tions of low-intensity light sources, so its use is quite wide. In this case, the color
and light refraction properties are critical because it can be easily seen at night,
allow us continuous monitoring with the help of an external source of illumination;
thus, reflective tapes become an alternative regarding the use of Near-Infrared (NIR)
cameras [3].
On the other hand, the location of the acquisition system, especially the camera,
must be installed in a high and fixed area [4] (in the study area of the Rimac River,
a height of four meters must be exceeded in order to visualize the other bank). The
camera view is in the direction of the other end of the riverbank, with a view capable
of capturing from the deepest area of the river to the top of the level indicator. These
conditions are taken into account when testing in a controlled environment as a water
tank. If it is desired to have monitoring when the light is low, it will be necessary
to implement a lighting system (flash or infrared) to take advantage of the reflective
properties of the water level indicator.

2.2 Processing Stages and Estimation of Water Level

2.2.1 Image Acquisition

For this study, the acquisition of images was done previously using a digital camera
NEX-7 SONY [5] with which video samples were taken during the time that the filling
of the conditioned room for the sampling of the simulated water level variation took.
Subsequently, the processing of the videos obtained was analyzed to select the
frames where the level variation was notable; the images are presented in the
results stage. An algorithm was implemented to visualize the values obtained from
measurement using the MATLAB software installed in a laptop that was used.

2.2.2 Pre-processing and Segmentation for Water Level Measurement

Indicator

At this point, we followed the steps presented in the scheme of Fig. 2d. From the
images obtained from a video, a cut of the area of interest [6, 7] will be made that
contains the measurement indicator (reflective tape) located in the most uniform and
stable area of the river. First, an analysis was made of the different color models that
194 L. E. López Huamán et al.

can be obtained from an image, such as RGB, HSV, YCbCr, L * a * b and NTSC.
The RGB model allows us to differentiate the colors of the tape, to highlight in one
case the red color was raised by SR, see the Eq. (1), and in the white case SB, see
the Eq. (2). It is necessary to determine the thresholds and the area of the submerged
segments over the total of the analyzed area, [8].

SR = R1−G1 (1)

SB = R1/(255 − B1 + 1) (2)

The objects out of the water will be displayed later on the determination of the
threshold. To reduce the number of pixels in resulting images, which form smaller
objects and do not belong to the white and red objects of the indicator, is performing
dilation, erosion, filling of internal pixels and deletion of unconnected objects. With
this, we obtained better binary segmentation of the water level indicator. Then, we
apply a technique for edge detection [9] and find the coordinates, on the new object
detected, of its centroid and a point located on each side of the edge that is collinear
with the centroids of the red and white segments. This procedure is applied to each
of the objects with the highest number of pixels present in the masks of the red and
white colors, obtained separately.
In total, four points are imported, among these are the centroid of the object of
the largest white segment that is one closest to the chamber (point C  ), the centroid
of the object of the red segment (point B ), the points located on the lower (D ) and
upper (A ) points of the river level indicator. The objects containing B and C  points
are imported because those have the least deformation of pixels in the image.

Fig. 2 a Scheme of the proposed methodology. b Each set of points (ABCD) is related to the
others (A B C  D ) by a line-to-line projectivity since the cross-ratio (CR) is an invariant under a
projectivity. c Location of four points on the reflective tape and d in the projective transformation
of its image
Method of Estimating River Levels with Reflective Tapes … 195

According to the concept of the cross-relationship invariance under projective

transformation and single-view metrology [10]. The four points A, B, C, and D (see
Fig. 2) remain invariant in an image and their distances have the relationship as
shown in Eq. (3) which is the function CR (cross-radio).

CR(A, B, C, D) = (AC/AD)/(BC/B D)
   
= A C  /A D  / B  C  /B  D  (3)

From the above equation A, B, C and D are points located on the referenced object
so AC, AD, BC, and BD would become the measured lengths in centimeters between
those points. Likewise, the points defined as A , B , C and D , are the projections of
A, B, C and D, which are found in the image and the lengths A C  , A D , B C  , B D
are in pixel unit. D is a point on the surface of the water (which varies according to
the water level). On the tape, the length of each segment, whether red or white, is
equal and we define half of this value as the variable S_r and the length between the
C and D points as the variable xx (in centimeters).
Then, considering that the water level will always be bellowing the first red–white
segments, AC will be 3 S_r, AD will be equal to 3 S_r + xx. The length of the AD
segment (segment length not submerged) will vary inversely proportional to the water
level and will depend on the values of xx and S_r.
Finally, replacing the known values in Eq. (3), where the values are in centimeters
for the segments AC and AD. The BC segment is equal to 2 S_r, BD is equal to 2 S_r
+ xx. For the second part of the Eq. (3), whose variables are in a pixel unit, we will
call it Rel_pixel as seen in the Eq. (4).

(3S_r/3S_r + x x)/(2S_r/2S_r + x x) = Rel_pixel (4)

We transform the Eq. (4) to solve for a variable xx, so we got the Eq. (5).

x x = 6S_r (1 − Rel_pixel)/(2 Rel_pixel − 3) (5)

Once the value of xx has been calculated, we can determine the length of the
segment AD in centimeters (Dt) and, likewise, the height of the water level knowing
the length of the belt (L), as shown in Eqs. (6) and (7).

Dt = x x + 3 S_r (6)

Water_level = L − Dt (7)
196 L. E. López Huamán et al.

3 Results and Discussion

As mentioned previously, the measurements were obtained in a controlled enclosure

or water tank (see Fig. 3), during the filling of the water tank. The water level fills
out about 90 cm for a period of an hour where we tried to simulate the variation
in the level of water that can happen in a river. The reflective tape was located on
one of the tank walls and then a measuring tape was used to manually register the
water level and its comparison with the estimated data obtained by the developed
algorithm. After this, the camera was installed at a certain height and distance (it is
important to consider the FOV and camera-resolution) where the length of the level
indicator can be displayed almost entirely. All these considerations must be taken
into account when acquiring the images and installing the system.
Once the images were obtained and saved in a computer, we continued with the
digital processing, starting with cropping the area of interest at the length of the tape
(with red–black pattern) as a resource to eliminate different objects that they do not
matter in our analysis.
On the new image, the determination of the thresholds is to differentiate the red
and white areas, with a threshold equal to 70, for segmentation on the red sections
present on the surface of the water. In the case of white, there was a greater difficulty
in determining the threshold, since the bottom used in the tank wall was also white.
Given this, an analysis of the histogram more focused on the area when the water
surface divides a segment level indicator, as shown in Fig. 4, obtaining a threshold
equal to 2.54.

Fig. 3 a Water tank for measurement level water by the NEX7 SONY camera (in INICTEL-UNI).
b Camera view of reflective tape and c tape measure
Method of Estimating River Levels with Reflective Tapes … 197

Fig. 4 Segmentation a, c of the tape obtained after the histogram analysis b, d to obtain the threshold
of white tape segment above the water surface, obtained at a different level of water

Better segmentation of both colors was obtained, using dilation and erosion tech-
niques for the morphological treatment of their masks, which were then integrated
to obtain a mask that contains the entire area of interest (region of the tape on the
surface of the water), where point D was obtained. For both the red and white masks,
the centroids of the object with the highest number of pixels were extracted, these
being pointed B and C  respectively, present in the Eq. (3). Likewise, the coordinates
of point A are analyzed by analyzing the object with the greatest dimension on the
mask obtained in the red color, and that is how it was possible to obtain the points
of interest A , B , C  and D necessary to estimate the distance in pixel unit at which
the highest water level is.
It was necessary to know the real distance values between these four points; in
our case, the measure of the length of each color segment was 15 cm. Some of the
procedures described in the methodology to obtain a mask of the segment of the
indicator that is located on the water surface of the acquired images are shown in
Fig. 5a.
The accuracy of the water level measurement depends on the resolution of the
image contained in the reference indicator and minimizing the deformation effect of
the pixels present in the image.
The segmentation performance in this type of indicator was evaluated with the
percentage of error presented with the measurement for 47 frames obtained from a
video of the monitoring of the variation of the water level in the controlled enclosure.
Some of the images and values are shown in Fig. 5.b. Figure 6 shows the experimental
results compared with those obtained by us manually, the level of water compared
to the estimated real value of different images taken at different times.
198 L. E. López Huamán et al.

Fig. 5 Values obtained from different water levels

Fig. 6 Curves with values obtained water level when filling a water tank. Level estimation with the
proposed method (dashed line), compared with the actual value and measured manually (orange
line)

The average error rate of 0.13% was obtained, being the highest error value 2.06%,
equivalent to 1.5 cm, and the lowest of 0.01%, equivalent to 0.01 cm, in 47 frames
acquired 1.5 m away from the tape and 2 m high with respect to the background.
A video provides frames converted to images for processing by the method
explained above. The values are shown in Fig. 7 were obtained with 5000 images
that were processed in the MATLAB environment. The graph shows the increase in
water level and some variations that are within the average error rate of the proposed
method for estimating the water level.
Method of Estimating River Levels with Reflective Tapes … 199

Fig. 7 Level water values obtained from 5000 images

4 Conclusion

In this work, it is demonstrated that a reference indicator with a red–white pattern

of colors provides sufficient information to estimate the water level from the RGB
images. It is worth saying that to find out a suitable threshold is necessary to properly
differentiate the non-submerged area of the indicator.
Additionally, the proposed method is capable to estimate the water level in a
water tank, with which a percentage error obtained is in the range of 0.01–2.06%,
equivalent to the range of 0.01–1.5 cm, acquired at 2 m away from the belt and 2 m
high from the bottom. That percentage of error is low; however, it increases with
distance. This is mainly due to the decrease in the resolution of images contained in
the reference tape.

Acknowledgements The primary support for this project was from INICTEL-UNI, which provided
the necessary equipment to carry out this study. We would like especially to thank the team of Signal
and Image Processing (G-PSI) of Research and Technological Development Department (DIDT)
for their valuable suggestions during this process.

References

1. Jia, L., Xue, B., Chen, S., et al.: A high-resolution ultrasonic ranging system using laser sensing
and a cross-correlation method. Appl. Sci. 9, 1483 (2019). https://doi.org/10.3390/app9071483
2. Kato, A., Sinde, R.S., Shubi, K.: Design of an automated river water level monitoring system
by using global system for mobile communications. Int. J. Comput. Sci. Syst. Anal. 13
3. Zhang, Z., Liu, Z., Wang, H.: Visual measurement of water level under complex illumination
conditions. Sensors 19(19), 4141 (2019). https://doi.org/10.3390/s19194141
4. Zhang, Zhen, Zhou, Yang, Liu, Haiyun, Gao, Hongmin: In-situ water level measurement using
NIR-imaging video camera. Flow Measur. Instrum. 67, 95–106 (2019)
200 L. E. López Huamán et al.

5. SONY (2011) Equipment camera model Sony NEX-7 SEL-1855. https://www.sony.es/electr

onics/support/e-mount-body-nex-7-series/nex-7/specifications. Accessed 1 July 2019
6. Yu, J., Hahn, H.: Remote detection and monitoring of a water level using narrow band channel.
J. Inf. Sci. Eng. 26, 71–82 (2010)
7. Lin, Y.-T., Lin, Y.-C., Han, J.-Y.: Automatic water-level detection using single-camera images
with varied poses. Measurement 127, 167–174 (2018). https://doi.org/10.1016/j.measurement.
2018.05.100
8. Hasan, I.: An effective camera based water level recording technology for flood monitoring.
In: INTERPRAEVENT Conference Proceedings, pp. 290–295 (2016)
9. Solomon, C., Breckon, T.: Fundamentals of Digital Image Processing a Practical approach with
examples in Matlab. pp. 63–76 (2011)
10. Hartley, R., Zisserman, A.: Multiple View Geometry in Computer Vision, 2nd edn, pp. 44–46.
Cambridge University Press, Cambridge (2003)
11. El comercio (2015). Chosica: no existen estudios de las quebradas de riesgo. https://elcomercio.
pe/lima/chosica-existen-estudios-quebradas-riesgo-356568-noticia/?ref=ecr. Accessed Nov
2019
Design and Implementation
of a Hydro-Energetic System in Water
Distribution Networks

Lenin Jiménez , Cecilia Luna , Manuel Quiñones-Cuenca ,

and Holger Benavides

Abstract This thesis details the design and implementation of a prototype for hydro-
energetic use in water distribution networks. The prototype is composed of two main
systems. On one hand, the generation system, which is in charge of transforming the
kinetic energy of water into electrical energy using microturbines, charge regulators
and batteries. The second one is the drive and telemetry system. It controls whether
the generation units turn on or off as well as allowing remote monitoring of electrical
parameters such as power, voltage and current of each unit. The device that monitors
the generation units is connected to the charge controller through an RS232 interface
and enables acquiring the desired electrical parameters to later store them in an
Internet of Things platform to visualize them through the Internet.

Keywords Drive and telemetry system · Electrical energy · IoT · Microturbines

1 Introduction

Nowadays, wireless sensor networks (WSNs) with a large number of sensor nodes
allow interaction with the physical world, which is why they are widely used in
industrial applications and also in monitoring physical and environmental variables

L. Jiménez (B) · C. Luna · M. Quiñones-Cuenca

Departamento de Ciencias de la Computación y Electrónica, San Cayetano Alto, Loja, Ecuador
e-mail: [email protected]
C. Luna
e-mail: [email protected]
M. Quiñones-Cuenca
e-mail: [email protected]
H. Benavides
Departamento de Geología y Minas e Ingeniería Civil, Universidad Técnica Particular de Loja,
San Cayetano Alto, Loja, Ecuador
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 201
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_20
202 L. Jiménez et al.

[1]. Like many technological advances, the WSNs had their origin in military appli-
cations. The Sound Surveillance System (SOSUS) was the first WSN application
developed by the United States Militia in 1950 to detect and track Soviet submarines
[2].
In the WSNs, the sensor nodes are responsible for capturing data and transmitting
it over the network. A large amount of captured data and the process of transmission
implies a continuous operation of each of the sensors [3]. These are powered by
batteries that hardly provide long-term operation and need to be replaced periodically,
making constant maintenance of the network nodes necessary [4].
Researchers have explored multiple alternative forms of energy acquisition to
power wireless sensor networks such as solar panels, electromagnetic radiation, water
or wind flows [5]. That is why the energy harvesting has emerged as a technique that
allows extending the lifetime of the sensors by making use of the available energy
in the environment to keep the nodes in constant operation within the network [1].
Among many sources of energy, water flow has been used to power WSN nodes [6]
using small hydro-generators placed in irrigation systems or water pipes [7, 8].
The use of a TRD microturbine, whose operation depends on a control system,
has been proposed to take advantage of water flow energy. The amount of energy
that microturbines can contribute is directly related to the pressure of the circulating
water [9].
The control system associated with the microturbine allows its continuous status
observation, this being of great importance to determine if it is supplying enough
energy and if its operation is optimal. So, it is convenient to store this information in
an Internet of Things (IoT) platform to remotely monitor the system.
In this context, an energy generation system designed and built to power telemetry
systems, as well as the control and monitoring system previously mentioned.

2 Methodology

This work aims to develop an autonomous system that uses TRD microturbines in
hydraulic installations to power a node of a wireless sensor network that monitors
water distribution network variables.
First, the TRD microturbine performance was characterized using a hydraulic
bench and a pressure-break tank (generation system) to contrast with the parameters
established in the manufacturer’s datasheet. From this, it was determined what is the
amount of energy generated by the microturbine as a function of the flow rate and
the pressure that passes through the pipe.
Then, the energy consumption of the central node of the “SMART WATER
NETWORK” project was obtained in each of the processes that it performs (active,
reception and transmission of data) to determine the battery that will be used to keep
the node active 24 h/day.
Design and Implementation of a Hydro-Energetic System … 203

Next, the telemetry and drive system of the generation units was designed and
built. This system obtains electrical parameters (voltage, current, power) of the micro-
turbines and sends this information to an IoT platform. First, the design was carried
out at hardware level, establishing the necessary components to obtain the electrical
parameters of the generation units and then the algorithm to be used was developed.
Finally, the generation system and the telemetry and drive system were integrated
into the UTPL Hydraulics Laboratory to observe the operation of these two systems
as a whole.

3 The Electrical Energy Generation System

3.1 TRD Microturbines Energy Efficiency

Table 1 shows the materials and equipment used in the project:

To characterize the microturbines, it was necessary to use a hydraulic bench. The
hydraulic bench was built in the UTPL Hydraulics Laboratory. A diagram of the
arrangement of the materials used for characterization within the hydraulic bench is
shown in Fig. 1.
In Fig. 1, two instruments for measuring hydraulic parameters are identified: flow
sensor and pressure gauge, which are used in order to know the pressure and flow
that reaches the turbine. There is also a pump that is responsible for sucking the water
located in a tank to push it toward the hydraulic circuit. In addition to this, a load is
connected to the turbine generator’s output.

Table 1 Equipment and

Detail Quantity
materials
TDR microturbine 5
Smart charger 5
Electrovalve 5
Phoenix inverter 5
12 V Bosch battery 5

Hydraulic bench
Flow sensor

Load

Hydraulic pump Manometer

Microturbine

Fig. 1 Microturbines characterization scheme

204 L. Jiménez et al.

Table 2 Obtained pressure

Pressure (bar) Flow (l/min) Hydraulic power (W)
and flow rate values
1 6.98 11.63
1.5 8.23 20.49
2 9.3 30.88
2.5 10.26 42.58
3 11.11 55.33
3.5 12 69.72
4 12.72 84.46
4.5 13.38 99.95
5 14.08 116.86

Considering the characterization made by the manufacturer of microturbines,

certain pressure values are taken as a reference to measure: voltage, current and flow
rate. Tables 2 and 3 show the results obtained regarding the variables of pressure,
flow rate, and power generated by each of the microturbines, respectively.
From the previous results, the performance curve of each microturbine is obtained.
Now we contrast the performance curve of each microturbine with the information
specified by the manufacturer in the datasheet. The performance curves obtained for
the five turbines and the one specified in datasheets from the manufacturer are shown
in Fig. 2.
The performance or efficiency of a turbine is defined by the relationship between
the input power (Pin) and the output power (Pout). In this case, they are hydraulic
power (Pin) and electric power (Pout). Knowing these values from previous results,
we have the turbine efficiency at an average of 15.15%; although it exceeds what
is specified by the manufacturer that is 10.41%, it is still low determined that the
efficiency of a DC generator is between 45 and 50%.

Table 3 Energy generated by the microturbines

Pressure (bar) Electric Power (W)
Turbine #1 Turbine #2 Turbine #3 Turbine #4 Turbine #5
1 1.05 1.13 1.43 1.35 1.18
1.5 2.63 2.38 2.93 3.20 2.66
2 4.42 4.18 5.05 5.30 4.51
2.5 5.64 6.44 7.02 7.93 6.61
3 8.04 8.91 9.32 9.98 8.77
3.5 9.97 10.95 11.70 12.87 11.54
4 12.12 13.61 14.88 15.31 14.02
4.5 14.40 15.57 17.30 17.76 16.37
5 18.03 17.95 17.86 19.89 18.72
Design and Implementation of a Hydro-Energetic System … 205

Fig. 2 Microturbines performance curve

3.2 Integration of Generation Units and Components

in the Pressure-Break Tank

This section describes the implementation of the system, that is, each generation unit
(microturbines) is installed, inside a pressure-breaking tank designed. This tank, like
the hydraulic bench, is located in the UTPL Hydraulics Laboratory. The materials
mentioned in Table 1 are used for the implementation.
The tank is built with galvanized iron sheets 0.9 mm thick. It has a layer of
anticorrosive paint that helps it not rust and remain in good condition. The pressure-
breaking tank is intended to dissipate energy and decrease hydrostatic pressure [10].
Figure 3 shows the energy generation system which consists of the following
parts:

Fig. 3 Arrangement of turbines and components in the pressure-breaking tank

206 L. Jiménez et al.

• 4 microturbines,
• 4 electro valves,
• 4 smart chargers,
• 4 batteries and,
• 4 inverters.
The smart chargers are located in a DIN rail that is attached to the side of the tank
as well as the batteries. Each battery has an inverter connected to it. Additionally,
the tank has iron hooks welded inside to support the weight of the turbines and keep
the pipes fixed.

4 Drive and Telemetry System

This section shows the design and implementation of the drive and telemetry system
of the generation units (GUs). This system must fulfill two objectives. First, to obtain
electrical parameters such as voltage, current and power of each GU automatically
and send this data to an IoT platform. And second is to guarantee a 24-hour functional
generation system that maintains and adequates rest margin for the GUs.
The electrical parameters are provided by the smart charger through an RS-232
interface and to establish communication with it is necessary to login with user and
password credentials. For the automatization, a relay board is used to activate and
deactivate the GUs using a microcontroller. A general scheme of the whole system
is shown in Fig. 4.

4.1 System Hardware and Connection

The drive and telemetry system is physically integrated by the following devices:
Waspmote Pro V1.2 data acquisition board, Atlas Scientific 8:1 UART port multi-
plexer, RS-232 to TTL converter, a communication module, RJ11 connectors and a
4 relay board.

4.2 PCB Design

Figure 5 shows the PCB design composed on three MAX-232 integrated circuits,
a 5 V regulation source, sockets for the connection between the MAX232 and the
multiplexer and sockets for the connection between the Waspmote Pro card and the
multiplexer.
Design and Implementation of a Hydro-Energetic System … 207

ENERGY GENERATION SYSTEM

DRIVE AND TELEMETRY SYSTEM

|DATA SC 1|
|DATA SC 2|
|DATA SC 3|
|DATA SC 4|

DRIVE

Fig. 4 Complete system general diagram

4.3 System Algorithm

Once the components have been established, an algorithm is developed to make use
of this hardware. The proposed algorithm activates the GUs in pairs; this way the
pressure in the hydraulic system remains almost constant and guarantees 24 h/day
energy generation. Also, it makes the login in each smart charger in order to obtain
the electrical parameters and sends this to the cloud.
As mentioned before, the GUs work in pairs, but the operation time is limited
to 30 min for each pair. This allows a great rest margin for the GUs (required by
the smart charger). When a transition is going to occur, the algorithm establishes a
15-second delay which is considered fundamental from the hydraulic point of view
to avoid any damage to the turbines or pipeline. This means that the hydraulic system
will not suffer pressure transients that at some point can lead to its collapse.

5 Results

5.1 Energy Generation System Implementation

Once the generation units are placed together with their components in the tank, the
wiring is carried out, based on the user manual provided by the manufacturer [11]. As
showed in Fig. 6 and 7, the microturbines are arranged inside the tank. PVC gutters
208 L. Jiménez et al.

Fig. 5 PCB design for hardware connection

are used in order to properly organize wiring and improve aesthetics. Switches are
also used to turn and off the smart chargers.

5.2 Power of the Installed System

For this analysis, the data obtained in Table 3 are used. It is worth mentioning that the
data taken into consideration are those corresponding to 3 and 3.6 bar pressure has
given that the smart charger works in that range. Table 4 shows the average power
generated by the turbines that are part of the generation system.
The installed power of the system is obtained by adding the power generated by
each of the turbines; therefore, an installed power of 42.51 W. is obtained. Each of
the generation units operates 12 h a day and in groups of two, therefore, the entire
system works 24 h a day. Taking this information into account, the energy generated
by the entire system is 510.12 Wh/day.
Design and Implementation of a Hydro-Energetic System … 209

Fig. 6 Generation system installed in the pressure-breaking tank at UTPL Hydraulics Laboratory,
top view

5.3 Hardware Integration and Software Implementation

Figure 8 shows the designed PCB, the multiplexer and the Waspmote Pro card
together with its battery. These devices are grouped inside an airtight box protected
from water and dust.
On the other hand, the algorithm was implemented through the Waspmote
IDE version 06.04. It consists of several functions that allow fulfilling everything
previously stated in the algorithm.
The main result of the implemented code is shown in Fig. 9. Parameters such as
voltage, current, power, time of use are considered. All the data obtained is stored
within a matrix and it is through the positions that these values occupy in the matrix
that the value of each parameter can be obtained.
210 L. Jiménez et al.

Fig. 7 Generation system installed in the pressure-breaking tank at UTPL Hydraulics Laboratory,
side view

Table 4 The average power

Detail Power (W)
generated by each GU
GU #1 10.15
GU #2 10.51
GU #3 10.43
GU #4 11.42

5.4 Pressure Measurement in GU Transition

The generation system consists of 4 GU which work in pairs. That is, once the
working units have reached 30 min of operation, they are turned off so that the other
pair of GUs starts to work. When a UG is switched on and off, a pressure transient is
produced in the hydraulic system. To avoid any damage to the system, the pressure
variation in the on and off process must be as small as possible.
In order to quantify the pressure variation at the time one GU started the transition,
the system was monitored with an MLH pressure sensor. The experiment consisted
of measuring the instantaneous pressure of the system every 5 seconds to obtain the
data while a GU was making a transition.
Figure 10 shows that the pressure ranges from 3 to 3.6 bars, with an average
pressure of 3.3 bars, which is within the range allowed by the smart charger. As view
in Fig. 10, there are two times when transitions occur (described as t1 and t2) where
the pressure value decreases to 2.73 and 2.69 bar, resulting in variations of 0.57 and
0.61 bar of the average system pressure. The implemented algorithm forces these
variations to be small in order to avoid any damage to the system.
Design and Implementation of a Hydro-Energetic System … 211

Fig. 8 Hardware implementation

5.5 Data in an IoT Platform

In this work, the ThingSpeak platform is used. After the drive and telemetry systems
obtain the electrical parameters, this is upload to the cloud. In the platform, four
channels were created for each GU. In each channel, variables of voltage, current,
power, time and state are stored. The write “APIKEYS” of each channel was added
to the algorithm so the code knows where to send the data.
212 L. Jiménez et al.

Fig. 9 Smart charger login and electrical parameters

Fig. 10 Pressure variation in GU transitions

Design and Implementation of a Hydro-Energetic System … 213

6 Conclusions

The implemented system uses the energy accumulated in water distribution lines to
generate electricity. The system, which operates twenty-four hours a day, generates
510.12 Wh/day. In addition, the drive and telemetry system allows monitoring the
status of the microturbines in real time as well as sending this data to an IoT platform.
In terms of electrical generation, the microturbines generate 10 to 11.5 W in
the pressure ranges in which the smart charger works. Although they exceed the
generation parameters established by the manufacturer, their energy efficiency, which
is 15.15%, is low considering that the efficiency of a DC generator is between 45
and 50%.
The drive and telemetry system implemented allows for a 24-h functional gener-
ation system. In addition, thanks to this system, turning on and off the GUs does not
produce large pressure variations, thus providing protection to the hydraulic system.
The use of other types of energy in sensor networks and IoT applications, apart
from solar energy, makes these systems more robust and independent. For the specific
case of the coordinating node, which cannot make use of rest functions due to the
role it plays within the network, it is important to guarantee its energy autonomy by
using energy generation systems such as the one developed in this paper.

7 Recommendations

VERNE, the manufacturer of the microturbines and smart chargers used in this
work, is currently not operating, so there is little availability of information about the
devices. It is recommended that for similar developments, the devices used should
be from a manufacturer that is active in the market in case of requiring detailed
information.
For the energy generation system, it is recommended for field applications to place
a bypass to limit the fluid pressure at the system entrance. Also, this is necessary to
carry out maintenance on the deployed electronic devices.
The pressure in the system must remain within a set range, and in a laboratory
environment, it is easy to monitor and regulate this parameter. However, if deployed, it
is recommended to implement an additional control system that constantly measures
the pressure in the system to regulate it if necessary. The information obtained may
also be available on an IoT platform.
Prior to the construction of the pressure-breaking tank, the design must consider
the measures of all the elements to use in the energy generation system. These
include microturbines, solenoid valves, pipes, elbows, joints, among others. Other-
wise, space, where the generation units are located, will be limited resulting in
additional adjustments to the tank.
In the case of deploying the system, protection for batteries and smart chargers
must be considered. The batteries should be stored in a place where they are protected
214 L. Jiménez et al.

from sudden changes in temperature. In addition, this place must have ventilation so
that toxic gases are not stored. As for the chargers, they must be protected from rain
as recommended by the manufacturer.
If the system is deployed and there is cellular coverage on the deployment area,
it is recommended to use the mobile network to send data to the IoT platform. In
addition, instead of HTTP, use the MQTT protocol in order to reduce costs.

References

1. Kamruzzaman, S.M., Fernando, X., Jaseemuddin, M.: Energy Harvesting Wireless Sensors for
Smart Cities, pp. 218–222 (2017)
2. Silicon Labs, The Evolution of Wireless Sensor Networks (2013)
3. K.Z. Panatik et al.: Energy Harvesting in Wireless Sensor Networks : A Survey, pp. 28–30
(2016). K. Elissa, Title of paper if known, unpublished
4. Azevedo, J.A.R., Santos, F.E.S.: Energy harvesting from wind and water for autonomous
wireless sensor nodes. IET Circ. Devices Syst. 6(6), 413–420 (2012)
5. Chaudhry, B.S.: Energy consumption and harvesting in WSNs: Challenges and solutions. In:
2017 International Conference on Innovations in Electrical Engineering and Computational
Technologies (ICIEECT) (2017) pp. 1–2
6. Mecocci, A., Peruzzi, G., Pozzebon, A., Vaccarella, P.: Architecture of a hydroelectrically
powered wireless sensor node for underground environmental monitoring. IET Wirel. Sens.
Syst. 7(5), 123–129 (2017)
7. Morais, R., et al.: Sun, wind and water flow as energy supply for small stationary data acquisition
platforms. Comput. Electron. Agric. 64(2), 120–132 (2008)
8. Mohamed, M.I., Wu, W.Y., Moniri, M.: Power harvesting for smart sensor networks in moni-
toring water distribution system. In: 2011 International Conference on Networking, Sensing
and Control, 2011, pp. 393–398
9. Grupo Verne.: Microturbina TRD (2015). (Online). Available: http://grupoverne.com/produc
tos/microturbina-trd. Accessed 01 Feb 2017
10. Vivanco, C., Comunicación oral: Tanque rompe-presión (2018)
11. Grupo Verne.:Microturbina TRD. Manual de Usuario (2015)
Hydraulic Fill Assessment Model Using
Weathered Granitoids Based
on Analytical Solutions to Mitigate Rock
Mass Instability in Conventional
Underground Mining

Cristhian Portocarrero-Urdanivia , Angela Ochoa-Cuentas ,

Luis Arauzo-Gallardo , and Carlos Raymundo

Abstract This study uses analytical solutions to assess a hydraulic fill model based
on weathered granitoid to increase underground opening stability and mitigate rock
bursts during mining operations in a conventional underground mining company
located in the Coastal Batholiths of the Peruvian Andes. This study assesses the
previous geological database provided by the mine, analyzes the on-site strengths
produced by the exploitation works that will subsequently be filled, identifies the
quality of the material used in the landfill (granitoids) through laboratory tests,
and compares compressive strength at different depths, all contemplated within the
landfill model used. This study focuses on the applicability of hydraulic fills in
conventional underground mine using natural geological material such as granitoid.

Keywords Hydraulic fill · Granitoids · Conventional underground mining ·

Instability · Analytical solutions · Rock bursts

1 Introduction

The mining industry is one of the economic pillars that drive the development of
several countries, such as Canada, Australia, and South Africa, relying on innovative

C. Portocarrero-Urdanivia · A. Ochoa-Cuentas · L. Arauzo-Gallardo · C. Raymundo (B)

Escuela de Ingeniería de Gestión Minera, Universidad Peruana de Ciencias Aplicadas (UPC),
Lima 15023, Peru
e-mail: [email protected]
C. Portocarrero-Urdanivia
e-mail: [email protected]
A. Ochoa-Cuentas
e-mail: [email protected]
L. Arauzo-Gallardo
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 215
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_21
216 C. Portocarrero-Urdanivia et al.

technologies in open pit or underground mines [1]. In South America, countries such
as Chile, Peru, and Brazil have exploited their mining resources for later commer-
cialization. Although large, medium, and small mining operations are developed in
some countries, accidents take place across all the mining operations. Some clear
examples of these accidents are rock mass instability produced by the extraction of
mineral ore, rockfalls, worker slips, trips, and falls, and poisoning–asphyxiation.
This study selected rock mass caused by mineral ore extraction as a key issue
affecting conventional underground mining. Within this analyzed context and due to
the different problems identified over time, this particular matter was deemed critical
for the mining industry because exploited opening instability generates empty spaces
subject to various stress and forces that increase with depth, thereby preventing
resource exploitation operations. Hydraulic fill is currently used in Peru with several
aggregates, such as cement, in large and medium mining operations. However, it is
not yet used in conventional underground mining owing to its high cost and additives
that must be added to the landfill to increase its resistance. Therefore, this study
proposes adapting hydraulic fills with granitoid in a conventional underground mine.
At the Marsa mining company located in the Pataz batholith, hydraulic fills were
supplemented with granitoid and rock bursts were successfully controlled and safe
operation was ensured [2].
This study aims to demonstrate the application of hydraulic fills with granitoid
because it is more affordable and adaptable for conventional underground mining
operations than other existing solutions, especially because of its short hardening
times.

2 State of the Art

2.1 Hydraulic Fill Method in Underground Mining

Several types of research have investigated the hydraulic fill properties in under-
ground mines and concluded that this type of landfill usually settles at a porosity
of approximately 40% and its relative densities are in the range of 40–70% [3]. In
addition, when settling as a suspension with 65–75% solid content, these landfills are
deposited at a dry density of approximately 0.56 times the specific gravity. Further-
more, the permeability of pulp (10–30 mm/h) is much lower than that desired in the
mining industry (100 mm/h) [4, 5].
As a part of their results, some authors agreed that hydraulic fills must ensure good
drainage so that the excess water may drain quickly. For achieving a good drainage
system, the effective size of the D10 grain must exceed 10 µm. Finally, the literature
has shown that when hydraulic fills are mixed as mud with 30% of water content, an
extremely dense hydraulic fill is obtained [6, 7].
Hydraulic Fill Assessment Model Using Weathered … 217

2.2 Using Granitoids in Hydraulic Fill Models

in Conventional Underground Mining

Very few researches have investigated the application of granitoid to hydraulic

fills; however, granite exhibits unique properties that provide better reinforcement
to hydraulic fills. In previous research, granite rocks were used in landfills, and
favorable results from the viewpoint of percolation, sedimentation, and permeability
were obtained [8]. These characteristics help us better identify the properties exhib-
ited by granites as an aggregate in hydraulic fills, thereby improving percolation,
permeability, resistance, and hardness, among others. [9, 10].

2.3 Using Analytical Solutions in Hydraulic Fills

in Conventional Underground Mining

Researchers have considered rock bursts as dynamic disasters that take place when
the energy accumulated in the rock mass of the mine is released suddenly, abruptly,
and violently. Therefore, these studies used analytical solutions to assess the quality
of the rock mass to be filled [10, 11].
Researchers have also claimed that the results obtained using analytical solutions
in the landfill, such as the constraining strength and pressures developing in the
landfill mass, must be determined according to the geometry of the exploded pits
and initial stress conditions [12].

3 Contribution

This hydraulic fill assessment model is based on two concepts: granitoid, i.e., the
material used as input, and analytical solutions, which will be a determining factor
in terms of design and implementation. The proposed model displays the developed
methodological approach and different steps used for predicting hydraulic fill perfor-
mance properties and streamline its mixing. The methodology includes three main
phases: experimental, modeling, and optimization, as presented in Fig. 1.

3.1 Model Components

Phase 1 This phase includes the experimental process of the pulp formed by water
and granitoid in laboratory tests based on the main parameters and analysis that
may affect its strength and consistency. This phase aims to generate inputs, such
as granulometry and percolation, through experimental and analytical assessments
aligned with the theoretical model.
218 C. Portocarrero-Urdanivia et al.

Fig. 1 Applied methodology

Phase 2 This phase includes simulation of the hydraulic fill model based on
analytical solutions and laboratory tests in the established domains and parameters.

Phase 3 Proposed model validation process: For this purpose, the results obtained are
assessed in terms of the following three basic factors: uniaxial compressive strength,
slump test, and the costs incurred for its subsequent application. The aim is to opti-
mize landfill improvements by creating optimum mixtures, thereby providing a tool
focused on landfills.

3.2 Process View

As a first step in the implementation of this type of landfill, the current situa-
tion must be analyzed to gather information on the mining site conditions through
studies focusing on the rock mass and exploited area that will subsequently be filled.
Gathering information regarding the exploitation method and rock mass parameters
will provide the basic knowledge related to the compressive strength for which the
hydraulic fill is subjected.
The geometric dimensions of the pits to be filled must be established according
to theoretical values of the arc, whose width can determine as follows:

W t = (0.15D + 18.3)3/4 (1)

where: Wt =Pit width, D =Pit depth from the surface.

Hydraulic Fill Assessment Model Using Weathered … 219

Any value between these parameters will allow the internal pit to withstand a load
equivalent to half the width of the pit. However, if the operational requirements are
exceeded, the safety factor of 2.5 will have to be used instead of 2.
If the room and pillar mining method are considered, the hydraulic fill using
granitoid will recover the alternate pillars by acting as ultimate active support that
facilitates the exploitation of the adjacent pit. For structures exceeding 4 meters,
mining operations must be performed by level or using another mining method, such
as a sublevel stopping method. To determine the volume to be filled, determining the
dimensions of an exploded and empty front would be ideal (Fig. 2).

x2
F(x) = − (2)
1.30
1.25
x2
S3 = dx
1.30
−1.25

S3 = 1 m2 (3)

S2 = 2.5 × 1.20 − S3
S2 = 2 m2 (4)

Fig. 2 Section pending of

filling
220 C. Portocarrero-Urdanivia et al.

Fig. 3 Rock mass assessment

S1 = 2.5 × 1.5
S1 = 3.75 m2 (5)

In addition, a detailed geomechanical assessment was conducted. All the relevant

parameters for a conventional underground mine are listed in Fig. 3.
The optimum particle size distribution for the landfill to reach a maximum on-
site density that may guarantee the design uniaxial compressive strength value is
determined by the Talbot curve because it provides the maximum density required
for landfill optimization.
Furthermore, other laboratory calculations were also made, as shown below:
• Specific gravity: To calculate the specific gravity or bulk density, two samples
of previously standardized and quartered weatherized granodiorite was extracted.
After assessing both samples, an average specific gravity of 2.58 was obtained.
• Material hardness: According to the MOHS scale of mineral hardness, the
hardness of weathered granitoid is 4.0. The critical speed of the pulp was 2.59 m/s.
• Pore ratio: It is defined as the ratio between the volume occupied by pores (V p )
and that occupied by solids (V s ). In a granular mass, the pore ratio is generally
30%.
• Porosity: It was calculated as 64.89%, revealing an excellent porosity rate, which
is positive for hydraulic fills since water is quickly percolated. This means that
water will flow out of the landfill in less time; therefore, the pulp will also dry
faster.
• Solid percentage: The number of solids present in the landfill was 75% of the total
weight.
• Sedimentation rate (slump test): First, the granodiorite and water must be mixed in
a ratio of 3:1 to form the pulp. Then, the pulp will be gradually filled in the slump
at a height of 19 cm, as per the requirements for the slump test. After standing for
3 min, the final height was 18 cm.
Hydraulic Fill Assessment Model Using Weathered … 221

GRANULOMETRIC CURVE
100

CumulaƟve Percentage
80
60
40

20

0
20000 2000 200 20
Mesh in microns
Granulometric curve of the material Talbot curve

Fig. 4 Comparison against the talbot curve

3.3 Indicator View

The function that indicates the amount of strength exhibited by the landfill with
granitoid for each meter of the height of the self-established slope was obtained
(Fig. 4).
Since the excellent quality of the granitoid exceeds in the Talbot curve, the strength
of the proposed models exceeds the required strength, thereby validating the proposed
model.

4 Validation

4.1 Scenario Description

The hydraulic fill system model supplemented by granitoid will be implemented at

the San Marcelo Mining project geographically located in the Paucartambo batholith
in the Eastern Andes. The study area covers an approximate area of 600 hectares of
weathered granitoid from the Quaternary age.

Current conditions of the San Marcelo Mining Landfill The required landfill
volume is determined based on the current production volume reported by the
company, listed as (Table 1).

Based on these results, an approximate landfill volume was determined for a

production level of 24,000 t/month with a specific gravity of 3 t/m3 .
We also identified the volume of the different landfill types currently used, listed
as (Table 2).
222 C. Portocarrero-Urdanivia et al.

Table 1 Monthly production

Method T.C.S. Percentage
Overhand cut and fill 16,468.00 76.26
Shrinkage 3154.00 14.61
Advances 1973.00 9.14
Monthly production 21,595.00 100

Table 2 Landfill progress in

Landfill type m3 Percentage
meters
Hydraulic fill 5745.00 77.65
Conventional fill 1654.00 22.35
Total 7399.00 100

5 Discussions

The main contribution of this study is that we are using a material that is not common
for use in hydraulic fillings, however, its use would bring operational benefits. A
detritic filling and a hydraulic filling with granitoid, the difference is that the detritic
filling is a mechanical filling, while the hydraulic filling with granitoid is a pulp that
will be transported by pipes providing more consistency in the empty spaces (intersti-
tium) of the exploited pits. The hydraulic filling with granitoid is a filler specifically
suitable for the method of exploitation cut and ascending filler where it will not be
necessary to use a lot of cement (cemented hydraulic filler), since the weathering
granodiorite in contact with water has a welding effect (clay repellent, plagioclase),
which after 5 h percolates all the water creating a consistent platform that allows
entering the next guard to continue the programmed exploitation, bringing produc-
tivity improvements, reducing operational times and costs. This type of hydraulic
filling can be implemented in other types of deposits located exclusively in batoliths
(precambrian rocks) since it is a place where granitoid can be found. This type of
hydraulic filling is beneficial in gold deposits since many gold companies started
with the recovery of the ore (gold) that goes to the tailings, and it would not be
feasible to use that same tailing for their mine filling, so they use the hydraulic filling
with granitoid, because even this filling allows the recovery of pillars.

6 Conclusions

The applicability of hydraulic fill conditions using weathered granitoid is governed

by the physical characteristics of the rock mass and exploitation method. Some of the
variables applied in this model are the presence of an incompetent roof cavity, subhor-
izontal veins, variable vein shoot, continuous mineralized structures, and mineral
recovery close to 100%. The recovery exhibited by this hydraulic fill model exceeds
Hydraulic Fill Assessment Model Using Weathered … 223

95%. Results obtained for this type of landfill using granitoid were the exposure of
filled pits exceeding 5 m without requiring cement because a uniaxial compressive
strength close to 0.8 kg/cm2 was achieved. The proposed hydraulic fill with grani-
toid would eventually reach the pit; however, considering a case in which the work
conducted at a level higher than that at the plant, we hereby propose a pumping
system. Different scenarios can be identified to assimilate the on-site stress that the
hydraulic fill may withstand. At the depth of 740 m, the horizontal stress considered
is at 19.98 and vertical stress is at 28.02 MPa.
We propose the expansion of this research as a contribution to the study of
hydraulic fillings in underground mining since its application would imply the
improvement of the company as an organization compared to the competition.

References

1. Dietz, M., Oremek, G.M., Groneberg, D.A., Bendels, M.H.K.: Was ist ein Gebirgsschlag?
Zentralblatt Für Arbeitsmedizin, Arbeitsschutz Und Ergonomie 68(1), 45–49 (2018). https://
doi.org/10.1007/s40664-017-0215-z
2. Farhad Howladar, M., Mostafijul Karim, M.: The selection of backfill materials for Barapukuria
underground coal mine, Dinajpur, Bangladesh: insight from the assessments of engineering
properties of some selective materials. Environ. Earth Sci. 73(10), 6153–6165 (2015). https://
doi.org/10.1007/s12665-014-3841-1
3. Köken, E., Özarslan, A., Bacak, G.. Weathering effects on physical properties and mate-
rial behaviour of granodiorite rocks. In: ISRM International Symposium - EUROCK 2016,
pp. 331–336 (2016). Retrieved from https://www.scopus.com/inward/record.uri?eid=2-s2.0-
85061015692&partnerID=40&md5=959e2c56486f0ebf5f3ddc715f387a67
4. Momeni, A., Karakus, M., Khanlari, G.R., Heidari, M.: Effects of cyclic loading on the mechan-
ical properties of a granite. Int. J. Rock Mech. Min. Sci. 77, 89–96 (2015). https://doi.org/10.
1016/j.ijrmms.2015.03.029
5. Strozik, G.: Design of underground mine voids filling operations in difficult flow conditions
of fly ash—water mixtures. In: IOP Conference Series: Earth and Environmental Science, vol.
174, p. 012014 (2018). https://doi.org/10.1088/1755-1315/174/1/012014
6. Yang, H., Liu, J., Zhou, X.: Effects of the loading and unloading conditions on the stress
relaxation behavior of pre-cracked granite. Rock Mech. Rock Eng. 50(5), 1157–1169 (2017).
https://doi.org/10.1007/s00603-016-1161-3
7. Zhao, K., Li, Q., Yan, Y., Zhou, K., Gu, S., Zhu, S.: Numerical calculation analysis of the struc-
tural stability of cemented fill under different cement-sand ratios and concentration conditions.
Adv. Civil Eng. 2018, 1–9 (2018). https://doi.org/10.1155/2018/1260787
8. Gupta, A.K., Paul, B.: Comparative analysis of different materials to be used for backfilling in
underground mine voids with a reference to hydraulic stowing. Int. J. Oil Gas Coal Technol.
15(4), 425 (2017). https://doi.org/10.1504/IJOGCT.2017.084830
9. Stone, D.: The evolution of paste for backfill. In: Proceedings of the 11th International
Symposium on Mining with Backfill: Mine Fill 2014, Australia, pp. 31–38 (2014)
10. Sivakugan, N., Rankine, K.J., Rankine, R.M.: Permeability of hydraulic fills and barricade
bricks. Geotech. Geol. Eng. 24, 661–673 (2006)
11. Beltran. W.: Estudio Experimental de relleno hidráulico en mina Atacocha (2014)
12. Webb, P.: Hydrated lime vs hydraulic lime (2018); Rankine, K.J., Sivakugan, N., Cowling,
R.: Emplaced geotechnical characteristics of hydraulic fills in a number of Australian mines.
Geotech. Geol. Eng. 24,1–14 (2006)
A Study on Investigations Carried Out
in Dams from the Perspective of Risk
Analysis

Gabriel Gomes de Oliveira , Yuzo Iano ,

Ana Carolina Borges Monteiro , Reinaldo Padilha França ,
Diego Pajuelo , and Pablo David Minango Negrete

Abstract With the population growth, the Petrobras modernization REPLAN

(Paulinia Refinery), the progressive region development and the shortage of resources
from new wellsprings, there was a need to increase the water availability of the region
of Campinas—SP and with that will be implanted two dams, one in the municipality
of Quarry—SP and another in the district of Two Bridges in the municipality of
Amparo—SP. The construction of the dams occurred for many reasons, one of which
is to benefit many cities through water reserves serves to about 5.5 million inhabitants
are that downstream of the reservoirs. These concepts, the present work AIMS to
present all the criteria and principles that are related to the construction of the dams.
Through research and consultations will be all the characteristics presented, direct
and indirect information,

Keywords Dam · Case study · Environmental impact · Water availability

G. G. de Oliveira (B) · Y. Iano · A. C. B. Monteiro · R. P. França · D. Pajuelo · P. D. M. Negrete

School of Electrical and Computer Engineering (FEEC), University of Campinas, UNICAMP, Av.
Albert Einstein-400, Barão Geraldo, Campinas, SP, Brazil
e-mail: [email protected]
Y. Iano
e-mail: [email protected]
A. C. B. Monteiro
e-mail: [email protected]
R. P. França
e-mail: [email protected]
D. Pajuelo
e-mail: [email protected]
P. D. M. Negrete
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 225
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_22
226 G. G. de Oliveira et al.

1 Introduction

The construction of dams has always been paramount to the growth and development
of mankind, from the oldest to which it is registered, such as Egypt, the Middle East,
and India. Currently, concern about water resources involving quantity and quality
has been accentuated due to drastic population growth and high pollution index, so
the water has become a precious commodity and the construction of dams is a way to
save her in order to have adequate human consumption without scarcity and quality
[1–4].
The implementation of dams directly affects areas of major proportions, and thus
end up being great potential works negative impact on the environment (flora, relief,
and fauna), to the cultural environment (historical, artistic, archeological, architec-
tural, heritage and people celebrations), artificial (towns, urban areas, villages) and
labor (and material and immaterial that are necessary for the performance of labor
exercises for humans) [5, 6].
These types of reservoirs are artificially made obstacles able to retain water for
the purpose of industrial or domestic supply, navigation, irrigation, recreation, sedi-
mentation control, and flood and are also intended for the production of electricity.
For water supply dams it is a necessary follow-up, monitoring, and also its mainte-
nance, it is one way to avoid any kind of accident that can occur and thereby ensure
the proper functioning of these reservoirs. Such containers are made with different
constructive techniques and varied sizes and this is set as required for the use, in
the case of farms are generally used small clumps of earth and often large structures
used for hydroelectric power supply and water supply, and between various other
purposes are constructed from landfills or concrete [7, 8].
The amount of water that is stored in the reservoir will depend on the needs to be
met. It should be set immediately after getting the planialtimetric survey of where
they will be flooded by the dam. The main dam construction means are concrete arch,
riprap, concrete, fill and gravity. These structures have further included the houses
of power, spillways, discharge structures, and control units.
Brazil is the largest country in Latin America (in territorial extension), having
approximately 209.3 million inhabitants, who live in 26 Brazilian states and the
Federal District. Its states are divided into five regions (North, Northeast, Midwest,
Southeast, and South), is the most populous southeast region, with approximately 80
million inhabitants, distributed among the states of Rio de Janeiro, Espirito Santo,
Minas Gerais, and São Paulo. The state of Sao Paulo has about 12 million individuals.
This high population concentration is due to large industrial and technological invest-
ments in these areas, thus generating greater financial circulation and consequently
more chances of employment [9–12].
Located in the state of Sao Paulo, the Campinas region is made up of 22
cities (Americana, Artur Nogueira, Campinas, Cosmopolis, Engenheiro Coelho,
Holambra, Hortolandia, Indaiatuba, Itatiba, Jaguariuna, Monte Mor, Morungada,
Nova Odessa, Paulínia, Pedreira, Santa Barbara d ‘Oeste, Santo Antonio de Posse,
Sumaré, Valinhos, and Vinhedo), 3,224,443 inhabitants, and total GDP (Gross
A Study on Investigations Carried Out in Dams … 227

Domestic Product) of R $ 142,301 million. This scenario has attracted more and more
people to live and work in this region. However, with warming and public miscon-
duct, many Brazilian cities have been suffering from water shortages at certain times
of the year, especially when temperatures can reach 37 degrees Celsius. Based on
this, there is a need to create dams to supply the entire population. Thus, in 2014
the process was started for the construction of the dams of the cities of Pedreira,
Duas Pontes and Amparo, which when completed will be dependent on the Jaguari
and Camanducaia rivers, which are located in the cities of Pedreira and Amparo,
respectively [9–12].
With the construction of dams stands out as a major benefit, increased water
availability for the population of the municipalities to be covered, which will require
well in water security, since studies show that if you do not adopt such measures in
2025 about 7 cities entered for water resource vulnerability list [8].
Environmental and socio-cultural impacts of water projects of this size require a
great preventive assessment, influencing the idea of design, formulation, and alter-
natives to the system, as well as in the design and detailed design. This assessment is
extremely important both for environmental and social viability as economic, high-
lighting the impact studies, expropriation, flooding so that future problems directly
related spending can be avoided. This assessment helps minimize problems, but not
extinguished, it is inevitable that some of the forests are devastated for project imple-
mentation causing the loss of biodiversity, degradation of upstream areas caused by
the flooding of the reservoir area, among other problems [13, 14].
Investigations based on geological studies are primary tools for dam analysis,
mainly as a form of prophylaxis for future human, environmental, financial and health
problems. However, due to the high costs resulting from this study often generates
the negligence of large companies in the supervision of their dams. Based on this,
the present study aims to perform an analysis of works related to the construction of
these dams in order to detect possible failures during the process as well as to expose
the possible damages and risks that populations face when there is negligence on the
part of companies.

2 Methodology

This study was conducted by gathering bibliographic data in the literature on dams
in Brazil, focusing on the water dam under construction in the Campinas region and
the two largest Brazilian environmental disasters involving ore tailings dams. The
results found are confronted with environmental, social and health issues.
228 G. G. de Oliveira et al.

3 Results and Discuss

The dam at the Two Bridges dominates site and reservoir are characterized by elon-
gated mirrors, medium to high slopes (>15%) and local amplitudes below 100 m. The
hills have rounded tops and straight slopes, south relief mountains, known as Serra
dos Feixos. The well where the future dam will be installed cuts the Camanducaia
River, characterized by stretches of rocks inserted in sandy soil, providing rapids in
the east-west direction. Drilling tests were performed at the locations on the edge of
the right SMDP probe 01, at a level of 1.60 m of residual soil followed by a change
in the soil rock composite by micaceous clayey sludge that overcame resistance 40
strokes in the test SPT after 15.00 m deep [15, 16].
In the left margin, it was necessary to carry out three points polls for more precise
information because there is greater instability in the soil. The 02-SMDP probe,
located near the river, pierced a layer of 4.00 m landfill, followed by a 1.00-m
alluvial layer comprising silt, sandy loam, with rock elements grayish-yellow in
pigment. Further down, at approximately 22.00 m deep, there is a change in rock
with silty-sandy texture micaceous. Continuing were performed on the left edge 03
and the SMDP-SMDP-04 polls, which is characterized in general by having layer of
1.00 m thick composed of colluvium bit sandy silt clay having small gravel; layer
5.60–6.00 m mature little residual soil sandy silt clay consistency and depth resistant
03-10 strokes; a ground-level shift of rock with silty-sandy texture micaceous found
some rock fragments. The SMDP-03 and-04 SMDP found groundwater surveys the
depth of 7.50 m and 20.00 m, respectively [15, 16].
In the report of the study pointed out that the quarry reservoir is inserted in a
mountainous relief Elongated hills, and its average slope characteristics and high
(>15%) and areas with ranges greater than 300 m. And it has shaped angular tops
with ravines and slopes with rectilinear shapes. The axis is intercepted by sharp bends
river Jaguari with portions of the bedrock originating tubing [15, 16].
During the surveys were carried out on the banks, both right and left, on the right
edge was performed SMPE-01 and found layers of rock changes, micaceous clay
and sandy silt clay micaceous. And by the results, it has resistance at SPT 40 strokes
starting at 12.00 m deep. In the left margin, analyzes were performed 4, 02-SMPE
polls the SMPE-05. The SMPE-02 probe was implemented on top of the rocky bed
of the water body and provided with mylonite gneiss depth of 15.90 m. Since the
SMPE-03 polls the SMPE-05, identified rock material layers (thickness varying from
1.00 to 4.50 m), silty clay-bit sandy soil, silt, clay and sandy micaceous the mylonite
gneiss. The resistance values presented a gradual penetration to the depth reaching
40 blows to 15 m. In the investigation pointed rock masses that show medians and
consistent changes at the top, passing the bedrock little changed in the final meters
[15, 16].
A Study on Investigations Carried Out in Dams … 229

4 Physical Impacts Analysis of Two Bridges and Reservoirs

Quarry

With the introduction of Quarry and Two Bridges dam, according to Article 2 of
CONAMA Resolution 349 (1986) to Directly Area Affected (DAA) is the place
where will be implemented the project with its settled structures, access roads that
may be increased produced or changed, with its unit procedures related to his work
infrastructure. Therefore the Direct Area of Influence (DAI) is the region which will
be reached directly from the impacts on the development coinciding with the (DAA)
may have positive or negative interference. The quarry dam, in turn, affects the EPA
(Environmental Protection Area) of Piracicaba/Juqueri-Mirim, the city of Campinas
also has an affected area and the same is located on the left edge of the reservoir
deployment. The APA of Piracicaba/Juqueri-Mirim, in turn, covers part of the left
edge of the tank and all directly affected and influenced area [17, 18].
With the implementation of the quarry dam, Annex K will be flooded about
105 ha APA corresponding part of Campinas being flooded. Thus all the DAI, which
corresponds to the left corner of the reservoir presented zoning in the management
scheme, the Hydro Conservation area Jaguari, which its procedures is to ensure the
preservation of sources and water, the Just as the recovery and protection of riparian
vegetation, and permanent banning of pesticides and chemical fertilizers [17].
This application boundaries EPA and DAI, follows the same design of the study
units for sub-basins and may require adjustments, it helps the planned environmental
programs Pedreira dam act in proportion to the guidelines and the EPA region of
objectives belonging to the Basin Jaguari. The introduction of the quarry will have a
dam reservoir with an area of approximately 2.02 km2 to about 82 ha overflow of the
storage units. In this case, the conservation area of Campinas will have about 33 ha
and the Piracicaba/Juqueri-Mirim have lost 97 ha. It tends to be noted that this total
is added to a flood area of approximately 72 ha area corresponding to the overlapping
of the two affected areas [18].
The EPA in Campinas will be affected by more than 105 ha which represent the
sum of the areas of 33 ha reservoir and the area lost with limits applied to 72 ha.
When added to the array of the tank, that area will have a total of 232 ha affected.
In turn, EPA Piracicaba/Juqueri Mirim 169 ha has affected areas occupied by the
quarry and reservoir 97 ha area is lost with overlapping boundaries 72 ha. When
added to the array of quarry dam the entire length will have a total of 370 ha affected.
In Table 1 is showed the affected storage units [17, 18].
The Two Bridges and quarry reservoir will occupy areas that are material goods,
residential and agricultural activities that are called a small scale, mainly for self-
consumption. In the case of the dam of Two Bridges, besides the presence of a small
proportion of activity, there is also the presence of large areas where there is major
business uses. In this case, there is the Farm Jaborandi industrial group Ypê which
carries out large-scale poultry farming jobs. There is also the Agriculture Tuiuti
(Shefa), which is a dairy, which will not be fully impaired. Since the establishment
230 G. G. de Oliveira et al.

Table 1 Protected areas affected by quarry dam

Conservation Unit Total area Flooded DAA Total Total with
area occupation overlapping
Campinas 22.278 33 53 86 232
Overlapping area between 2966 72 74 146 –
conservation units
Piracicaba/Juqueri-Mirim 287.000 97 127 224 370

Resorts, 4 Angels who works as a research laboratory and studies of animal vaccines
and is a Research Center for Animals of Brazil will be completely flooded [18].
Regarding the commercial activities can point to the known “Ecological Reserve
Mundão of Trails,” which develops tourism activities, relying primarily on the area
and structure to track and camping near the river Camanducaia, the District of
Arcadas, Amparo. Although titled “Ecological Reserve” is not protected conser-
vation unit or established by law, unless that focus on permanent preservation areas
and reserves. On the left edge of the river is located Jaguari the “Aunt Bar,” down-
stream from the dam quarry, which is likely to suffer interference by dam construction
[17, 18].
There are also two small main dams that will be flooded. The river Jaguari is
located the PCH White Monkey, with 2.36 MW of installed capacity owned by
Electricity company and should be affected by the formation of the quarry dam and
river Camanducaia, SHP Feixos, with 1.15 MW installed capacity of the Company
Energy Wolf Jump Ltda, which will be hindered by the Two Bridges bus. In popu-
lation surveys in areas to be occupied by the dam quarry and two bridges a total of
82 families were identified, of these 27 families are located in the area to be flooded
quarry and 55 families in the dam Amparo [17, 18]. Table 2 shows the families living
in the area that will be compromised by construction.
The compensation for expropriation can be performed in two ways: through
payment in cash or government bonds (in urban environments or in the case of
land reform). The impacts in rural areas with productive activity, considering the
loss of areas dedicated to agriculture and forestry, will be enormous because these
areas are in the direction of the dam spillway. The quarry dams and bridges two
reservoirs should form with total area respectively of 202 and 486 ha2 , the areas to
be flooded effectively correspond to 181 ha and 463 ha and other areas are occupied
by the river.

Table 2 Families living in the area directly affected (DAA)

Families Two bridges Quarry Total
Families interviewed 44 25 69
Family no information (*) 11 2 13
Total families 55 27 82
* indicates without information about these families
A Study on Investigations Carried Out in Dams … 231

In recent years the region near these rivers is increasing its population rate signif-
icantly, so has increased pressure on water demand, finally, states have adopted
measures to be no lack of water at the time and in the future for the population
and aimed at development of the region, thereby improving the quality of life of
the population. The quarry reservoir is located near Highway John Beira, known as
SP-095 and the town road Basil Vieira de Godoi, already the Two Bridges is close
to the highway John Beira, also known as SP-095 and the highway Aziz Lian, also
known as SP-107 [17].
The quarry dam is 3 km from the urban center of the city of Pereira and will be
installed in the river Jaguar, located in the municipalities of Pedreira and Campinas,
such reservoir will have an area of 2.02 km2 and can store about 32 billion liters of
water. Already the dam of Two Bridges is 8 km from the city of Amparo, and will
be introduced in the river Camanducaia and will run until Ribeirão Pantaleon such a
reservoir will have an area of 4.86 km2 and can store about 53 billion liters of water
[18].
Both dams are in regions with rapid economic development and population, and
that was by agribusiness enterprises and large national and multinational companies
in the region, so it has been the need to increase the water demand in the region.

5 Environmental Disasters in Brazil: The Mariana

and Brumadinho Dams

Actually in Brazil, when the word “dam” is said, it is associated with the word
catastrophes, deaths, environmental crime, ecological disaster. This is because, on
November 5, 2015, there was a breakdown of an ore tailings dam in the city of
Mariana, Minas Gerais (Southeast Brazil). The Mariana tailings dam is owned by
Samarco, which is managed by Vale S. A. and BHP Billiton. The disruption resulted
in 43.7 million cubic meters of tailings dumped and a total of 19 deaths. The mud-
covered a total of 663 km until it reached the sea in the state of Espírito Santo. In
addition, the mud has reached the Doce River, which covers 230 municipalities that
have their bed as a subsistence tool. According to environmentalists, the effects of
tailings at sea will be felt for at least 100 years. One month after this catastrophe
occurred, 11 tons of dead fish were removed in both Minas Gerais and Espírito
Santo. Lives were taken, districts were destroyed and thousands of residents were
left without water, homeless and without work. The rupture of the tailings dam
(Fundão) in Mariana is considered the biggest environmental disaster in the country’s
history. Until the moment the region remains uninhabited and covered with mud.
Many residents still receive government assistance. The feasibility of building a new
district for the relocation of these families is studied [19–22].
However, isolation from communities and loss of access to health services can
aggravate existing chronic diseases in the affected population (hypertension, diabetes,
232 G. G. de Oliveira et al.

kidney failure, tuberculosis, etc.), as well as triggering new deleterious health situa-
tions such as mental illness (depression). and anxiety), hypertensive crises, respira-
tory diseases, domestic accidents and outbreaks of infectious diseases. The increased
incidence of pre-existing diseases in the region, such as yellow fever, diarrhea, and
schistosomiasis may be a consequence of the medium-term disaster. In addition,
contact with mud and water may lead to cases of leptospirosis [23].
On January 29, another tragedy struck the city of Brumadinho, Minas Gerais. This
dam belongs to the Brazilian multinational mining company Vale S.A. As much as
this breach has dumped a smaller amount of waste compared to Mariana, its social and
environmental impacts have been as great. According to the National Mining Agency,
the dam had no documentary backlogs, and was considered inactive, meaning it was
not receiving any new tailings. The volume of excrement manure was about 12
million m3 (1 m3 equals 1000 L) and the speed of the mud reached 80 km per hour.
At the time of the tragedy, security sirens should have been sounded to alert Vale
workers and local residents. This, however, did not happen [19–22].
The mud, which contained iron, silica, and water, reached the Paraopeba River,
which in turn negatively affected water quality, as this water present in the river
presented risks. By the end of August 2019 (about 7 months later), 241 deaths and
21 people still missing are counted [19–22].
It is necessary to examine the presence of heavy metals in the tailings and their
monitoring along these rivers to avoid the consumption and use of contaminated
waters in the coming years. Heavy metal-enriched sediment can be remobilized to
rivers with heavy rainfall, dredging and hydroelectric dam operation over the next
few years [23].
It is important to point out that both environmental disasters were due to human
failures, one of them based on a communication failure between the responsible
companies, employees and city dwellers. Therefore, it is necessary to implement
more modern communication systems that present low abstraction level, speed and
low computational memory consumption [23–26], regardless of the operating system
used by the company [27]. Moreover, it is important to invest in low cost, high
accuracy medical methodologies that do not depend on specific equipment, so that
the exams can be performed through a simple computer [28, 29].

6 Conclusions

By analyzing all the necessary procedures for the implementation of these dams, it
was found that the project will benefit several cities in the region with the supply
of water, which is recurrent shortages in drought times, and also will bring, all
matters involving environmental impacts in the area, and the physical impacts that
are associated with the expropriation of the owners of homes and loss of heritage
sites to be flooded.
With work and the technical site, visit can understand the size and grandeur of
the buses that aims to help society as a whole, and also comprehend and understand
A Study on Investigations Carried Out in Dams … 233

the frustration of residents who are expropriating their homes because their lands
were some generations of family heritage and for the construction of dams, but
cannot observe reality broadly and understand that this project will benefit millions
of people.
Since it is extremely important to build these dams even if it works with consid-
erable environmental impacts in the affected area and physical impact on people’s
lifestyle directly injured with expropriations, but that in the future will benefit because
there are no setbacks with the lack of d ‘ water, it will be one of the viable alternatives
to the water crisis that is regular and that affects most municipalities that are down-
stream of future works, and as this will minimize the dependence of water coming
from the Cantareira System to supply the region.

References

1. Kutzner, C.: Earth and Rockfill Dams: Principles for Design and Construction. Routledge
(2018)
2. Reynolds, N.Y.: City of the high dam: Aswan and the promise of postcolonialism in Egypt.
City Soc. 29(1), 213–235 (2017)
3. De Chatel, F.: Water Sheikhs and Dam Builders: Stories of People and Water in the Middle
East. Routledge (2017)
4. Bidoglio, G., Berger, M., Finkbeiner, M.: An environmental assessment of small hydropower
in India: the real costs of dams’ construction under a life cycle perspective. Int. J. Life Cycle
Assess. 24(3), 419–440 (2019)
5. England, M.I.: India’s water policy response to climate change. Water Int. 43(4), 512–530
(2018)
6. Garchitorena, A., Sokolow, S.H., Roche, B., Ngonghala, C.N., Jocque, M., Lund, A., Andrews,
J.R.: Disease ecology, health and the environment: a framework to account for ecological and
socio-economic drivers in the control of neglected tropical diseases. Philos. Trans. Roy. Soc.
B Biol. Sci. 372(1722), 20160128 (2017)
7. Zhang, L.M., Xu, Y., Jia, J.S.: Analysis of earth dam failures: a database approach. Georisk
3(3), 184–189 (2009)
8. Junior, A.M.L.: Environmental licensing and cultural heritage: the licensing aspects of Serro
Azul dam in Pernambuco. 68 2016. f. Dissertation (Training Course of Cultural Managers of
the Northeastern States)—Federal University of Bahia, Recife (2016)
9. Salgado, A.A.R., Santos, L.J.C., Paisani, J.C. (eds.).: The Physical Geography of Brazil:
Environment, Vegetation and Landscape. Springer (2019)
10. Dantas, E.W.C., Silva, J.B.D.: Brazilian urban geography: from a diachronic analysis to the
most cited works in the last decade. Confins. Revue franco-brésilienne de géographie/Revista
franco-brasilera de geografia 38 (2018)
11. Cano, W., Brandão, C.A.: A Região metropolitana de Campinas. Urbanização, Economia,
Finanças e Meio Ambiente. Campinas: Editora da Unicamp (2002)
12. Minuzzi, R.B., Sediyama, G.C., Barbosa, E.D.M., Melo Júnior, J.C.F.: Climatologia do compor-
tamento do período chuvoso da região sudeste do Brasil. Revista Brasileira de Meteorologia
22(3), 338–344 (2007)
13. Durst, J., Neumann, L., Smith A.: Contested development: the Belo Monte Dam, Brazil. In:
Water, Climate Change and the Boomerang Effect, pp. 84–99. Routledge (2018)
14. Poff, N.L., Olden, J.D.: Can dams be designed for sustainability? Science 358(6368), 1252–
1253 (2017)
234 G. G. de Oliveira et al.

15. Cancellara, M.A.: Two Bridges basic dam project: Final report. Paulinia: Projectus (2013). 184
f
16. Cancellara, M.A.: Final report: Basic quarry dam project. Paulinia: Projectus (2013). 190 f
17. EIA—Pedreira Two Dams and Bridges. Volume V—impacts and environmental pro-grams
take 1—text. Department of Sanitation and Water Resources Department of Water and Power
(2015)
18. EIA—Study Quarry dams and Two Bridges environmental impact. Volume III—diagnosis of
biotic take 1—text (part 1). Department of Sanitation and Water Resources Department of
Water and Power (2015)
19. Sousa, V.S., de Freitas, V.M.: Revisão Teórica Sobre os Desastres Da Mineração Brasileira
Incididos em Mariana-Mg (2015) E Brumadinho-Mg (2019). In: Anais Colóquio Estadual de
Pesquisa Multidisciplinar (ISSN-2527-2500) & Congresso Nacional de Pesquisa Multidisci-
plinar (2019)
20. Freitas, C.M.D., Barcellos, C., Asmus, C.I.R.F., Silva, M.A.D., Da Xavier, D.R.: Samarco
em Mariana à Vale em Brumadinho: desastres em barragens de mineração e Saúde Coletiva.
Cadernos de Saúde Pública 35, e00052519 (2019)
21. Almeida, I.M.D., Jackson Filho, J.M., Vilela, R.A.D.G.: Razões para investigar a dimensão
organizacional nas origens da catástrofe industrial da Vale em Brumadinho, Minas Gerais.
Brasil. Cadernos de Saúde Pública 35, e00027319 (2019)
22. dos Santos, D.A., Martins, J.E.R., Rodrigues, M.R.: Breve ensaio sobre a responsabilidade
civil do Estado no contexto dos desastres ambientais em Mariana e Brumadinho. Revista do
Curso de Direito da Uniabeu 12(1), 158–161 (2019)
23. Padilha, R., Iano, Y., Monteiro, A.C.B., Arthur, R., Estrela, V.V.: Betterment proposal to
multipath fading channels potential to MIMO systems. In: Brazilian Technology Symposium,
pp. 115–130. Springer, Cham (2018)
24. França, R.P., Iano, Y., Monteiro, A.C.B., Arthur, R.: Improvement for channels with multipath
fading (MF) through the methodology CBEDE. In: Fundamental and Supportive Technologies
for 5G Mobile Networks, pp. 25–43. IGI Global (2020)
25. França, R.P., Iano, Y., Monteiro, A.C.B., Arthur, R.: A proposal of improvement for transmis-
sion channels in cloud environments using the CBEDE methodology. In: Modern Principles,
Practices, and Algorithms for Cloud Security, pp. 184–202. IGI Global (2020)
26. França, R.P., Iano, Y., Monteiro, A.C.B., Arthur, R.: Improvement of the transmission of
information for ICT techniques through CBEDE methodology. In: Utilizing Educational Data
Mining Techniques for Improved Learning: Emerging Research and Opportunities, pp. 13–34.
IGI Global (2020)
27. França, R.P., Peluso, M., Monteiro, A.C.B., Iano, Y., Arthur, R., Estrela, V.V.: Development
of a kernel: a deeper look at the architecture of an operating system. In: Brazilian Technology
Symposium, pp. 103–114. Springer, Cham (2018)
28. Monteiro, A.C.B., Iano, Y., França, R.P., Arthur, R., Estrela, V.V.: A comparative study between
methodologies based on the hough transform and watershed transform on the blood cell count.
In: Brazilian Technology Symposium, pp. 65–78. Springer, Cham (2018)
29. Monteiro, A.C.B., Iano, Y., França, R.P., Arthur, R.: Methodology of high accuracy, sensitivity
and specificity in the counts of erythrocytes and leukocytes in blood smear images. In: Brazilian
Technology Symposium, pp. 79–90. Springer, Cham (2018)
Numerical Simulation of Hydrodynamic
Conditions in Rivers Facing Extreme
Events Due to the “El Niño” Phenomenon

Gabriela Alvarez , Alvaro Moreno , Emanuel Guzmán ,

and Sissi Santos

Abstract Disasters caused by floods are increasingly common around the world,
such as the case in the western part of North and South America, due to the presence
of the “El Niño” phenomenon, which causes an increase in rainfall and flow rates of
the rivers causing floods. For this reason, this present work has carried a numerical
simulation of flooding in the face of extreme events such as “El Niño,” with the
purpose of validating a hydrodynamic model that subsequently allows to predict
the magnitude of a future event, for this, hydrological data of average flows and
instantaneous maximums over a period of 40 years has been used. The hydrological
information was processed using the Gumbel method to obtain simulation flows for
return periods of 2, 5, 10, 50, 100 and 500 years. As a result, an approximation of 75%
of the model was obtained in comparison with the historical event. Additionally, an
increase of 103% in water elevation was found for discharge with a return period of
500 years, compared with the event of March 27, 2017, which represents a discharge
with a return period of 36 years.

Keywords Flood · Hydraulics · “el niño” phenomenon

G. Alvarez · A. Moreno (B) · E. Guzmán · S. Santos

Universidad Peruana de Ciencias Aplicadas, Av. Prolongación Primavera 2390, Santiago de
Surco, Lima, Peru
e-mail: [email protected]
G. Alvarez
e-mail: [email protected]
E. Guzmán
e-mail: [email protected]
S. Santos
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 235
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_23
236 G. Alvarez et al.

1 Introduction

Floods caused by river overflows are recurring problems worldwide and are related
to increased rainfall, structural failures in dams or protection structures, as well as
the occurrence of extreme phenomena such as the “El Niño” phenomenon (FEN).
It affects the western part of North and South America, causing large economic and
human losses. Facing this situation, the authorities carry out contingency plans to
mitigate the negative impacts caused by these phenomena.
Several studies and investigations have been carried out in order to study the floods
and prevent the disasters they cause [1–13], one of the techniques used for this is
hydraulic modeling where programs such as IBER [5], HEC-RAS [13], FLO2D [11],
among others are used.
In order to develop the modeling, it is common to use digital terrain elevation
models (DEM) to characterize the topographic configuration of the study area. The
DEM is a computational representation of the triangular irregular networks (TIN)
of the surface that can be obtained from satellites that take images obtaining from
horizontal (x, y) and vertical (z) vectors in such a way that it generates the digital
terrain models with different resolutions [14, 15]. From the DEM obtained, TINs
are created, which are a necessary requirement for hydraulic simulation, since from
this it is possible to generate the mesh that covers the study area, the river cross-
sections, among others [6]. A joint work of data collected in the field and a digital
model facilitates and enriches the information extracted from the satellite at the time
of the extraction of cross-sectional sections of the river for 1D or 1D/2D modeling
[16–18]. The hydrological data (rain, river or both) implemented in the models help
determine flood plains that occur around the river or within the study basin. For the
calibration of the generated model, it is possible to make use of statistical parameters
that start from physical properties of the place or the digital calibration is carried out,
which consists in the comparison of models of different satellites to obtain the best
available model with the least possible error and closely related [10, 16]. On the other
hand, to carry out the validation of the model there are different studies that compare
the numerical models obtained in a combination of the digital elevation models and
the hydraulic information of historically observed events, comparing mainly their
physical differences. Maswood & Hossan [14] use the comparison of the bathymetry
water levels of the flooded days and the numeric values obtained, while Abdul Kadir
[1] compares the longitudinal profiles observed with the numerical ones, and in the
key points are compared to verify if they have a significant difference. In addition,
Surwase [17] compares flood plains by changing roughness parameters, proving that
for a higher Manning n value, an increase in the flood area will be generated.
Therefore, the use of numerical modeling is an important tool since it allows eval-
uating real and hypothetical scenarios, considering return periods of up to 500 years,
which allow obtaining extreme events and assessing and identifying the most vulner-
able areas to flood risks, which allows competent authorities to prepare management
plans for these risks. The present investigation provides a hydrodynamic model in
Numerical Simulation of Hydrodynamic Conditions … 237

HEC-RAS applied in Peru, to simulate possible future extreme events caused by “El
Niño” phenomenon.

2 Area of Interest

The study area is located in the northern part of Perú, in the department of Piura
and includes the section between the Los Ejidos dam to the Bolognesi Bridge
(Fig. 1). Likewise, other bridges are shown within the study area, such as the Andrés
Avelino Cáceres Bridge, Eguiguren Bridge, Sánchez Cerro Bridge, and San Miguel
Suspension Bridge, being its distribution with downstream direction.

Fig. 1 Study section from Los Ejidos dam to Bolognesi Bridge

238 G. Alvarez et al.

The district of Castilla and the district of Piura represent a part of the urban area
of Piura. In these areas, there are homes, schools, parks, etc. Points of importance
due to flooding are places such as the main square of Piura and the open plaza of
Piura due to its proximity to the river.
The Sánchez Cerro Bridge hydrometric station is administered by the National
Water Authority (ANA) [4] and is in the district of Miraflores, at a height of 34 m.a.s.l
in the coordinates of latitude = 05° 11 37.15 , longitude = 80° 37 24.39 . The station
records average normal flow conditions of 752.6 m3 /s, while in extreme events it
reaches 4424.0 m3 /s. In the structure of the Sánchez Cerro Bridge, its cross-section
supports a maximum avenue of 2000 m3 before overflowing; this flow is recorded
and usually exceeded when natural phenomena such as the “El Niño” phenomenon
occur.

3 Methodology

The work began with the collection of data from different official institutions such
ANA [4], Regional Emergency Operations Center of Piura (COER-PIURA) [3]
and the Special Project Chira-Piura (PECHP) [12], from where the information of
maximum daily, daily average and hourly discharged was obtained. For the morpho-
logical characterization of the area, information has been used on a digital terrain
elevation model (DEM), obtained from the SPOT-7 satellite, with a resolution of 6
× 6 m, which was collected from National Aerospace Research and Development
Commission (CONIDA) [2]. DEM has been processed, the land data and the corre-
sponding cross-sections of the river were obtained and exported to HEC-RAS for
simulation.
For the generation of the model, after adding the input data in the HEC-RAS, in
the geometry section, the mesh was defined according to the DEM resolution. Also,
the upstream edge conditions were defined, where the hydrograph corresponding to
the date and time of the event was entered, while the average slope of the river in the
study area was placed downstream.
Subsequently, the values of Manning n were chosen according to the river’s char-
acteristics as observed in the study area, since it is a coefficient that represents the
roughness of the riverbed and was used to calibrate the model, also, when modifying
that parameter, it will change substantially the results of the model. To have a better
fit, the calculation intervals of the model were determined according to the Courant
stability criterion, which is indicated in the HEC -RAS 5.0.7 manual. The hydrody-
namic simulation was visualized in RAS Mapper, and the validation was carried out
by means of a comparison of the flood area with the high-resolution digital image
of the same date and time selected. Finally, an extreme flood event was modeled
with the data obtained of discharge for 500 years of the return period, where the
flooded area is visualized, and this allowed to generate control and prevention plans
in the vulnerability area to these disasters. This methodology was adapted according
to Surwase for the respective purposes [17]. The extremes discharge values were
Numerical Simulation of Hydrodynamic Conditions … 239

obtained using Gumbel Distribution, which passed the Kolmogorov-Smirnov test to

verify that it has a normal distribution, and the flows for the return periods of 10, 20,
50, 100 and 500 years.

4 Simulations

The DEM for the required study area was adjusted and modified since it has elevation
data. For the study, the contour lines were extracted at an interval of every 3 m to
provide the necessary precision for the model. The contours achieved to consider
the topography of the terrain, from them a tin (irregular triangulation networks)
was created, showing the different elevations in the area. Then, the cross-sections
were created, following the axis alignment and the defined river limits. The terrain
geometry was entered in the HEC-RAS in the “RAS Mapper” window to display the
respective elevations. In the “Geometric Data” window, the calculation mesh within
the flood area was defined, which must contain a cell size that conforms to the DEM
resolution for best results. Different boundary conditions were considered for the
calculation mesh, for example, for the upstream the respective input hydrograph was
entered and for the downstream, the slope of the land was entered, which was 1.2%.
Finally, for the calculations of computation interval, the cell parameters and speed of
the unsteady flow for different levels of river water with a Courant coefficient equal
to 1 were considered, according to the equations defined in the HEC-RAS Manual
5.0.7.
Simulations were developed for the event of March 27, 2017, date that a flood
event was reported due to the FEN, in this stage, it was proceeded to calibrate the
model and validate with a satellite image. In the second stage, an extreme event was
modeled for the discharge rate obtained for a return period of 500 years, so the flood
hazard area was obtained.

4.1 Event Simulation: FEN March 27th, 2017

For the simulation, a Manning roughness coefficient of 0.035 was considered as

roughness for dredged material with short grass and some weeds [7] and the hydro-
gram of March 27th from 00:00 h until 21:00 h taken from the Ejidos dam (Table 1;
Fig. 2).
The modeling result for March 27th, 2017 (Fig. 3b) shows a flooded area of
3.18 km2 , where two mainly affected areas such as the main square and the Open
Plaza Shopping Center are observed due to its large audience attendance.
Figure 2 shows the difference between (a) the observed flood, by calculating
the Normal Differential of Water Index NDWI index, obtained from the Sentinel-2
satellite, image from April 4th, 2017 and (b) the modeled flood of the DEM obtained
from CONIDA. For the main square, the observed event and the modeling have
240 G. Alvarez et al.

an area of 0.98 km2 and 0.74 km2 , respectively, this represents an approximation
of 75.5%. Also, for the Open Plaza, the observed event and the modeling have
an area of 0.30 km2 and 0.23 km2 , respectively, this represents an approximation
of 79.31%. In addition, results were obtained from the depths and average speeds
in areas such as the open plaza Piura, the main square, Sánchez Cero Bridge and
Bolognesi Bridge, which are consistent with the historical information of the “Niño
Costero” phenomenon of the year 2017 (FEN Costero 2017) (Table 2).

Table 1 Hourly discharge of March 27th, 2017 at the Ejido dam station. Source COER [8]
Ejidos dam
Time Flow m3 /s Time Flow m3 /s Time Flow m3 /s
00:00 2270 08:00 3095 16:00 3133
01:00 2325 09:00 3095 17:00 3076
02:00 2396 10:00 3264 18:00 3035
03:00 2501 11:00 3378 19:00 3005
04:00 2574 12:00 3468 20:00 2923
05:00 2864 13:00 3264 21:00 2850
06:00 2895 14:00 3216
07:00 3016 15:00 3174

Fig. 2 a Flood observed by the SENTINEL-2 satellite for April 4th, 2017 [9]. b The simulated
flood of March 27th, 2017
Numerical Simulation of Hydrodynamic Conditions … 241

For calibration, adjustments were made to both the digital elevation model and the
roughness coefficients, so that the results of the flood are more in line with reality, for
the calibration of our model. For this, several simulations of the scenario of March
27th, 2017 (FEN Costero 2017) are carried out, verifying that the information about
water levels is consistent with what happened historically.

Fig. 3 Flood simulation of the city of Piura for discharge with a return period of 500 years

Table 2 Results of depth and speed obtained from modeling in HEC-RAS

Location Depth (m) Speed (m/s)
Open Plaza Piura 2.57 1.26
Main square 1.80 0.099
Sánchez Cerro bridge 3.10 4.25
Bolognesi bridge 2.18 9.90
242 G. Alvarez et al.

Table 3 Extremes discharge obtained for different return periods

T (years) Q (m3 /s)
10 2335.29
20 2950.13
50 3745.97
100 4342.34
500 5720.47

Table 4 Depth and speed for discharge for the return period of 500 years
Location Depth (m) Speed (m/s)
Open Plaza Piura 3.55 2.42
Main square 4.72 1.54
Sánchez Cerro bridge 7.37 10.20
Bolognesi bridge 3.80 11.75

4.2 Extreme Event Simulation with a 500 Year of the Return

Period

To achieve the extreme event with a return period of 500 years, the distribution of
Extreme Values Type I of the maximum instantaneous discharge of the Chira-Piura
project was made from 1971 until 2017 when the last El Niño phenomenon occurred.
Once the probability of exceedance for this distribution was realized, the flows were
calculated for the return periods of 10, 20, 50, 100 and 500 years. To do this, the
flows were obtained as shown in Table 3.
Then, the modeling of the extreme event was performed for a 500-year return
period, where an increase in the extent of the flood could be seen (Fig. 3). The results
of the depths and speeds with a flow rate for the return period of 500 years are
presented in Table 4.

5 Results

The simulation performed with HEC-RAS shows an accuracy between 75.5 and
76.67% respect Sentinel Satellite image for April 4th, 2017. For 500 years of the
return period, water levels show a significant increase of 103% respect the last FEN
2017, it will generate a greater impact on vulnerable areas and the population.
The simulation carried out for the extreme event provides data on the most affected
areas to be damaged by flooding, thus to mitigate the impact of this situation is neces-
sary to develop a contingency, that involves building hydraulic protection structures,
dredging plans for water depth that can control these effects.
Numerical Simulation of Hydrodynamic Conditions … 243

Water surface elevation (WSE) vary along the stretch of the Piura River, obtaining
values between 55 and 63 m.a.s.l (0–15 m water depth), due to the geometry of its
channel and the discharge obtained and simulated for a return period of 500 years.
The highest WSE found were located by the Open Plaza Piura, being an average
of 61 m.a.s.l, since it is upstream and the riverbed was narrower and while going
downstream the channel decreases.
The velocity found for a 500-year return period event would vary between 0 and
4 m/s in public places such as sidewalks, streets, the main square, open plaza, etc.,
and the velocity would increase in the riverbed becoming values between 10 and
17 m/s as in the Sanchez Cerro Bridge and Bolognesi Bridge which predicted speeds
of 10.2 and 11.75 m/s, respectively.

6 Conclusions

The main contribution of the work is to make a model as close as possible to the
events that occurred historically. As a result of the investigation, a model with a 75%
approximation to the facts was obtained, in this way, simulations can be carried out
for possible future extreme events of different return periods, as can be done in the
simulation for 500 years, where an increase of the flood area of 80% was obtained in
comparison with a simulation for 10 years, which represents a normal condition of
the river. Also, based on the results of this study, once the most vulnerable areas are
found, risk maps and contingency plans will be developed, such as the implementation
of riverine defense structures to mitigate the economic and human impact caused by
floods.

References

1. Abdul Kadir M.A., Abustan I., Abdul Razak M.F.: 2D flood inundation simulation based on a
large scale physical model using course numerical grid method. Int. J. GEOMATE 17, 230–236
(2019). https://doi.org/10.21660/2019.59.icee17
2. Agencia Espacial del Perú. (n.d.). Comisión Nacional de Investigación y Desarrollo Aeroes-
pacial. Lima: CONIDA. Available in: http://www.conida.gob.pe/
3. Agencia Peruana de Noticias. (2014). Inauguran en Piura Centro de Operaciones de Emergencia
Regional. Piura. Available in: https://andina.pe/agencia/noticia-inauguran-piura-centro-operac
iones-emergencia-regional-499055.aspx
4. Autoridad Nacional del Agua (ANA). (n.d.). Autoridad Nacional del Agua. Lima: ANA.
Available in: https://www.ana.gob.pe/
5. Blade, E., et al.: Iber: herramienta de simulación numérica del flujo en ríos. Revista internacional
de métodos numéricos para cálculo y diseño en ingeniería 30(1), 1–10 (2014). https://doi.org/
10.1016/j.rimni.2012.07.004
6. Chakraborty, S., Biswas, S.: Application of geographic information system and HEC-RAS in
flood risk mapping of a catchment. 33, 215–224 (2020). https://doi.org/10.1007/978-981-13-
7067-0_17 (unpublished)
7. Chow, V.T.: Hidráulica de Canales Abiertos. McGraw-Hill, Bogotá, Colombia (2004)
244 G. Alvarez et al.

8. COER.: Reporte de caudales del Río Piura para el 27/03/2017”, Centro de Operaciones de
Emergencia Regional—Piura. Marzo (2017)
9. COPERNICUS (n.d). Copernicus Open Access Hub.: Copernicus, Europe’s eyes on Earth.
Available in: https://scihub.copernicus.eu/dhus/#/home
10. Farooq, M., Shafique, M., Shahzard Khattak, M.: Flood hazard assessment and mapping
of River Swat using HEC-RAS 2D model and high-resolution 12-m TanDEM-X DEM
(WorldDEM). Nat. Hazards 97, 477–492 (2019). https://doi.org/10.1007/s11069-019-03638-9
11. FLO-2D Software, INC. (n.d.). FLO-2D. Florida: FLO-2D Software, INC. Available in: https://
www.flo-2d.com/
12. Gobierno Regional de Piura.: Proyecto Especial Chira Piura. Piura (2019). Available in: http://
www.chirapiura.gob.pe/
13. Hydrologic Engineering Center (HEC). (n.d.). HEC-RAS. Washington: HEC. Available in:
https://www.hec.usace.army.mil/software/hec-ras/
14. Maswood, M., Hossain, F.: Advancing river modelling in ungauged basins using satel-lite
remote sensing: the case of the Ganges-Brahmaputra-Meghna basin. Int. J. River Basin Manage.
14, 103–117 (2016). https://doi.org/10.1080/15715124.2015.1089250
15. Mohamad Faudzi, S.M., Abustan, I., Abdul Kadir M.A., Wahab, M.K., Abdul Razak M.F.: Two-
dimensional simulation of Sultan Abu Bakar Dam release using hec-ras. Int. J. GEOMATE 16,
124–131 (2019). https://doi.org/10.21660/2019.58.icee18
16. Patel, D.P., Ramirez, J.A., Srivastava, P.K., Bray, M., Han, D.: Assessment of flood inun-dation
mapping of Surat city by couple 1D/2D hydrodynamic modeling: a case application of the new
HEC-RAS 5. Nat. Hazards 89, 93–130 (2017). https://doi.org/10.1007/s11069-017-2956-6
17. Surwase, T., SrinivasaRao, G., Manjusree, P., Begum, A., Nagamani, P.V., JaiSankar, G.: Flood
inundation of Mahanadi River, Odisha during September 2008 by using HEC-RAS 2D Model,
pp. 851–863 (2018). https://doi.org/10.1007/978-3-319-77276-9_77
18. Vozinaki, A.K., Morianou, G.G., Alexakis, D.D., Tsanis, I.K.: Comparing 1D-and com-bined
1D/2D hydraulic simulations using high resolution topographic data, the case study of the
Koiliaris basin, Greece. Hydrol. Sci. J. 7, 642–656 (2016). https://doi.org/10.1080/02626667.
2016.1255746
Dynamic Amplification Factor Proposal
for Seismic Resistant Design of Tall
Buildings with Rigid Core Structural
System

Eder Quezada , Yaneth Serrano , and Guillermo Huaco

Abstract Currently, there is an increase in the demand for tall buildings in the city
of Lima. This research proposes to reduce the dynamic amplification factor through
the seismic design of tall buildings based on the requirements of Peruvian code
considering that they are regular in plan and height. Minimum base shear values
according to the comparison of static seismic shear and dynamic shear from the
spectral modal analysis were reviewed for cases of buildings larger than 120 m.
The study of 28 reinforced concrete buildings was proposed, with different heights
- varying from 24 to 36 floors, with different floor configurations, as well as the
arrangement of the walls considering as a rigid core structural system. Additionally,
the characteristics of the materials, the loads and combinations were defined. The
responses of these buildings were determined by the response spectrum analysis
(RSA) and then compared with those obtained by the lineal response history analysis
(LRHA), for the last analysis, five Peruvian seismic records were used and scaled to
0.45 g. The seismic responses of the LRHA procedure were taken as a benchmark.
The result of this study is the analysis and proposal of the C/R factor for high-rise
buildings, as well as obtaining the base shear and drift verification. Minimum base
shear values can be reduced for high or long-term buildings, being regular in plan
and height.

Keywords High-Rise buildings · Long natural period · Minimum seismic base

shear force · Rigid core

E. Quezada · Y. Serrano · G. Huaco (B)

Peruvian University of Applied Sciences UPC, Av. Prolongación Primavera 2390, Santiago de
Surco, Lima, Peru
e-mail: [email protected]
E. Quezada
e-mail: [email protected]
Y. Serrano
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 245
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_24
246 E. Quezada et al.

1 Introduction

A study by the United Nations estimates that by the end of 2050, approximately 64%
of developing countries and 86% of the population of developed countries will live
in cities. To solve the increase in population density in urban areas, these must be
expanded, or high-rise buildings built [1]. For this reason, in the last decade, Lima has
experienced an increase in the verticality of its buildings. Therefore, it is necessary
to use a structural system resistant to wind and earthquake forces. Based on this,
several structural systems have been developed for high-rise buildings. The rigid
core structural system constitutes such a solution and offers the advantage of faster
construction, flexible architecture and open space availability [1–8]. To design these
structures that resist earthquake loads, there are several numerical methods [2]. One
of them is the response spectrum analysis (RSA) which is widely used in various
seismic codes such as ASCE 7, UBC-97, FEMA-356 and ATC-40 to determine design
forces and displacement demands [3]. In addition, must be taken into consideration
that tall buildings are complex due to the numerous structural components and several
vibration modes [4]. In the case of the Peruvian Code, this method is related to the
Equivalent Lateral Force (EFL) by means of the dynamic amplification factor, which
guarantees that the current base is not less than 80% of the value calculated by EFL
for regular structures. To determine the shear force by means of the EFL, the Standard
establishes a minimum C/R factor. The subject of this analysis is to corroborate this
parameter in the case of tall buildings by comparing inelastic responses with the
LRHA. In addition, few studies prove that the RSA established by the standard is
adequate for high-rise buildings. The advantage of using an adequate value for this
factor will avoid the design of the building with unnecessary robust elements that
meet the seismic requirements, which would significantly increase the cost of such
infrastructure since it would cause a lack of investment due to the cost overrun.
The methodology adopted in this work is proposed by our authorship and is
described as follows:
1. The case studies are defined taking into consideration the following criteria:
use of the building, dimensions, predominant material, mezzanine height, slab
thickness and variation in compressive strength.
2. Selection of five seismic acceleration records according to soil type and moment
magnitude range.
3. Proceed with the EFL and RSA, as established by the Seismic Resistant Design
Code E.030.
4. The inelastic responses of the spectral modal dynamic analysis are compared with
the linear time-history analysis to determine the adjustment of the C/R factor.
5. The results analysis is carried out and an adjusted C/R value is proposed for
high-rise reinforced concrete building.
Dynamic Amplification Factor Proposal for Seismic Resistant … 247

2 Methodology

2.1 Description of the Buildings

Twenty-eight tall buildings located in Lima, Peru, are considered in this study (Fig. 1).
The number of floors of the cases analyzed is 24, 26, 28, 30, 32, 34 and 36 floors.
Fourteen of these buildings correspond to a square plan and the other fourteen to
a rectangular plan. Likewise, the structural configuration was modified considering
frames and post-tensioned slabs.
Each beam and column was modeled as an elastic frame element. The slabs and
the cutting walls are modeled with a thin shell elastic element and the foundation was
idealized as recessed base support. For all buildings, the columns on each floor resist-
less than 30% of the total lateral force; therefore, the structural system is considered
as structural walls in accordance with Code E.030. The P −  effects are included
in all the methods [5] (Table 1).

Fig. 1 Floor plans and 3D models. a PC-POR, b PR-POR, c PC-POS y d PR-POS

248 E. Quezada et al.

Table 1 Coding of the square and rectangular floor model

Number of stories Walls and Columns Post-tensioned slab
PC-POR PR-POR PC-POS PR-POS
Square Rectangular Square Rectangular
24 N24-PC-POR N24-PR-POR N24-PC-POS N24-PR-POS
26 N26-PC-POR N26-PR-POR N26-PC-POS N26-PR-POS
28 N28-PC-POR N28-PR-POR N28-PC-POS N28-PR-POS
30 N30-PC-POR N30-PR-POR N30-PC-POS N30-PR-POS
32 N32-PC-POR N32-PR-POR N32-PC-POS N32-PR-POS
34 N34-PC-POR N34-PR-POR N34-PC-POS N34-PR-POS
36 N36-PC-POR N36-PR-POR N36-PC-POS N36-PR-POS

2.2 Selected Seismic Records

In this study, a set of five seismic records was selected and the information was
obtained from the Japanese Peruvian Center for Seismic Research and Disaster Miti-
gation (CISMID) [7]. The selection of these registers took into consideration the type
of soil where the building is to be cemented (Very rigid soil) and with magnitude in
the range of 6–8.5-moment magnitude. In addition, these were scaled to 0.45 g cm/s2
(Fig. 2).

600

500 SCO
AcceleraƟon (cm/s2)

400 PQR_ANC

300 PQR66

200 TAC2005

100 MOL74

0
0.5 1 1.5 2 2.5 3 3.5 4
Period (s)

Fig. 2 The response spectrum of selected earthquakes with a damping of 5%

Dynamic Amplification Factor Proposal for Seismic Resistant … 249

2.3 Methods of Analysis for Earthquake Resistant Design

2.3.1 Equivalent Lateral Force Procedure (ELF)

The ELF procedure in Code E.030 is adopted in this study [6]. This method represents
the seismic solicitations as a set of lateral forces that act in the center of mass of each
level of the building. It consists of determining five seismic parameters that are: zone
factor (Z), soil factor (S), use factor (U), seismic amplification factor (C), seismic
force reduction coefficient (R). The determination of the shear force at the base of
the structure for each direction considered is calculated as follows:
ZU C S C
V = · P; ≥ 0.11 (1)
R R
This methodology cannot be applied to structures of reinforced concrete bearing
walls for buildings with a height greater than 30 m. However, it is necessary to carry
out this analysis in the design of reinforced concrete elements by what establishes
the norm as a minimum base shear force.

2.3.2 Response Spectrum Analysis (RSA)

The RSA method in Code E.030 is adopted in this study [6], where the elastic response
is combined for several vibration modes to obtain the total elastic response, using the
complete quadratic combination (CQC), which is then reduced by a force reduction
factor (R) to obtain the design forces of the structure.

3 Results

3.1 Comparison of Design Demands Between RSA

and LRHA

The following graphs represent the base shear through different types of analysis
in both directions. The first bar represents the RSA analysis with the C/R factor
calculated, while the second bar depicts the base shear obtained with the RSA analysis
with the minimum C/R factor proposed by the Peruvian Code. Addition-ally, the other
bars represent the base shear calculated from LRHA analysis with the seismic records
mentioned above (Figs. 3, 4, 5, 6, 7, 8, 9 and 10).
250 E. Quezada et al.

The shear forces calculated from the LRHA are significantly lower than those
calculated from RSA established by the Peruvian Standard of Seismic-Resistant
Design E.030, only a few exceptions for the seismic record of Surco1974 for the
N34PC_POR model. Therefore, it can be deduced that the C/R value is lower than
what is established in the regulations. The results of the LRHA were reduced by a
factor R. There are greater requests in terms of cut in square and rectangular plants
of 36 levels.

2500

2000 DINX
C/R=0.089 C/R=0.083
C/R=0.13 C/R=0.0904 C/R=0.075
C/R=0.116
Base Shear(tonf)

C/R=0.104
DINX_DIS
1500
PQR66X

1000 PQRANCX

SCOX
500
MOL74_X

0
TAC2005_X
24 26 28 30 32 34 36
Number of stories

Fig. 3 Base Shear for PC-POR in the X-direction

2500

DINY
2000
DINY_DIS
Base Shear (tonf)

C/R=0.117 C/R=0.104 C/R=0.078 C/R=0.073

C/R=0.093 C/R=0.084 C/R=0.064
PQR66Y
1500
PQRANC_Y

1000 SCO_Y

MOL74_Y
500 TAC2005_Y

0
24 26 28 30 32 34 36
Number of stories

Fig. 4 Base Shear for PC-POR in the Y-direction

Dynamic Amplification Factor Proposal for Seismic Resistant … 251

2500

2000
DINX
Base Shear (tonf)

DINX_DIS
1500
C/R=0.023
C/R=0.018 PQR66X
C/R=0.021
1000 C/R=0.026 C/R=0.012 PQRANCX
C/R=0.013 C/R=0.02
SCOX
500
MOL74_X

0 TAC2005_X

24 26 28 30 32 34 36
Number of stories

Fig. 5 Base Shear for PC-POS in the X-direction

2000

1500 C/R=0.022 DINY

Base Shear (tonf)

C/R=0.017
C/R=0.02 DINY_DIS
C/R=0.016
C/R=0.032
C/R=0.02 PQR66Y
C/R=0.024
1000
PQRANC_Y

SCO_Y
500
MOL74_Y

TAC2005_Y

0
24 26 28 30 32 34 36
Number of stories

Fig. 6 Base Shear for PC-POS in the Y-direction

2000

1500 C/R=0.022 DINY

Base Shear (tonf)

C/R=0.017
C/R=0.02
C/R=0.016 DINY_DIS
C/R=0.032
C/R=0.02
1000 C/R=0.024 PQR66Y

PQRANC_Y

SCO_Y
500
MOL74_Y

TAC2005_Y

0
24 26 28 30 32 34 36
Number of stories

Fig. 7 Base Shear for PR-POR in the X-direction

252 E. Quezada et al.

3000
C/R=0.113 C/R=0.1
C/R=0.088
2500 C/R=0.078 DINY
C/R=0.072 C/R=0.063
Base Shear (tonf)

C/R=0.052
2000 DINY_DIS

PQR66Y
1500
PQRANC_Y
1000 SCO_Y

500 MOL74_Y

TAC2005_Y
0
24 26 28 30 32 34 36
Number of stories

Fig. 8 Base Shear for PR-POR in the Y-direction

3500

3000
Base Shear (tonf)

2500 DINX
C/R=0.069
2000 DINX_DIS

PQR66X
1500 C/R=0.054 C/R=0.051
C/R=0.045 C/R=0.038
PQRANCX
C/R=0.05 C/R=0.04
1000
SCOX

500 MOL74_X

0 TAC2005_X
24 26 28 30 32 34 36
Number of stories

Fig. 9 Base Shear for PR-POS in the X-direction

3500
DINY
3000
DINY_DIS
Base Shear (tonf)

2500 C/R=0.042
C/R=0.087 C/R=0.075 C/R=0.064 C/R=0.049 C/R=0.042
2000 C/R=0.051 PQR66Y

1500 PQRANC_Y

1000
SCO_Y
500
MOL74_Y
0
24 26 28 30 32 34 36
TAC2005_Y
Number of stories

Fig. 10 Base Shear for PR-POS in the Y-direction

Dynamic Amplification Factor Proposal for Seismic Resistant … 253

The general procedure to determine the value of C/R for high-rise buildings
consists of computing ratios that are based on the base shear obtained by the RSA and
the LRHA. Then the ratios are multiplied by the C/R values obtained from the LRHA.
Finally, the most critical values were chosen as the final value (Tables 2 and 3).

Table 2 Summary of proposed C/R values in the X-direction

C/R proposed
Level Tx C/R calc. PQR 66 PQR70 SCO74 MOL74 TACNA
PC-POR 24 1.284 0.13 0.055 0.079 0.052 0.047 0.067
26 1.441 0.116 0.054 0.054 0.077 0.044 0.052
28 1.602 0.104 0.069 0.050 0.085 0.072 0.067
30 1.779 0.094 0.061 0.054 0.092 0.044 0.072
32 1.873 0.089 0.041 0.039 0.077 0.051 0.063
34 2.01 0.083 0.032 0.043 0.081 0.045 0.058
36 2.23 0.075 0.060 0.041 0.086 0.065 0.070
PC-POS 24 2.666 0.059 0.044 0.033 0.031 0.049 0.054
26 3.848 0.028 0.020 0.017 0.022 0.013 0.017
28 3.074 0.044 0.025 0.023 0.031 0.034 0.049
30 3.91 0.027 0.024 0.021 0.019 0.024 0.018
32 2.863 0.051 0.029 0.030 0.041 0.045 0.040
34 2.96 0.048 0.024 0.027 0.034 0.036 0.037
36 3.229 0.04 0.020 0.026 0.026 0.029 0.030
PR-POR 24 1.111 0.15 0.059 0.097 0.069 0.100 0.097
26 1.229 0.136 0.055 0.064 0.054 0.063 0.071
28 1.353 0.123 0.060 0.072 0.085 0.047 0.052
30 1.542 0.108 0.058 0.042 0.066 0.053 0.052
32 1.557 0.107 0.070 0.053 0.079 0.064 0.063
34 1.696 0.098 0.066 0.040 0.097 0.069 0.072
36 1.835 0.091 0.045 0.045 0.084 0.052 0.073
PR-POS 24 2.901 0.05 0.033 0.025 0.043 0.047 0.050
26 3.227 0.04 0.024 0.020 0.025 0.030 0.035
28 3.045 0.045 0.025 0.020 0.032 0.027 0.045
30 3.317 0.038 0.022 0.018 0.029 0.025 0.038
32 2.789 0.054 0.040 0.031 0.031 0.046 0.071
34 2.853 0.051 0.040 0.021 0.027 0.038 0.054
36 2.401 0.069 0.038 0.026 0.045 0.064 0.069
254 E. Quezada et al.

Table 3 Summary of C/R proposed values in the Y-direction

C/R proposed
Level Ty C/R calc. PQR 66 PQR70 SCO74 MOL74 TACNA
PC-POR 24 1.426 0.117 0.066 0.064 0.064 0.039 0.052
26 1.605 0.104 0.067 0.044 0.068 0.054 0.065
28 1.786 0.093 0.047 0.041 0.079 0.060 0.057
30 1.975 0.084 0.046 0.047 0.072 0.044 0.062
32 2.152 0.078 0.063 0.039 0.063 0.048 0.054
34 2.298 0.073 0.038 0.035 0.080 0.057 0.059
36 2.55 0.064 0.064 0.045 0.080 0.064 0.074
PC-POS 24 2.32 0.072 0.052 0.036 0.074 0.051 0.056
26 3.04 0.045 0.035 0.037 0.057 0.045 0.049
28 2.78 0.054 0.045 0.036 0.038 0.070 0.055
30 3.446 0.035 0.023 0.032 0.028 0.048 0.026
32 2.904 0.049 0.040 0.047 0.047 0.046 0.050
34 3.04 0.045 0.023 0.029 0.035 0.038 0.038
36 3.349 0.037 0.019 0.025 0.019 0.033 0.026
PR-POR 24 1.472 0.113 0.065 0.058 0.048 0.092 0.065
26 1.672 0.1 0.066 0.053 0.062 0.048 0.059
28 1.886 0.088 0.056 0.041 0.062 0.040 0.061
30 2.149 0.078 0.060 0.034 0.061 0.060 0.046
32 2.321 0.072 0.054 0.036 0.082 0.069 0.056
34 2.58 0.063 0.038 0.046 0.080 0.064 0.058
36 2.843 0.052 0.030 0.031 0.071 0.042 0.073
PR-POS 24 1.928 0.087 0.050 0.046 0.041 0.071 0.063
26 2.235 0.075 0.059 0.043 0.061 0.042 0.059
28 2.549 0.064 0.049 0.029 0.065 0.036 0.062
30 2.872 0.051 0.025 0.035 0.058 0.037 0.042
32 2.906 0.049 0.029 0.034 0.059 0.063 0.051
34 3.162 0.042 0.027 0.031 0.070 0.042 0.053
36 3.146 0.042 0.025 0.029 0.037 0.050 0.041

4 Discussions

This work incentive the design of buildings that meet the seismic requirements and is
economically feasible, besides that it was resolved the lack of information regarding
high-rise reinforced concrete buildings in Peru and investigates the advantages of
regular seismic buildings as seen in other international buildings such as the case of
Burj Khalifa.
Dynamic Amplification Factor Proposal for Seismic Resistant … 255

The range of periods found for the PC-POR and PR-POR buildings is between
1.5 and 3 s, while for the PC-POS and PR-POS structures their range corresponds to
2.5–4 s, being the largest period, the cross-section that is slenderer.

5 Conclusions

The value of C/R is analyzed in the case of tall reinforced concrete buildings by
comparing the design shears indicated by Peruvian seismic regulations with the base
shear obtained from an LRHA, which have been scaled to 0.45 g.
It was found lower drifts of the LRHA smaller than what is indicated in Code
E.030, except for specific cases, but all are less than 0.7%, which establishes said
norm. The dynamic amplification factor is sensitive to the C/R value. It was noticed
that for PC-POR and PR-POR buildings, considering the minimum C/R equal to
0.11, the shear does not need to be amplified for buildings smaller than 30 floors,
however, for larger buildings the factor of amplification would be in the order of
150%. On the other hand, for PC-POS buildings, it is not necessary to amplify the
shear, and for PR-POS it is amplified from the 32-story buildings, in the order of
250%.
The main finding of this paper proposes a minimum value of C/R less than 0.11
for each of the 4 studied cases of the high-rise building. It was obtained due to the
regularity in the plan and height of the building and the use of a rigid core structural
system. As for PC-POR buildings, it is established that C/R values are much lower
than 0.08 for buildings with up to 32 levels, for buildings with a higher number
of levels this value is close to 0.08; something similar happens for structures type
PR-POR. In addition, for PC-POS and PR-POS buildings, the value of C/R is less
than 0.08 at all levels. Two C/R values differentiated by the structural system are
proposed, C/R equal to 0.075 for PC-POS and PR-POS and for 0.10 for PC-POR
and PR-POR.
There is a pressing need to extend the research to study the implications of the
wind force due to the predominant effect for high-rise buildings such as this one
because of its high natural period. It is suggested to study it. Furthermore, the value
of R must be improved or studied, being able to be less than 6, for this, it is suggested
to perform a nonlinear response history analysis (NLRHA).

References

1. Rahman, A., Zaman, Q., Irshad, M.: Evaluation of Simplified Analysis Procedures for a High-
Rise Reinforced concrete core wall structure. Adv. Civil Eng. 16, 599–614 (2019)
2. Khy, K., Chintanapakdee, C., Wijeyewickrema, A.: Modified response spectrum analysis to
compute shear force in tall RC shear wall buildings. Eng. Struct. 180, 295–309 (2019)
256 E. Quezada et al.

3. Najam, F, Warnitchai, P.: A modified response spectrum analysis procedure to determine

nonlinear seismic demands of high-rise buildings with shear walls. Struct. Des. Tall Special
Build. 27 (2017)
4. Mehmood, T., Warnitchai, P., Suwansaya, P.: Seismic evaluation of tall buildings using a
simplified but accurate analysis procedure. J. Earthq. Eng. 22(3), 356–381 (2017)
5. Liu, Y., Kuang, J.: Spectrum-based pushover analysis for estimating seismic demand of tall
buildings. Bull. Earthq. Eng. 15, 4193–4214 (2017)
6. Peru Seismic Code E.030.: Seismic Design Requirements for Buildings, SENCICO, Peru (2018)
7. Accelerometers Network Database, Center for Earthquake Engineering Research and Disaster
Mitigation, CISMID, Lima Peru (2019)
8. Ali, M., Moon, K.: Advances in structural systems for tall buildings: emerging developments
for contemporary urban giants. Buildings 8, 104 (2018)
A Mechanical Development of a Dry Cell
to Obtain HHO from Water Electrolysis

Gustavo Salazar , Wilmer Solis , and Leonardo Vinces

Abstract This article proposes a mechanical development of a dry cell in order to

obtain HHO through water electrolysis. Calculations and technical specifications of
the materials used for implementation are supported by mathematical, physical and
chemical formulas and theories (Faraday´s Law, electrolysis process and mechanical
design). The importance of mechanical design is focused on achieving efficient use
of the energy provided to the cell that allows the H2 and O2 molecules to be separated
without overheating the cell, evaporating the water, loss of current due to the geom-
etry of the electrodes (Foucault Current). Moreover, choosing materials for proper
implementation and physical robustness is mandatory. In addition, the mechanical
design is not justified in different articles. Nevertheless, the mechanical design of the
cell and the efficiency in the production of HHO are related. Therefore, the mechan-
ical design and the calculations were performed, as well as the construction of the dry
cell to obtain HHO. The results of the implementation and production were placed
and compared with what theoretically the dry cell should produce from the law of
Faraday. Finally, the volumetric flow of HHO obtained was 2.70 L per minute. It
means a production efficiency of 98.68%. It is higher than the majority of the dry
cells.

Keywords HHO generation · Electrolysis · Dry cell · Faraday · Mechanical design

G. Salazar · W. Solis · L. Vinces (B)

Universidad Peruana de Ciencias Aplicadas, Av. Prolongación Primavera 2390, Santiago de
Surco, Lima, Peru
e-mail: [email protected]
G. Salazar
e-mail: [email protected]
W. Solis
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 257
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_25
258 G. Salazar et al.

1 Introduction

Currently, global energy demand has increased and will continue for 25% more by
2040, according to the International Energy Agency [1]. In many countries, most
of the energy comes from fossil sources like energy generated by coal, oil or its
derivatives. An example of this is the energy balance carried out in Peru by the
Ministry of Energy and Mines in 2016 recovered by the Latin American Energy
Organization [2]. The results of this study show that the main source of energy is
natural gas and oil (47% and 30%, respectively), both non-renewable and polluting
energy sources. Because of this, there is a need to use a renewable and less polluting
alternative fuel. Hydroxy gas or brown gas (HHO) is a clean and renewable source of
energy due to the absence of hydrocarbon compounds; therefore, carbon dioxide or
other potentially dangerous gases are not released after combustion, it only produces
water vapor [3], as shown in (1).

HHO(g) + Calor → H2 O(l) + Energy (1)

HHO is produced by dividing water into hydrogen and oxygen molecules through
electrolysis. An electrolyzer, also called a hydrogen generator, is necessary for this
process. Some works on this field have been developed, the ones mentioned below
are the most important.
Rusdianasari et al. [4] used a software capable of calculating the production of
HHO gas and the power consumed of a wet cell according to the number, dimensions
of the electrodes, the voltage and current provided. In addition, the authors arrived
and compared the value of the HHO flow produced with the value calculated by
the software. However, the authors did not justify the characteristics chosen for the
design of the wet cell test resulting in an efficiency of 89.13%.
Shah et al. [5] compare the energy efficiency, production of HHO, the maintenance
and safety considerations of the dry and wet cell. As a result, they conclude the dry
cell being better in all the fields described above with the same input conditions as
the cell wet. The authors described the components of the dry cell and the physical
characteristics of the mechanical system that influenced the production of HHO. In
addition, they present a detailed design of the dry cell. However, the design presented,
and the components used for implementation don´t present technical specifications
or mathematical calculations that justify them.
Nabil et al. designed in [6] several cells with different numbers of electrodes and
configurations. In addition, the authors show a table with all the designs made and
they compare the energy consumption, the production of HHO gas, the molarity used
and the temperature of each one. However, despite using Faraday’s first law in their
article, the authors do not compare the production flow of each mechanical design
with what the cell should theoretically produce. Moreover, they don’t demonstrate
the efficiency of each configuration. In addition, the measurements and the shape of
the electrodes are not justified.
A Mechanical Development of a Dry Cell to Obtain HHO … 259

Because of the cases seen before, it is proposed to make a design based on math-
ematical calculations, physical and chemical formulas (Faraday´s Law, electrolysis
process and mechanical design). Furthermore, the implementation of a correct selec-
tion of dry cell components. Finally, the production of HHO of the designed cell will
be compared with what should theoretically be produced from Faraday’s law.

2 Description of the Proposed Method

To have an overview of the dry hydrogen cell that was designed, the P&ID diagram
of the electrolysis process can be seen in Fig. 1.
The diagram shows the instrumentation to be used in the process and the parts
that comprise it. In the dry cell is where the separation of water molecules occurs.
Moreover, it is where the research will focus on. In order to obtain an efficient
mechanical design, the following phases shown in Fig. 2 must be followed.

2.1 Calculations of Preliminary Cell Dimensions

For the design of the dry cell, calculations had to be made based on the power
available for the process. In this case, we have a 12 V battery and the current will
be regulated at 20 A. From this data, we can configure the cell plates. In the same
way that series circuits can be created by connecting various chained elements to
each other and connecting the power supply only at the ends, it is also possible to do
the same with the electrodes. Nevertheless, with the difference that the connection
between them occurs through the electrolyte. The plates that remain unconnected are
called neutral or bipolar, these plates are not connected to any source of electricity
and are placed between two plates connected to different polarities. In this way, it
is achieved that in each separation created between the connected plates there is a
voltage equal to the total divided by the number of spaces, without the need to adjust
the voltage output of the source or use cables.
260 G. Salazar et al.

Fig. 1 P&ID diagram of the electrolysis process

Fig. 2 Block diagram of the proposed method

According to Gamez [7], to calculate the number of plates, the voltage between
electrodes and the total source voltage must be considered. The number of plates is
obtained by applying Eq. (2):

VT
VC = (2)
n−1

The n is the number of plates per group, VT the total cell voltage and VC the
calculated voltage between electrodes. The result of n will be 7, two of which will be
polarized (cation and anion) and the other five will be neutral. For greater production,
the following configuration will be chosen:

−NNNNN + NNNNN−

The N´s are the neutral plates between the polarizing plates (+ −). In addition, it
must be considered that the current delivered by the source will be divided into two
groups. However, having two negatively polarized plates, hydrogen production will
be higher because it has a positive ion.
The geometry chosen for the plates is circular since there will be no electrical
losses due to vertices of the electrode. These currents are called Foucault currents.
In certain cases, these currents can produce undesirable effects. For example, they
increase internal energy and temperature. A conductive material with angular edges
and corners has electrical losses (corner effects) as a result of the accumulation of
electromagnetic fields in said vertices caused by a non-uniform current density [8].
A Mechanical Development of a Dry Cell to Obtain HHO … 261

Moreover, the current that will give energy to the cell will be regulated with a PWM
wave, the current will be variable, and the current density will be non-uniform.
Gámez defines the current density as the amount of current that circulates per unit
area of an electrode, and is calculated using (3):

IE IE
dC = = (3)
SE NA × SE

where dC is the current density (A/m2 ), IE the intensity of the current flowing through
the electrode (A), SE the surface of the electrode (m2 ) and NA the number of cell
groups.
Assuming the same electrode surface, by increasing the current flowing through
the cell the production of gas will be increased, but also the voltage necessary to
circulate this current, so that the efficiency of the cell will decrease. To avoid this,
Gámez mentions that the idea is to maintain a flow of 1000 A/m2 , since, at a smaller
area, a higher voltage will be needed to circulate the same amperage, which will
generate more heat.
Having an electrode surface (SE) of 0.01 m2 and a circular geometry of the
electrode, an electrode radius equal to 56 mm is obtained. That is the theoretical
radius that the cell’s circular plate should have. However, for experimentation and
ease of machining, a plate with a radius of 75 mm will be implemented. In addition,
the excess area can serve as a heat sink for the cell.

2.2 Cell Design

The software called Inventor [9] was used to design the plates. Two types of plates
will be designed: neutral and polarized. The polarizing plates will have an extension
in one corner, as seen in Fig. 3 so that it is attached to the stud that will pass through
the cell to assemble it. This stud will be connected to a pole of the power supply.
However, the extension made on the plate will be with a rounded corner to avoid the
phenomenon of stray currents. The neutral plate will not be connected to any pole
of the power supply and will be fully circular as seen in Fig. 3.
For the separation between electrodes and to ensure the sealing of the same, nitrile
O-ring will be used. Due to the current that will pass through the cell, a separation
of 4–8 mm is recommended. Because of the current that will pass through the cell
and the high gas productions, the electrolyte will bubble and foam, appearing spaces
without liquid between the electrodes that will reduce the efficiency of the cell.
Consequently, 6.99 mm thick O-rings were chosen, standardized measurement of
this material [10]. To give greater physical strength to the cell, 18 mm thick acrylic
will be used. The design of these acrylic plates is seen in Fig. 4.
262 G. Salazar et al.

Fig. 3 Detailed plane of the polarized and neutral plate

Fig. 4 Detailed plane of acrylic plate

2.3 Cell Implementation

To carry out the implementation of the cell it was necessary to make a previous study
of the materials to be used. The 316 L stainless steel was used in the plates since the
dry hydrogen cell will be in constant contact with salts and water and this material
offers a high electrical conductivity for electrolysis (electrical resistivity: 70–78 ×
10–8  × m). In addition, it has a high resistance to corrosion due to its chemical
composition [11]. In the case of the O-rings, nitrile was used, since it is not high
priced, its working temperature is adequate (−20 to 90 °C) and it is very durable
[12]. For the electrolyte, KOH was chosen with a concentration of 10%, since it is
the best electrolyte for any application and at this concentration, the best results are
obtained. This is because it supports more varied conditions and leaves no residue.
Sodium hydroxide, on the other hand, generates a brown residue that clouds the
solution and soils the cell, progressively reducing its efficiency if it is not cleaned
periodically [13, 14]. Finally, to seal the cell, 18 mm thick transparent acrylic and
Teflon couplings were used to enter the solution and exit hydrogen. After choosing
A Mechanical Development of a Dry Cell to Obtain HHO … 263

Fig. 5 HHO dry cell assembled

all the materials, the cell assembly is carried out. Subsequently, tightness tests were
performed to avoid leaks. In Fig. 5, the fully assembled cell can be seen.

2.4 Cell Test According to Faraday´s Law

The electrolysis process is basically based on Faraday’s laws for electrolysis, thus
having both the first (4) and the second law that refers to the fact that the amount of
altered substance in an electrode directly depends on the electric current in Coulombs.
In addition, the mass deposited in an electrode is proportional to its molecular weight
or chemical equivalent.

Q = I ×t (4)

where Q is the necessary energy (C), I the current intensity supplied (A) and t the
time (s). Then, we proceed to calculate the number of chemical equivalents of a
substance deposited in an electrode with (5):
264 G. Salazar et al.

Q
n ◦ eqq = (5)
F
In this equation, “n° eqq” is the number of chemical equivalents of a substance
deposited in an electrode, which is directly proportional to “Q” (1200 C). The
necessary energy calculated with Faraday’s first law (4), divided between Faraday’s
constant, “F”. The chemical equivalent found with (5) must be transformed to mass
of H2 , considering the molecular weight of H2 (2 g/mol) and its valence electrons
(+2) we can obtain that its “eqq” is 1. With this calculation, the mass of H2 (m H2 )
can be transformed using the following relation (6):

1g
n ◦ eqq × = m H2 (6)
1eqq

Knowing the mass of hydrogen to be obtained per minute, 0.0124 g H2 , now it is

desired to pass this mass to volume, specifically to liters, for this the ideal gas equation
is used, replacing the values at normal conditions resulting in 152 mL per minute to
calculate the production of O2 we can use the proportionalities of the volume of H2
(V̇H2 ) and O2 (V̇O2 ) that have a ratio of 2: 1. Therefore, the volume of HHO (V̇HHO )
produced per compartment can be obtained. Since there are 12 compartments, the
equation is obtained in (7):
 
V̇HHO = V̇H2 + V̇O2 × 12 (7)

Assuming 100% efficiency, this value found in (7) is approximately 2.73 L/min.

3 Results

In order to obtain the production efficiency of the dry cell, we implemented the P&ID
shown in Fig. 1, using a rotameter as a flow meter. In the test, the rotameter indicated
an HHO flow rate of 2.7 L/min with a peak current of 20 A. This result gives us an
efficiency of 98.68% as shown in Table 1.

Table 1 Production comparison

Rusdianasari [4] Our work
HHO production (L/min) 0.1028 2.73
Efficiency 89.13% 98.68%
A Mechanical Development of a Dry Cell to Obtain HHO … 265

4 Conclusion and Future Work

Our development can increase efficiency by 9.55% compared to the cell of Rusdi-
anasari. Moreover, the mechanical design is just the beginning of this work. The
purpose of developing this dry cell is to generate electricity by using an engine
without any kind of gasoline.

References

1. International Energy Agency. https://www.iea.org/newsroom/news/2018/november/world-

energy-outlook-2018-examines-future-patterns-of-global-energy-system-at-a-t.html. Last
Accessed 21 Aug 2019
2. Ministerio de Energía y minas. http://www.minem.gob.pe/minem/archivos/BNE_2016.pdf.
Last Accessed 10 Act 2019
3. Choodum, N., Sangwichien, C., Yamsaengsung, R.: Optimization of a closed-loop HHO
production system for vehicles and houses. Environ. Progress Sustain. Energy 38(1), 268–277
(2019)
4. Rusdianasari, R., Bow, Y., Dewi, Y.: HHO gas generation in hydrogen generator using
electrolysis. Earth Environ. Sci. 258 (2019)
5. Shah, S., Ali, Z., Larik, J., Kaimkhani, A.: Comparative study of dry and wet cell for HHO
gas generation as a supplement fuel for I.C. Engine. Int. Conf. Comput. Math. Eng. Technol.
(2018)
6. Nabil, T., Mohamed, M., Dawood, K.: Enabling efficient use of oxy-hydrogen gas (HHO) in
selected engineering applications; transportation and sustainable power generation. J. Clean.
Prod. 237 (2019)
7. Gámez, D.: El hidrógeno y sus aplicaciones energéticas. INS La Ferrería, España (2010)
8. Montrose, M.: Time and frequency domain analysis for right angle corners on printed circuit
board traces, 551–556 (1998)
9. Autodesk Homepage. https://www.autodesk.com/products/inventor/overview. Last Accessed
28 Nov 2019
10. Parker Homepage, https://www.parker.com/literature/o_ring.pdf. Last Accessed 20 May 2019
11. King, M.: Water electrolyzers and the zero-point energy. Phys. Procedia 20, 435–445 (2011)
12. Komariah, L., Arita, S., Aprianjaya, F., Novaldi, M., Fathullah M.: O-rings material dete-
rioration due to contact with biodiosel blends in a dinamic fuel flow. J. Phys. Conf. Serie
(2019)
13. Streblau, M., Mihail, S., Aprahamian, B., Dimova, T.: The influence of the electrolyte param-
eters on the efficiency of the oxyhydrogen (HHO) generator. IEEE Electr. Apparatus Technol.
(SIELA), 31–39 (2014)
14. Fiala, J., Kuracina, M., Hrusovsky, I., Soldan, M.: Study of basic characteristics of hydrogen
generator. Appl. Mech. Mater. 10, 3078–3081 (2013)
An Automated System for the Stage
of Hydrolysis and Filtration
in the Extraction of Pectin
from the Cocoa Shell

Maritza Ccencho , Valeria Quijada , and Leonardo Vinces

Abstract Pectin obtained from cocoa husks has recently been investigated because
of its gelling and stabilizing properties that have great potential for the food, cosmetic
and pharmaceutical industries. However, its production at the industrial level has not
been studied or developed in Peru. A fundamental part of the extraction process is
the stage of hydrolysis and filtering of the cocoa shell. Because of this, an automated
system for acid-thermal hydrolysis and shell filtration is proposed. The control of both
processes is of great importance because the quality and efficiency of the extracted
pectin will depend on it. The tests will be carried out in a cylindrical taper with a
100 L capacity which is adapted to contain a 20 L cylindrical filter, both made of
AISI 304 stainless steel. The filter has a motor to homogenize the temperature and
pH of the mix. The pH of the process is 2 and the temperature is 90 ± 2 °C, since
in these ranges the pectin is released more easily from the cocoa shell. The method
consists of the structural design of the tank and the filter, and in the design of the pH
regulator and the heating system. The yield of the extraction process was achieved
by 10%.

Keywords Hydrolysis · Shell filtering · Pectin extraction · Cocoa shell

M. Ccencho · V. Quijada · L. Vinces (B)

Universidad Peruana de Ciencias Aplicadas, Av. Prolongación Primavera 2390, Santiago de
Surco, Lima, Peru
e-mail: [email protected]
M. Ccencho
e-mail: [email protected]
V. Quijada
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 267
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_26
268 M. Ccencho et al.

1 Introduction

Pectin is made up of complex molecules formed by polysaccharides, mainly by galac-

turonic acid units, esterified or by groups of methoxyl. These molecules are classi-
fied as high methoxyl and low methoxyl according to their degree of esterification,
providing different properties and gelation power to each of them [1]. These charac-
teristics have allowed it to be used primarily as a gelling and stabilizing agent in the
food and cosmetic industries, and it can also be used in the pharmaceutical industry
for its hydrocolloid, therapeutic and increased properties of the active ingredients.
New uses have also been investigated in the plastics industry and in the manufacture
of foaming products, such as clarifying agents and binders [2]. The global pectin
market has grown at a compound annual rate of around 6% during 2011–2018 [3]
and is estimated to reach USD 1.87 billion by 2026 [4]. At present, this product is
not produced in Peru, and demand is fully covered with imports from Mexico and
China. Currently, 365 t of pectin are imported into the country per year [5].
Likewise, it was decided to extract pectin from the cocoa shell to use this product
that is wasted nowadays. In Peru, 134,676 tons of cocoa are produced per year [6].
Of that amount, only 10% represents the fruit and the rest to the shell that is later
discarded.
Finally, the main objective of the project is to extract pectin from the cocoa shell
automatically and effectively. For this, a tank is designed where the hydrolysis stage
is allowed to be carried out at 90 °C and with a pH of 2. Also, a filter containing the
cocoa shell is designed to prevent contamination of the pectin when the extraction is
finished.

2 Literature Review

Previously, works on the subject have been developed. And now, the most important
ones are mentioned:
Andersen et al. have identified in [7] the variation of the properties and chemical
composition of the peels of different citrus fruits that are used to make pectin. The
dynamic model of the pectin extraction process was developed by determining the
main phenomena involved in a batch extraction reactor to predict the effect on the
performance and properties of the pectin in the key process conditions: temperature
and pH. Pectin was extracted from 0.5 kg of dried lime peel at different temperatures
(60, 70, 80 °C) and pH (1.5, 2.3, 3.1). However, this work only focuses on the stage
of acid hydrolysis. Likewise, the regulation of pH and filtering is done manually.
In addition, the proposed method is performed in two tanks, which has a greater
investment.
Casas-Orozco, Villa, Bustamante, and González propose in [8] the dimensioning
and programming of a semi-continuous plant for the production of 348 t/year of pectin
from Valencia orange peels in optimal experimental extraction conditions (90 °C,
An Automated System for the Stage of Hydrolysis … 269

75 min, pH = 1.5). The designed plant includes the acid hydrolysis, the filtration
of the hydrolyzed suspension, the precipitation of pectin, the gel pressing, the gel
drying and the reduction of pectin size. Total production of 337.610 kg/year of pectin
is achieved in the designed facility, with a lot size of 3383 kg of shell and a cycle
time (3.61 h) determined by the hydrolysis stage. However, this work only focused
on the simulation of the process and the implementation stage was not carried out.
Marsiglia et al. provide [9] information on the extraction and characterization of
the pectin from the cocoa shell. The cleaning, drying, grinding and sieving of the
shell were performed. For the extraction, dried and ground cocoa husk was used.
Each sample was mixed with acidified water to reach pH levels (2, 2.5 and 3). For
hydrolysis, the sample was placed in a three-necked flask with a mechanical stirrer,
a thermometer, and a heating mantle. Then, the mixture was cooled, centrifuged,
precipitated with 96% ethanol and dried at 40 °C. The highest yield was 15.97% and
it was obtained with citric acid at a pH of 2. However, this work is carried out at a
laboratory in small quantities and manually. Also, pretreatment of the cocoa shell is
needed.
Torres and Gonzales [10] developed a stainless steel agitator type tank with a
volume of 4.2 L was designed to carry out thermal, enzymatic and thermal-acidic
hydrolysis. The tank has a heating system through a 2000 W immersion resistance
to achieve constant temperatures of 50, 80 and 95 °C with a control system that
includes a thermostat, voltmeter, and autotransformer also has a Rushton agitator
connected to a reduction motor with frequency converter. The price of the equipment
was $15,000. However, this work only focused on the theoretical design of the tank
but not on its implementation.
In the absence of the implementation, automation and economic viability of the
other projects in the extraction of pectin, we propose to make an automatic system
that meets the appropriate characteristics for the realization of acid hydrolysis and
filtering of the cocoa shell. These stages are important in the extraction of pectin
since they determine the yield and quality of it.

3 Description of the Proposed Method

Figure 1 shows the block diagram of the development of the proposed method.

Fig. 1 Block diagram of the process

270 M. Ccencho et al.

3.1 Mechanical Design of the Tank

The tank size is calculated taking into account the volume that the filter would occupy
and the 50 L of water that will enter the tank. Therefore, a 100-L tank will be used.
Also, so that the container is resistant to corrosion, high temperatures and is sanitary,
stainless steel of 2 mm thick will be used in manufacturing.
The bottom of the tank is shaped like a cone trunk with a volume of 7 L. Therefore,
the cone trunk apothem is found with the following relationship.
 2

 V1 D − dt 2
Ap =   D+dt 2 + (1)
π 2
4
× 2

where V1 is the volume of the cone trunk, D is the diameter of the tank, dt the lower
diameter of the cone trunk and Ap is the cone trunk apothem.
Knowing that the diameter of the tank is 0.5 m, the bottom diameter is for a 1
tube and the volume is 0.007 m3 . The data is replaced in (1) and obtained Ap =
0.270 m.
The volume of the cylindrical part of the tank is expressed in (2).
2
D
V2 = π × × h2 (2)
2

where V2 is the volume of the cylinder part and h 2 is the height of the cylinder.
Knowing what is determined in (2) and that V1 is 0.007 m3 , the expressions (3),
(4) and (5) are found.

VT = V1 + V2 (3)

2
D
VT = 0.007 + π × × h2 (4)
2
VT − 0.007
h2 =  2 (5)
π × D2

where VT is the total volume of the tank.

Knowing that the total volume of the tank is 0.1 m3 and the diameter of the tank
is 0.5 m, it is replaced in (5) and obtained h 2 = 0.474 m.
Once the measurements are obtained, the tank is shown in Fig. 2.
An Automated System for the Stage of Hydrolysis … 271

Fig. 2 Hydrolysis tank

3.2 Mechanical Design of the Filter

The filter size for 5 kg of the cocoa shell is calculated. The reference is that the
density of the cocoa shell is 0.3 kg/L. Therefore, the filter would have a capacity of
20 L. Also, the same tank material will be used. As the filter vessel is a cylinder, the
following relationship is obtained.
2
df
V =π× ×h (6)
2

where V is the volume of the cylinder, df is the diameter and h is the height.
Knowing that the volume is 0.020 m3 and the diameter is 0.4 m, this data is
replaced in (6) and obtained h = 0.160 m.
Once the measurements are obtained, the filter is shown in Fig. 3.
Also, the filter has holes of 1 mm in all its cylindrical structure.
272 M. Ccencho et al.

Fig. 3 Hydrolysis tank filter

3.3 Design of the Regulator of PH

A certain pH is required in the acid hydrolysis stage. This variable is controlled with
the amount of citric acid dosed in the tank and with the electrode that measures the
pH in the mixture. For this process, an electric valve is used that drops citric acid
into the tank.
In the laboratory tests with 500 ml of water, approximately 40 g of citric acid
were needed. With these data, the ratio of water and acid is 25:2 (volume/mass).
Therefore, approximately 50 kg of citric acid is needed with the 50 L of process
water.
First, the citric acid is dropped in portions of 500 g and then the response of the
pH electrode is expected. Once a pH of 2.2 or less is reached, the dosage portion is
reduced to 100 g.
A pH of 2 is set as a setpoint. Once the desired pH is reached, the dosing system
stops completely.

3.4 Design of the Heating System

In the hydrolysis stage, a heating system is required, which has a temperature trans-
mitter that measures the temperature variables throughout the process. Also, there
is an electrical resistance of the band type around the tank. This will be the actuator
that will be used to generate the temperature change, which will vary with the power
An Automated System for the Stage of Hydrolysis … 273

given to the resistor by means of a PID control. Accuracy of ± 2 °C is necessary for

this heating system.
To obtain the necessary power for a transfer of energy in a certain time, the
following relation (based on the law of Fourier for conduction) is used [11].

m × Ce × (Tf − Ti )
P= × 4.184 (7)
t
where P is the power, m is the mass, Ce is the specific heat, Ti is the initial temperature,
Tf is the finale temperature and t is the time.
A mixture of 50 kg of water at a pH of 2 with 5 kg of cocoa was used for the
pectin extraction process. The mixture was heated from 20 to 90 °C and kept constant
for 90 min. The specific heat of the water is 1 kcal/kg°C and of the cocoa shell is
0.4 kcal/kg°C. Also, the heat that the filter will remove inside the tank is taken into
consideration. Therefore, the specific heat of stainless steel that is 0.12 kcal/kg°C is
taken into account. Replacing the data in (7), the expressions (8), (9) and (10) are
obtained.
For the water:
50 × 1 × (90 − 20)
PWater = × 4.184 = 4.067 kW (8)
3600
For the cocoa shell:
5 × 0.4 × (90 − 20)
Pcocoa = 4.184 = 0.163 kW (9)
3600
For the filter:
10 × 0.12 × (90 − 20)
PFilter = × 4.184 = 0.098 kW (10)
3600
Knowing what is determined in (8), (9) and (10), the sum of the powers is obtained
in (11).

PTotal = 4.067 + 0.163 + 0.098 = 4.328 kW (11)

A total power of 4.328 kW is obtained. Therefore, a 5 kW commercial value

resistor is used.
274 M. Ccencho et al.

100
90
80
Temperature (°C)

70
60
50
40
30
20
10
0
0 4 8 12 16 20 24 28 32 36 40 44
Time(min)

Fig. 4 Temperature variation during extraction time

4 Results

The tests were carried out in a place without air current, so as not to affect or impair
the process of hydrolysis of the cocoa shell. For the tests, 500 ml of water with 36 g
of critical acid was used to obtain acidified water with a pH of 2. Then, it was mixed
with 50 g of dried cocoa husk. The temperature was checked to be at 90 °C for 40 min
(Fig. 4). During all the time, the mixture was in constant agitation.
50 ml of liquid pectin with a weight of 50 g was obtained from 50 g of the cocoa
shell.
To obtain the dry pectin mass for yield, 96% ethanol was added and dried to obtain
5 g of dry pectin.
To obtain the yield the following relationship was used.

mass of dry pectin

%Yield = × 100 (12)
mass of dry cocoa shell

Knowing that the mass of dry pectin is 5 g and that the mass of dried cocoa shell
is 50 g, it was obtained that the yield was 10%.

5 Conclusions

According to the analysis performed for the hydrolysis and filtration stage in the
pectin production process from the cocoa shell, it was determined that is feasible
the implementation of an automatic system that allows the pectin extraction from
the cocoa shell. In this way, due to the high demand for this input, this market could
begin to be covered with a natural product. Also, it would be of great importance
An Automated System for the Stage of Hydrolysis … 275

that future studies automate the remaining stages of the process of obtaining pectin
powder.

References

1. Canteri, M., Moreno, L., Wosiacki, G.: Pectina; da matéria-prima aoproducto final. Polímeros
22(2), 149–157 (2012)
2. Bazarte, H., Sangronis, E., Unai, E.: La cáscara de cacao (Theobroma Cacao L.): una posible
fuente comercial de pectinas. Archivos Latinoamericanos de nutrición 58(1), 64 (2008)
3. IMARC Services Private Limited Report List: Pectin Market: Global Industry Trends, Share,
Size, Growth, Opportunity and Forecast 2019–2024 (2019)
4. Polaris Market Research: Pectin Market Size, Share, Trends, & Industry Analysis Report,
(By Sources (Thickener, Stabilizer, Gelling Agent, Fat Replacer), By Application (Food &
Beverage, Pharmaceuticals, Bakery, Dairy Products), By Regions: Segment Forecast, 2017–
2026 (2018)
5. Comercio exterior para el agro V.1.1. http://sistemas.minagri.gob.pe/siscex/usuarios/home.
Last Accessed 10 Sep 2019
6. Andina.: https://www.andina.pe/agencia/noticia-minagri-proyecta-una-produccion-cacao-149
000-toneladas-2019-753060.aspx. Last Accessed 9 Sep 2019
7. Andersen, N., Cognet, T., Santacolma, P., Larsen, J., Armagan, I., Larsen, F., Gernaey, K., Abild-
skov, J., Huusom, J.: Dynamic modelling of pectin extraction describing yield and functional
characteristics. J. Food Eng. 192, 61–71 (2017)
8. Casa-Orozco, D., Villa, A., Bustamante, F., González, L.: Process development and simulation
of pectin extraction from orange peels. Food Bioprod. Process. 96, 86–98 (2015)
9. Marsiglia, D., Ojeda, K., Ramírez, M., Sanchéz, E.: Pectin extraction from cocoa pod husk
(Theobroma cacao L.) by hydrolysis with citric and acetic acid. Int. J. ChemTech Res. 9(7),
497–507 (2016)
10. Torres, K., Gonzales, M.: Diseño de un tanque agitado para la etapa de hidrólisis en la
producción de miel de agave. Jóvenes en la ciencia 1(3), 22–26 (2015)
11. Çengel, Y.: Transferencia de calor y masa, 3rd edn. McGraw-Hill/interamericana editores, S.A.
de C.V., Mexico (2007)
Automatic Leaf Segmentation
from Images Taken Under Uncontrolled
Conditions Using Convolutional Neural
Networks
Itamar Franco Salazar-Reque
and Samuel Gustavo Huamán Bustamante

Abstract Automatic leaf segmentation from images taken in-field in uncontrolled

conditions is a very important problem that has not been properly reviewed and that
is crucial due to its possible use as a previous step in classification algorithms that can
be used in agriculture applications. In this work, a CNN architecture (LinkNet) was
trained to solve the isolated leaf segmentation problem under natural conditions. To
do so, an open dataset has been modified and augmented, using rotations, shearing,
and artificial illumination changes, in order to have a proper amount of imagery
for training and validation. We have tested the CNN in two different datasets: The
first belongs to the original open dataset that shares some visual characteristics with
training and validation dataset. The second one contained its own imagery from a
different set (images from different plants and with different illumination conditions)
in order to evaluate the CNN model generalization. We obtained a mean Intersection
Over Union (IoU) value of 0.90 for the first test and a 0.92 for the second one. An
analysis of these results has been made and some problems regarding classification
applications were commented.

Keywords Leaf segmentation · CNN · In-field acquisition

1 Introduction

Leaves are biological structures that play a crucial role in trees life due to the photo-
synthesis process they are in charge of, a process which is necessary to produce
energy. Most of the year leaves are present on trees, even more than fruits or flowers.
For this reason, farmers often use them to inspect the development of their crops.

I. F. Salazar-Reque (B) · S. G. Huamán Bustamante

Instituto Nacional de Investigación y Capacitación de Telecomunicaciones, Universidad Nacional
de Ingeniería, Lima 15021, Peru
e-mail: [email protected]
S. G. Huamán Bustamante
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 277
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_27
278 I. F. Salazar-Reque and S. G. Huamán Bustamante

In fact, it is possible to observe over leaves visual manifestations triggered by nutri-

tional deficiencies, diseases, or pests attacking the tree. However, the process to
identify these visual manifestations tend to be highly subjective and time-consuming.
Computer vision techniques applied to leaves digital images have been developed to
tackle this problem, some of those use classical techniques like color components
ratios [1], histogram over new color models [2] or multilayer perceptron (MLP)
trained with superpixels [3] to classify or to segment visible diseases manifestations
over the leaves. Recently, convolutional neural networks (CNN) have shown great
performance in classification problems, for this reason, in [4], the authors used that
approach to tackle the plant disease detection problem. Nevertheless, some works
have reported that CNN’s are using image background information to perform their
identifications (i.e., parts of the image different than the leaf of interest) [4, 5].
For example, in [5], the author found that removing image background (uniform or
complex background) alters classifier performance, that is, CNN’s found spurious
correlations [6] in data and used those to classify between classes. This motivates the
need for a previous step to eliminate image background information. By doing this,
we ensure that CNN is focusing only on characteristics from the leaf of interest. A
review of leaf segmentation in the uniform background can be found in [7], in that
research authors also proposed a method that used Expectation-Maximization (EM)
clustering in saturation-value space to provide priors that serve as energies used in
a graph cut step to perform the task. In [8], a modification of that technique, which
used superpixels and k-means clustering, was proposed achieving similar results
over images taken in-field with a white sheet of paper under the leaf of interest. Even
though, capturing pictures under this type of semi-controlled conditions is a laborious
and unpractical procedure. Other works, focused on the segmentation of all individual
leaves present in an image, this is a challenging task due to the presence of occlusions
and shape variability. In [9] a review of techniques to segment all individual leaves in
images of rosette plants taken under lab conditions was made. However, for practical
applications to be made, image segmentation should be performed under natural
conditions, exposed to illumination changes, variety of shapes, occlusion among
others. With this in view, in [10] the authors presented Dense-Leaves, a dataset of
images captured in dense foliage with a ground-truth segmentation label. They also
proposed a CNN architecture to automatically discriminate leaf boundaries that were
used in a watershed-based algorithm to perform segmentation of all leaves. In this
work, we have used the same dataset to train a CNN to directly segment the leaf of
interest, which we assumed is always centered in the image.
Automatic Leaf Segmentation from Images Taken Under … 279

Fig. 1 A sample of augmented images using the procedure described. The first row, from left
to right: original, rotated, and flipped image. Second and third-row show images with random
shearings. Each transformation was followed by a random illumination change

2 Method

2.1 Image Datasets

Our dataset is a modification of the one provided in [10]. The original dataset is
composed of 146 foliage images, each one with labels from leaves present in the
image, although, no all leaves were labeled. We used labels provided to generate new
images by isolating and centering leaf of interest. After that, we manually selected
best-isolated images, this was a necessary step due to the fact that some leaves
were highly occluded so the resulting isolated image was incomplete. Thus, we
obtained 910 images for training, 187 for validation, and 412 for the test, each of
280 I. F. Salazar-Reque and S. G. Huamán Bustamante

those with respective masks. We performed data augmentation techniques in order

to have a proper amount of data to train a CNN. Data augmentation was made by
rotating images 90°, 180°, and 270° degrees, images were horizontally and vertically
flipped, we applied random shearing transformations over the leaves and finally, we
performed illumination changes in every image (excepting the original) by modifying
luma component by a maximum value of 0.3. Thus, we obtained 32,670 images
for training, 6732 for validation, and 14,832 for testing, all of them with a spatial
resolution of 224 × 224 pixels. A sample of this procedure is shown in Fig. 1, in
which we performed one rotation (180°), one flipping (vertical), two shearings, and
illumination changes to one centered leaf.

2.2 CNN Architecture and Training Procedure

Some architectures have been proposed to solve the leaf segmentation problem [11–
13], all of them with images captured in lab conditions as the techniques reviewed
in [9]. In this work we used LinkNet [14], a CNN architecture developed to perform
segmentation in complex scenes that reuses CNN parameters efficiently so that the
total number of parameters is small. This architecture has four residual encoding
blocks and four decoding blocks using full-convolutions. Due to spatial information
loss caused by subsampling in encoding blocks this architecture links each encoding
block input to corresponding decoding block output. In Fig. 2, we show encoding
blocks in blue and decoding blocks in green, see how encoding blocks have residual
units and green blocks have full-convolutional steps. The kernel size (k), the stride
(s), and the up-sampling (up = 2) hyperparameters are also indicated. Gray boxes
correspond to pre-processing and post-processing blocks described in [14]. We have
used augmented RGB images (see Sect. 2.1) with corresponding binary masks to
train and validate (test images are defined in the next section) the CNN architecture
by 100 epochs.

Fig. 2 a Residual encoding block b Decoding block c Convolutional neural network architecture
based on residual encoding and decoding blocks
Automatic Leaf Segmentation from Images Taken Under … 281

Adam [15] and AMSGrad [16] were used as optimizer techniques, both with
learning rate of 0.0005, β 1 coefficient of 0.9 and β 2 coefficient of 0.999; parameters
updating was made every 16 images (N = 16) and, as loss function, we have used
binary cross-entropy (BCE) whose formulation is defined in Eq. (1). Additionally,
we have monitored mean intersection over union (IoU) metric defined in Eqs. (2)
and (3).
⎛ ⎞
1  1 ⎝
N P
BC E = − (T i ( j) log(Y i ( j)) + (1 − T i ( j)) log(1 − Y i ( j)))⎠
N i=1 P j=1
(1)
1 
IoU = IoU(γ ) (2)
K γ ∈

(Y i ≥ γ ) ∩ T i
IoU(γ ) = (3)
(Y i ≥ γ ) ∪ T i

where N is the batch size, P is the number of pixels of the target binary mask (in our
case P = 50,176 since binary masks dimension is 224 × 224), T i (j) is the value of
pixel “j” from target binary mask “i” (T i ). Y i (j) is the value of pixel j from CNN
output image (Y i ), Y i is the CNN result, the parameter γ is a threshold value used
to binarize Y and  is the set of K discrete values that γ can take, in our case  =
{0.5, 0.55, 0.60, …, 0.95}.

2.3 Test Datasets

We used two image datasets to test CNN efficiency. Test dataset 1: Images with similar
characteristics that training and validation sets, i.e., they were obtained from the
original Dense-Leaves dataset. Test dataset 2: Images with different characteristics
that training and validation datasets. They were captured in-field under uncontrolled
conditions from avocado trees having different nutrient deficiencies and plagues.
We used them to assess whether the trained CNN maintained its performance with
images other than those used for training and validation.

3 Results

The BCE and IoU training values per epoch are shown in Fig. 3. In general, a lower
BCE value implies less difference between values from target binary masks (Ti ) and
outputs (Yi ). A higher IoU value implies a greater similarity between Ti and binarized
output (Yi ≥ γ ).
282 I. F. Salazar-Reque and S. G. Huamán Bustamante

Fig. 3 a IoU metrics and b BCE loss during training

Table 1 Average IoU(γ = 0.5) for training, validation, test 1 and test 2 datasets
Train Val Test 1 Test 2
ADAM—best 0.953 0.906 0.897 0.909
AMSGrad—best 0.962 0.911 0.903 0.913
ADAM—100 0.976 0.910 0.902 0.921
AMSGrad—100 0.975 0.907 0.899 0.915

In general, AMSGrad made BCE values lower than those obtained with ADAM. It
can see how, independently of the optimization technique, BCE loss tends to increase
around the 20th epoch. This would imply a decrease in performance, however, this
trend contrasts with IoU the metric tendency which, instead of decreasing, increases
for both optimization techniques (a discussion of this is made in Sect. 4). Table 1
shows average IoU values for γ = 0.5 for training, validation, tests 1 and 2 datasets
using best CNNs, which are the ones with the lowest BCE, and CNNs of the 100th
epoch. We contrasted these results to verify whether there was indeed a decrease in
CNN performance by letting BCE continue to increase. A sample of the segmentation
visual results is shown in Fig. 4. This figure is divided into eight groups. Groups A,
B, C and D have images from test data 1 and groups E, F, G and H have images
from test data 2. Each group has an RGB image that is accompanied by five binary
images. The corresponding ground-truth image (Ti ) is shown just below the RGB
image. The remaining four binary images correspond to binarized results (Yi ≥ 0.5)
for ADAM and AMSGRAD and for best CNN or 100th epoch CNN.
Automatic Leaf Segmentation from Images Taken Under … 283

Fig. 4 Sample of the segmentation visual results

4 Discussion

From Fig. 4 and Table 1, we noted that masks obtained from CNN’s optimized using
ADAM and AMSGRAD do not differ considerably. In addition, we do not observe
substantial differences between “best” CNNs and “100th” CNNs, even though the
latter had a higher BCE (see Fig. 3). This can be explained by Eqs. (1), (2), and (3).
284 I. F. Salazar-Reque and S. G. Huamán Bustamante

The IoU metric is formulated so that a binarized output (Yi ≥ γ ) is necessary, while
BCE equation uses directly CNN’s output (Yi ). Therefore, for IoU computation there
is no need for Yi and its corresponding Ti to be close. This would explain why IoU
continues to improve while BCE loss function worsens (see Fig. 3). Images from
test 2 have different characteristics that ones used in training and validation. This
was made to assess CNN’s ability to generalize. We found that results in test 2
datasets were similar to those in validation and test 1. That is, it maintained a similar
behavior even with different images, this contrasts with the behavior that CNNs
trained for classification tasks have shown [4, 5]. We argue that this happens because
leaf segmentation is a more complex problem than enforces CNN to include context
information which makes generalization easier. However, CNN segmentations are
not perfect, it should be studied if these imperfections impair the performance in
classifications tasks. There are two main imperfections. First, sometimes CNN masks
included elements outside the leaf. For example, in Fig. 4, masks from groups D,
F, and G add some elements to the output mask that differ from GT. If we use
these masks to segment RGB images and use those as inputs to a CNN trained
for classification this could lead to uncontrolled errors. Even when some floating
elements can be handled with morphological transformations, these transformations
will be image-dependent. A second problem is that in some cases CNN eliminates
damaged areas from the leaf. Take the example of the image in group H in Fig. 4,
in which brown areas from leaf apex were confused with an image background and
therefore were not included in the final mask. Usually, diseased regions in a leaf
represent a small area, thus, segmentation performance metrics are not very affected.
However, in classification tasks, the no inclusion of these areas could seriously affect
the expected results.

5 Conclusions

In this work, we have trained a CNN to segment leaves images acquired in-field under
uncontrolled conditions with complex backgrounds. CNN generalization was eval-
uated using two test sets: the first belongs to the original Dense-Leaves dataset that
shares some visual characteristics with training and validation sets and the second
one, containing own imagery from different plants and with different illumination
conditions. We found that, despite having some overfitting, the CNN kept its valida-
tion results constant for the two test sets, indicating good generalization. This was
true for the two optimization techniques used (ADAM and AMSGRAD) and for
CNN at two different states of training. Also, it has been reported that an increase in
the loss function does not necessarily imply a decrease in IoU. Additionally, we have
observed that, in some cases, the resulting mask has floating elements; in others,
it eliminates part of the leaf, especially injured parts. These problems can generate
errors if resulting masks are used for classification purposes.
Automatic Leaf Segmentation from Images Taken Under … 285

Acknowledgements The authors thank FONDECYT PERU for the funds allocated to the project
97-2018-FONDECYT-BM-IADT-AV.

References

1. Barbedo, J.G.A: A new automatic method for disease symptom segmentation in digital
photographs of plant leaves. Eur. J. Plant. Pathol. 147, 349–264 (2017)
2. Camargo, A., Smith, J.S.: An image-processing based algorithm to automatically identify plant
disease visual symptoms. Biosyst. Eng. 102, 9–21 (2009)
3. Salazar-Reque, I.F., Huamán, S.G., Kemper, G., Telles, J., Diaz, D.: An algorithm for plant
disease visual symptom detection in digital images based on superpixels. Int. J. Adv. Sci. Eng.
Inf. Technol. 9, 194–203 (2019)
4. Mohanty, S.P., Hughes, D.P., Salathé, M.: Using deep learning for image-based plant disease
detection. Frontiers Plant Sci. 7, 1–10 (2016)
5. Barbedo, J.G.A: Impact of dataset size and variety on the effectiveness of deep learning and
transfer learning for plant disease classification. Comput. Electron. Agricul. 153, 46–53 (2018)
6. Arjovsky, M., Bottou, L., López-Paz, D.: Invariant risk minimization. arXiv preprint arXiv:
1907.02893 (2019)
7. Soares, J.V.B., Jacobs, D.W.: Efficient segmentation of leaves in semi-controlled conditions.
Mach. Vis. Appl. 24, 1623–1643 (2013)
8. Salazar-Reque, I.F., Pacheco A.G., Rodriguez. R.Y., Lezama, J.G., Huamán, S.G.: An image
processing method to automatically identify Avocado leaf state. In: 22nd Symposium on
Image, Signal Processing and Artificial Vision (STSIVA). pp. 1–5. IEEE Press, Bucaramanga,
Colombia (2019)
9. Scharr, H., Minervini, M., French, A.P., et al.: Leaf segmentation in plant phenotyping: a
collation study. Mach. Vis. Appl. 27, 585–606 (2016)
10. Morris, D.: A pyramid CNN for dense-leaves segmentation. In: 15th Conference on Computer
and Robot Vision (CRV). pp. 238–245. IEEE Press, Toronto, Canada (2018)
11. Aich, S., Stavness, I.: Leaf counting with deep convolutional and deconvolutional networks.
In: Proceedings of the IEEE International Conference on Computer Vision. pp. 2080–2089.
IEEE Press, Venice, Italy (2017)
12. Romera-Paredes, B., Torr, P.H.S.: Recurrent instance segmentation. In: European Conference
on Computer Vision. pp. 312–329. Springer, Cham, (2016)
13. Ren, M., Zemel, R.S.: End-to-end instance segmentation with recurrent attention. In: Proceed-
ings of the IEEE Conference on Computer Vision and Pattern Recognition. pp. 6656–6664.
IEEE Press, Honolulu, Hawaii (2017)
14. Chaurasia, A., Culurciello, E.: Linknet: Exploiting encoder representations for efficient
semantic segmentation. In: 2017 IEEE Visual Communications and Image Processing (VCIP).
pp. 1–4. IEEE Press, St. Petersburg, FL, USA (2017)
15. Kingma, D.P., Ba, J.: Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.
6980 (2014)
16. Reddi, S.J., Kale, S., Kumar, S.: On the convergence of adam and beyond. arXiv preprint arXiv:
1904.09237 (2019)
Method of Anomalies Detection in Persea
Americana Leaves with Thermal
and NGRDI Imagery

Bruno D. Rivadeneyra Bustamante

and Samuel G. Huamán Bustamante

Abstract Among the main agricultural export products of Perú is the Persea Amer-
icana (Hass avocado). However, pests and deficiencies cause a decrease in avocado
crop yield. For that reason, anomalous visual characteristics detection of a leaf has
been studied for thermal and multispectral images. In this study, we propose a proce-
dure for imagery acquisition and evaluate factors that can affect the acquisition of
thermal imagery; furthermore, we propose a method for the extraction of charac-
teristics to classify between a healthy leaf and a leaf with some visible anomalies
that indicate some deficiency or pest. This method consists of the following stages:
segmentation of thermal and NGRDI images, resampling, clustering with k-means
algorithm and classification with SVM. Likewise, we test three cases of image acqui-
sition to analyze the effects on statistical descriptors. Finally, the classification stage
was tested with leaves processed; as a result, we got out an average accuracy of
82.67%, from ten experiments with the same set of images.

Keywords Thermal imagery · NGRDI · K-means · SVM

1 Introduction

Peru is one of the countries with the highest export of Persea Americana (avocado)
in the world, as well as in its production [1]. The most common avocado varieties in
Peru are Hass and “Fuerte” [2]. However, 95% of avocados exported by Peru are of
the Hass variety.

B. D. Rivadeneyra Bustamante (B) · S. G. Huamán Bustamante

Instituto Nacional de Investigación y Capacitación de Telecomunicaciones, Universidad Nacional
de Ingeniería, Lima 15021, Peru
e-mail: [email protected]
S. G. Huamán Bustamante
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 287
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_28
288 B. D. Rivadeneyra Bustamante and S. G. Huamán Bustamante

The growing demand for both the production and export of this fruit is reflected
in the consolidation of trade with the USA and the European Union and with the
opening of new markets such as Chinese since 2015 [1].
However, the yield of avocado tree crops is not as expected. Peru produced an
average of 11.8 t ha−1 in 2017, which places it out of the top ten countries in avocado
yield [1]. One of the causes of low agricultural yield was the poor control and care
of the vegetation. For this reason, hard work is necessary to achieve the reduction of
pests and deficiencies in avocado crops.
The vegetation type detection or classification in multispectral images using arti-
ficial neural networks has been studied in recent years. For instance, the images have
been segmented to restrict the area where a leaf or tree is located and, from this
result, the color characteristics are extracted in order to train a classifier [3–7]. In
that sense, the addition of thermal images into training has improved the accuracy
of those classifiers [7].
Thermal images in crops have been studied [8, 9], demonstrating that, in general,
any leaf increases the temperature in the diseased areas of its beam surface [8] and,
in addition, the average of the temperature over a tree canopy varies depending on
the hour of acquisition [7].
For this work, the acquisition and processing of thermal and RGB images is
performed to detect anomalies in Hass avocado leaves, using a proposed method
that considers the temperature distribution in the beam of the leaves, as a descriptor
of the anomalous characteristics that may occur on a leaf, and color characteristics,
through the Normalized Green-Red Difference Index (NGRDI).
Section 2 details the procedure for acquiring images using an infrared camera. In
Sect. 3, the method used in extracting information from images, using the k-means
algorithm, and the classification stage, with a Support Vector Machine (SVM), is
explained.
Section 4 has the results and discussions of experiments related to the average
surface temperature variation of the Hass avocado leaf beam at different times of the
day and the resultant performance of the proposed method for the classification of
healthy leaves and leaves with anomalies.

2 Image Acquisition

2.1 Instruments and Environmental Considerations

Image acquisition was carried out in the avocado cultivation area of the National
Institute of Agrarian Innovation (INIA), located in Lima (Peru), during this time, 499
thermal images and 499 RGB images of Hass avocado leaves were acquired, of which
100 of each type were selected, according to the visible and thermal characteristics
useful for anomalies detection. These images were acquired with the FLUKE TiS
45 9 Hz infrared camera [10], of which 20 correspond to healthy leaves and 80 to
Method of Anomalies Detection in Persea Americana Leaves … 289

diseased leaves or with visible anomalies. Thermal images have a size of 320 × 240
pixels and RGB images have a size of 640 × 480 pixels.
Regarding the configuration of the infrared camera, this was adjusted to an emis-
sivity of 0.95, as an average of the accepted values for the surface of a leaf [9]. In addi-
tion, the specific background temperature is considered for each of the acquisitions.
This temperature was measured with a FLUKE 80bk-a thermocouple.

2.2 Image Acquisition Procedure

The proposed procedure for image acquisition is the result of various field tests and
consists of the following steps:
1. Locate a leaf, which must not be directly exposed to the Solar light but has
adequate lighting.
2. Gently wipe the surface of the leave with a cloth until dust and dirt are removed,
then wait a few minutes until the leaf surface again acquires thermal equilibrium
with the environment.
3. Insert the background temperature in the infrared camera. This will be the
temperature measurement that thermocouple marks at leaf position. For this,
the measurement is made by placing the thermocouple as close to the leaf beam
surface as possible.
4. Position the camera so that the leaf extension is located vertically inside the
camera’s viewfinder. Then, focus, capture and finally, store.
5. Repeat the acquisition procedure for all the chosen leaves.
In conditions of high air temperatures, it will be necessary not to keep the camera
on for more than an hour to avoid electronic system overheating. In these cases, it
is recommended, after one hour of activity, turn it off for 15 min and then continue
with the acquisition.
The thermographic camera configuration only allows manual focus adjustment.
Then, a leaf will be correctly focused as soon as the RGB and thermal images fit
as well as possible. Likewise, it is recommended that the camera position does not
obstruct direct radiation that is toward the leaf surface.
The hands and, in general, the human body emits thermal radiation that can
alter the surface measurement of the leaf in the infrared camera. Therefore, it is
recommended to be placed at one meter away from the leaf at the acquisition time.

3 Proposed Image Processing and Acquisition Method

The thermal images acquired with the thermographic camera are exported as a text
file, which contains a matrix of representative temperature values for each acquisition.
290 B. D. Rivadeneyra Bustamante and S. G. Huamán Bustamante

Within the mentioned matrix, the leaf is segmented, selecting four points (on a
displayed image) that determine a rectangle of minimum dimensions but that encloses
the leaf. Then, the k-means technique [11] is used in order to divide the leaf into k =
15 clusters. From each of these clusters, the statistical descriptors, mean and standard
deviation, are extracted to obtain a total of 30 descriptors of the thermal image of the
avocado leaf.
On the other hand, RGB images acquired by the infrared camera are also exported,
but in jpg format. As in thermal imaging, these are segmented in each of their bands,
following the same procedure.
Once a leaf is cut, it is resized to the size of the corresponding thermal image,
then, Normalized Difference Green-Red Index (NGRDI) [12], is calculated in each
pixel and It is defined by the following expression:

NDG − NDR
NGRDI = (1)
NDG + NDR

where NDG is the digital value of each pixel in the green band and NDR is the digital
value of each pixel in the red band. While the resulting image of a single color band
whose pixel values correspond to the respective NGRDI value will be called the
NGRDI image.
Similar to the thermal image processing, the K-means algorithm is used to divide
the NGRDI image into 15 clusters and for each one, the mean and standard deviation
is calculated. These values add a total of 30 descriptors.
Finally, the number of pixels of each cluster is calculated, both in the thermal
image and in the NGRDI image, adding 15 descriptors. In total, there is a set of 90
descriptors for each leaf, they are inserted into an SVM network [13], which uses
a linear Kernel transformation function. Likewise, the images were distributed in
training, validation and test groups (70%, 15%, and 15% respectively).
In Fig. 1, the diagram of the stages of the proposed method is shown.

4 Results and Discussions

The experiments carried out and shown below are proposed to analyze the char-
acteristics of descriptors and the effects of environmental conditions on these and,
consequently, on the detection of the anomalous and non-anomalous condition of a
tree leaf.
These experiments were carried out in the experimental station of the National
Institute of Agrarian Innovation (INIA), in Lima. Thus, in the proposed first experi-
ment to verify the effects due to direct solar radiation, images were acquired of two
healthy Hass avocado leaves, one under the shade and another exposed to direct solar
radiation, both belonging to different trees. The results are shown in Fig. 2.
Method of Anomalies Detection in Persea Americana Leaves … 291

Fig. 1 Diagram with the processing steps of the proposed method

During this experiment, the sky remained cloudy during the first five acquisitions,
in which, there was a minimum temperature difference over time between both
leaves. However, from the following acquisitions, the solar radiation modified the
temperature of the exposed leaf significantly. In Fig. 2, the hourly temperature is
shown where there was an incidence of direct solar radiation on the leaf, as well as
the temperature of a leaf under the shade, called an unexposed leaf. To quantify the
relationship between these temperatures, the Spearman correlation coefficient [14]
was used, whose value is 0.4857; this represents a moderate correlation. In addition,
temperature differences between both measurements are of 4–10 °C.
292 B. D. Rivadeneyra Bustamante and S. G. Huamán Bustamante

Fig. 2 a Variations in the average surface temperature of b a healthy leaf exposed to direct solar
radiation and c a healthy leaf under the shade and in different Hass avocado trees

The results are shown in Fig. 2, require caution to avoid direct solar radiation
in the beam of Hass avocado leaves during acquisition procedure in the following
experiments. The second experiment consisted of acquiring images from healthy
leaves and leaves with anomalies. Thus, the acquisition of 22 thermal images and
22 RGB images of the Hass avocado leaf was carried out, in the range from 11 h
to 15 h on May 31, 2019, at INIA. Each acquisition was simultaneously carried
out for a healthy leaf and an anomalous leaf, both were under shade. The resulting
temperature curves, shown in Fig. 3a, represent the variation of the average surface
temperature of the healthy and the anomalous leaf beam. The temperature of the curve
that represents the healthy leaf is below the temperature of the curve that represents
the leaf with an anomaly, for each acquisition. Both leaves, of the same tree and very
close to each other, always remained in the shade. In addition, the curves show a
similar trend over time. Spearman’s correlation for this case has a value of 0.9018,
with significance value p = 0.0001; this result indicates a high correlation.
On the other hand, Fig. 3b shows the comparison of the same leaf with an anomaly,
this time with a healthy leaf belonging to another tree also under shade. In this case,
Spearman’s correlation of 0.8064, with significance value p = 0.0027, is still high.
In spite of a reduction in the correlation, a significant difference was maintained
between the healthy leaf temperatures and the leaf anomalous temperatures, which
allows those to be differentiated.
Consequently, a third experiment was carried out to analyze the variation of the
temperature distribution in three different acquisition positions on the Hass avocado
leaf beam. In this experiment, three thermal images were acquired by rotating on
the azimuth: front and two lateral sides, all of them correspond to a healthy leaf
under a shade, but with sufficient illumination. The acquired thermal images were
preprocessed, cropping the region bounded by the leaf, then, they were divided into
k = 15 significant clusters, following the method mentioned in Sect. 3. Next, an
analysis of the regions of the leaf that belong to clusters with the highest number of
pixels, as well as a surface distribution of clearly identifiable temperatures.
Method of Anomalies Detection in Persea Americana Leaves … 293

Fig. 3 a Average surface temperature variation of c a healthy leaf and d an anomalous leaf of the
same avocado tree and, b the average surface temperature variation of e a healthy leaf of different
avocado tree and the same anomalous leaf in 11 different hours of the day

Figure 4 shows the temperature distribution of three clusters on the two healthy
leaves for the three views (one front and two sides), where one is under a shade and
the other exposed to direct solar radiation.
Table 1 shows that there is a variation of pixels between the different views,
which varies from 26 to 28% for cluster 1 and 33–35% for cluster 2; Despite this,
the average temperature does not vary. While the number of pixels for cluster 3, for
all three views, varies from 35 to 62%, in addition to the temperature variation up
to 0.3 °C. This means that different views have an effect on temperature for clusters
with the least number of pixels; consequently, this variation may have an effect on
the performance of the classification algorithm.
In the case of Table 2, in cluster 1 it is observed that the average temperature varies
from 0.1 to 1.4 °C, in cluster 2 from 0.5 to 1.7 °C and in cluster 3 from 0.5 to 1.4 °C.
In this case, when acquisition in any different view is carried out, the variations in
average temperature are significant, so that it is necessary to perform the acquisition
under shade.
294 B. D. Rivadeneyra Bustamante and S. G. Huamán Bustamante

Fig. 4 Images of two healthy leaves in three views. From left to right, the RGB image and three
views (front, side 1 and side 2, respectively) are shown. In (s) the top row, (s, a) a leaf under a shade
and their respective clusters for (s, b) the front view, (s, c) side 1 and (s, d) side 2. In (r) the bottom
row, (s, a) a leaf exposed to solar radiation and their respective cluster for (r, b) the front view, (r,
c) side 1 and (r, d) side 2. Clusters 1, 2 and 3 can be observed in green, red and blue respectively

Table 1 Effects of the infrared camera rotation on three clusters that belong to a healthy leaf under
the shade and obtained with the K-means algorithm. In the upper region of the avocado leaf cluster
1 and in the lower region cluster 2. In addition, cluster 3 with a variation of the average temperature
Leaf under shade View Number of pixels Mean (o C) Standard deviation (°C)
Cluster 1 Front 1464 19.10 0
Side 1 1122 19.10 0
Side 2 1082 19.10 0
Cluster 2 Front 747 19.70 0
Side 1 500 19.70 0
Side 2 484 19.70 0
Cluster 3 Front 258 20.00 0
Side 1 165 20.26 0.06
Side 2 98 20.02 0.06
Method of Anomalies Detection in Persea Americana Leaves … 295

Table 2 Effects of the infrared camera rotation on three clusters that belong to a healthy leaf
exposed to direct solar radiation and obtained with the K-means algorithm
Leaf exposed to solar View Number of pixels Mean (°C) Standard deviation (°C)
radiation
Cluster 1 Front 1290 23.70 0.06
Side 1 493 25.03 0.09
Side 2 749 24.90 0.06
Cluster 2 Front 690 24.64 0.09
Side 1 605 26.38 0.06
Side 2 509 25.85 0.09
Cluster 3 Front 345 25.42 0.12
Side 1 299 26.85 0.09
Side 2 508 26.31 0.17

In both Tables 1 and 2, uniform temperature distributions are assumed on the

leaf and, consequently, standard deviations of uniform distribution. For the analysis
of the three experiments described above, only the region bounded by the leaf is
considered, where temperatures have a distribution similar to a uniform distribution.
However, for the method proposed in Chap. 3, the rectangular re-cutting of the leaf
justifies the use of a standard deviation for a normal distribution, due to the variability
of the environment surrounding the leaf.
It should be noted that the standard deviation of cluster 3 increase in a leaf under
a shade, for a smaller number of pixels. While its behavior is erratic in a leaf exposed
to solar radiation.
The SVM classifier to healthy leaves and anomalous leaves was trained and tested
with 100 thermal and NGRDI images, which are images that correspond to leaves
not exposed to solar radiation, each of which was processed following the method
proposed in Sect. 3. This classifier uses a linear Kernel transformation function, as
well as a distribution of seventy-five training images, fifteen for validation and fifteen
for the test. The result of an average of 10 experiments for the same set of 100 images
shows an accuracy of 82.67%.

5 Conclusions

In this work, we propose a new method to process, cluster and classify Hass avocado
leaves condition using thermal and NGRDI images and a procedure to acquire thermal
and RGB images of leaves in the field, using a handheld infrared camera.
The four experimental results about the effects of solar radiation on avocado leaves
show alteration in the temperature measurements and, consequently, to clustering of
healthy or dead zones of a leaf.
296 B. D. Rivadeneyra Bustamante and S. G. Huamán Bustamante

In future works, the accuracy of the classification stage can be improved with a
greater variety and quantity of images obtained by the image acquisition process and
processed by the method proposed in this work.

Acknowledgements The authors thank Innóvate Peru and the Ministry of Production of Peru
for financing with contract No. 119-INNOVATEPERÚ-IDIBIO-2018 and the National Institute of
Agrarian Innovation (INIA) for access to their experimental stations.

References

1. MINAGRI Dirección General de Políticas Agrarias: La Situación del Mercado Internacional

de la Palta. Boletín técnico, 8–35 (2019)
2. Huaman, N.L., Zapana, V., Granados, C.E.: Aislamiento e identificación de Phytophthora
cinnamomi Rands en el cultivo de palto variedades Hass y Fuerte. Journal de Ciencia y
Tecnología Agraria 1, 57 (2015)
3. Blanco, J.A.U.M.E.: Desarrollo de una aplicación móvil para la detección y clasificación de
hojas de árboles. Dissertation, Universitat Politècnica de València (2015)
4. Galárraga, J.L.: Clasificador de hojas mediante Deep Learning. Dissertation, Universidad
Politécnica de Madrid (2017)
5. Garnique, L.A.V., Cabrera, H.I.M., Chavarry, K.L.V., Barco, M.J.V.: Visión artificial: Apli-
cación de filtros y segmentación en imágenes de hojas de café. Ingeniería: Ciencia, Tecnología
e Innovación. 1(2), 71–81 (2014)
6. Salazar-Reque, I. F., Pacheco, A. G., Rodriguez, R. Y., Lezama, J. G., Huamán, S. G.: An image
processing method to automatically identify Avocado leaf state. Paper presented at the 2019
XXII Symposium on Image, Signal Processing and Artificial Vision (STSIVA), Universidad
Industrial de Santander, Bucaramanga, 24–26 Apr 2019
7. Calderón, R., Zarco, P., Navas, J., Blanca, M.: Detección de enfermedades de cultivos me-
diante imágenes hiperespectrales y térmicas de alta resolución espacial. Grandes cultivos, 18
(2018)
8. Granum, E., Pérez, M., Calderón, C., Ramos, C., de Vicente, A., Cazorla, F., Barón, M.:
Metabolic responses of avocado plants to stress induced by Rosellinia necatrix analysed by
fluorescence and thermal imaging. Eur. J. Plant Pathol. 142(3), 625–632 (2015)
9. Jones, H.: Application of thermal imaging and infrared sensing in plant physiology and eco-
physiology. In: Advances in botanical Research, pp. 107–163. Academic Press (2004)
10. FLUKE Corporation. FLUKE TiS45 Infrared Camera. https://www.fluke.com/en-us/product/
thermal-cameras/tis45. (2019). Accessed 1 Dec 2019
11. Jain, A.K.: Data clustering: 50 years beyond K-means. Pattern Recogn. Lett. 31(8), 651–666
(2010)
12. Hunt, E.R., Doraiswamy, P., McMurtrey, J.E., Daughtry, C.S., Perry, E.M., Akhmedov, B.: A
visible band index for remote sensing leaf chlorophyll content at the canopy scale. Int. J. Appl.
Earth Obs. Geoinf. 21, 103–112 (2013)
13. Vapnik, V.: The nature of statistical learning theory. Springer Science and business media
(2013)
14. Zhang, W.Y., Wei, Z.W., Wang, B.H., Han, X.P.: Measuring mixing patterns in complex
networks by Spearman rank correlation coefficient. Physica A 451, 440–450 (2016)
Development of a Machine to Control
the Level of Washing in Panca Chili Seeds

Anthony De La Cruz , Jaime Cardenas , and Leonardo Vinces

Abstract The washing of Panca chili seeds requires innovative solutions that
allow controlling this process. It is necessary to handle variables (conductivity, pH,
colorimetry) in the face of the challenge of working with small seeds. At present,
there are no machines that are dedicated to the washing of this type of seeds, since in
many companies this work is done manually, which is not the one indicated because
this technique cannot guarantee homogeneity in the seed washing. In addition, direct
handling of this type of seeds can cause irritation to the eyes and skin of the person
who maintains contact with the seeds. That is why, it is proposed to make a machine
to scale by means of a motorized rotary agitator inside a tank, in order to guarantee
the homogeneity of the mixture when washing seeds. The present work will allow
to determine, among two different types of agitators (axial and radial), which type
of agitator is the most efficient in the washing of seeds of Panca chili, to achieve
this objective the measurement of pH and electrical conductivity to the water will be
carried after the mixture, after stirring. Finally, the analysis of the tests performed
on the mixture obtained and washed by each type of agitator allowed to identify the
turbine-type radial agitator, like the one that obtained greater efficiency in the washing
of seeds, with respect to the helical agitator and pallets, designed for development of
this work, in turn, could also confirm that this type of palette with the conductivity
control allows to guarantee the homogeneity of the mixture during washing.

Keywords Stirring · Mixing · Washing seeds · Capsaicin · pH measurement ·

Electrical conductivity measurement

A. De La Cruz · J. Cardenas · L. Vinces (B)

Universidad Peruana de Ciencias Aplicadas, Av. Prolongación Primavera 2390, Santiago de
Surco, Lima, Peru
e-mail: [email protected]
A. De La Cruz
e-mail: [email protected]
J. Cardenas
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 297
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_29
298 A. De La Cruz et al.

1 Introduction

Currently, chili pepper is the sixth most exported product in the non-traditional
agricultural sector. In 2017, Peru exported Capsicum for more than 238 million soles,
according to the Association of Exporters (ADEX). In addition, it is estimated that
the per capita consumption of chili peppers and peppers in Peru is 3.9 kilos per year,
the most prominent categories were canned chili peppers (57.4%) and dried chili
peppers (38.4%) [1]. Peru has a wide variety of chili peppers, as this is valued from
pre-Hispanic cultures for its high nutritional value, exquisite taste, and exotic aroma,
native peppers descend from common ancestors that originated in South America,
in heights of Bolivia and Peru or in southern Brazil, Peru being the country that has
the greatest diversity of this product in the world, among which is the Panca pepper
[2].
Seed oil production occurs from dried peppers such as the Panca pepper. This
type of pepper is grown in all regions of Peru, especially on the north coast, this
product is a food with low caloric intake since it is formed in almost 90% by water.
In addition, it has largely capsaicin. This component is what gives it the spicy flavor
and serves as an analgesic and anticoagulant, ideal for people at risk of cardiovascular
diseases. Likewise, its consumption helps to combat the pain generated by arthritis,
stimulates the nervous system, because it causes the body to produce endorphins,
regulates blood sugar levels, generates a bactericidal effect, eliminating stomach
bacteria, which helps decrease the possibility of stomach diseases [3].
It is thanks to these positive health effects that many products are made from chili
peppers, such as oil production, which uses the seeds of the Panca pepper as raw
material, which is carefully processed to get to the final product. Thus, to produce
chili seed oil it is necessary to perform the washing of the seeds, then the drying and
finally the dehydration of the seed. The great challenge within this whole process
is to perform the washing, this procedure should be the most appropriate so that at
the end of the process a clean seed is obtained and that it always loses a constant
amount of itching (capsaicin). Capsaicin is a compound that is found naturally in
fruits, although in different proportions. In chili, it usually varies between 0.1 and
1% by weight, although it seems small, that amount is enough to produce the typical
sensation of itching [4]. This compound is not uniformly distributed in the fruit; It
usually concentrates on the seeds and on the cover that surrounds them (pericarp),
these are the hottest areas of chili pepper [5].
The application of capsaicin on the skin or mucous membranes causes burning
and hyperalgesia, but repeated application leads to loss of sensitivity to capsaicin;
the application of higher doses causes a blockage of the C fibers that lead to a long-
term sensory deficit. This property has been used therapeutically in neuropathic
pain as an option when the other drugs are ineffective. Thus, capsaicin has demon-
strated its efficacy after repeated administration in postmastectomy pain, stump pain,
reflex sympathetic dystrophy, oral neuropathic pain, fibromyalgia, and especially in
diabetic neuropathy and postherpetic neuralgia. Consequently, there is a high interest
in finding new drugs analogous to capsaicin [6].
Development of a Machine to Control the Level of Washing … 299

While some studies have been developed, such as the one carried out in 1912, by
chemist Wilbur Scoville, who developed a scale with his surname, which measures
the degree of itch of a pepper. Scoville assigned the value of zero to sweet peppers,
which do not bite [6]. At the other end of the scale, he placed capsaicin to which he
gave a value of sixteen million, as the spiciest substance. At the time of washing the
seeds, the capsaicin molecules are mixed with the water reducing the concentration
of capsaicin in the seeds, therefore, the oil itching (final product) should decrease.
On the other hand, the continuous development of automatic and semi-automatic
machines has allowed the performance in the production, operational, and quality
control areas to be improved [7]. However, in the processing of seeds, to obtain
derivative products (oils, resins, etc.), great advances in automatic machines have not
been achieved. This is mainly due to the limitation of the sensors when controlling
variables [8]. For example, in Ecuador, the washing of sesame seeds is generally
also done manually, due to the complexity of handling the small size of the seed and
the little use of sesame seeds in this part. of the world. Given this, projects are also
being carried out that allow the washing of this type of seeds and in turn allow the
production needs to be met [9].
In the absence of a specific sensor intended to measure the concentration of
capsaicin in a mixture, it will be necessary to indirectly measure the amount of
capsaicin that is removed in the water during each wash. For this purpose, it has been
considered to measure the level of pH and electrical conductivity (Ppm) of the water
that has been mixed with the seeds, in this way it is expected to be able to control the
level of washing in the seeds, without affecting the organoleptic properties of this.
Additionally, it is guaranteed that the experiment will always keep the proportions
constant throughout the test (Rpm, exposure time, etc.), all with the intention of
guaranteeing the result of the tests performed.
The main objective of this study is to identify the design, process, and control of
variables that allow obtaining an efficient machine to wash the seeds of the pepper. To
achieve this objective, the washing tank and two types of agitators (axial and radial
flow) have been designed, the appropriate water temperature has also been identified,
the control of a variable that allows me to know the level of washing and the time
of agitation necessary for washing. With each of the agitators, a process of washing
pepper seeds will be carried out in a machine made to scale that allows controlling
the speed of rotation of the agitator, as well as the washing time for a mixture of two
hundred grams of seeds for a liter of water.

2 Description of the Design of the Machine for Washing

Chili Seeds

Figure 1 shows the block diagram of the design of the chili seed washing machine.
The details of each of the process steps will be described in the following sections.
300 A. De La Cruz et al.

Fig. 1 Block diagram of the machine design for chili seed washing

2.1 Characteristics of the Seeds

Figure 2 shows the diameter and thickness measurements of the chili seed.
In order to find the density of the seeds, we perform the corresponding division.
We divide the measured mass of 50 ml of test tube seeds over 50 ml. In order to
perform the following calculation, it is necessary to use the data in Fig. 3, which
represents the weight of 50 ml of chili seeds.
m Seeds
ρsemillas = (1)
50ml

Fig. 2 Dimensions of
diameter and thickness of
chili seeds

Fig. 3 Weight of 50 ml of
seed
Development of a Machine to Control the Level of Washing … 301

2.2 Design of Agitators

The base of the agitated tanks affects the efficiency of the mixture. Several base forms
are shown in Fig. 4. The base should be rounded at the edges instead of the flat; This
eliminates sharp corners and bags in which fluid flows may not penetrate, reducing
the formation of stagnant areas. The energy required to suspend solids in agitated
tanks is sensitive to the shape of the vessel base: according to the type of impeller
and the generated flow pattern, the modified geometries are shown in Fig. 4b through

Fig. 4 Different profiles for

the base of agitated vessels.
Extracted from [10]
302 A. De La Cruz et al.

(e) are can be used to improve particle suspension compared to the flat bottom tank
of Fig. 4a. In contrast, sloping sides or a conical base like the one shown in Fig. 4f
promote sedimentation of solids and should be avoided if a suspension of solids is
required [8].
To perform the sizing of the tank, it is necessary to know the total volume that
the tank must contain. To carry out the design of the washing tank, the use of two
hundred grams of seeds will be considered. The ratio between the mass of seeds
and water will be 1–5. Therefore, the mass of water needed to wash 200 g of seeds
will be 1000 g of water. Finally, the machine must be sized correctly considering the
appropriate volume of the container for raw material to be processed in addition to
assigning an additional percentage of volume for safety. The water ratio for washing
is as follows:

m Seeds = 5 × m Seeds (g) (2)

Volume calculation:
Step 1: Calculation of the volume of 200 g of seeds.

vSeeds = m Seeds × ρSeeds (3)

Step 2: Calculation of the volume of tolerance for washing.

Vtolerance = vSeeds × 0.25 (4)

Step 3: Calculation of the total volume.

Vtotal cylinder = Vtolerance + VSeeds + Vwashed water (5)

2.3 Design of Agitators

At this stage of the process, three types of agitators were designed, which are; straight
pallet agitator (Rushton), helical type agitator, and inclined pallet type agitator, which
are shown in Fig. 5a–c, respectively.
Table 1 presents the equations for the design of the three wash pallets. These three
palettes will be designed in Autodesk Inventor [9] and subsequently manufactured in
3D printers, as shown in Fig. 6. The design and manufacturing process was carried
out with the aim of testing in the next stage and thus being able to determine which
of them are the most suitable for washing chili seeds.
Development of a Machine to Control the Level of Washing … 303

Fig. 5 Types of agitators used in the project

Table 1 Patterns for agitator design

Agitator Di /DT HL /DT Ci /DT WBF /DT Number
Turbine Rushton 0.33 1 0.33 0.1 6
WB /Di = 0.2, L B /Di = 0.25
Inclined turbine 0.33 1 0.33 0.1 6
WB /Di = 0.2, L B /Di = 0.125
6 blades, 45°
DT /Di C/Di Hi /Di WB /Di
Helical 1.02 0.01 1 0.1

Fig. 6 Agitators designed in the project

2.4 Test Design

In order to establish the level of washing of the seeds, it is necessary to define some
measurement variables. For this, the seeds will not be directly analyzed, the water
in the mixture will be analyzed at the time of washing. The level of pH and the
level of electrical conductivity in the water will be measured after each wash. It is
considered that, by washing the seeds, the water can become more acidic. This is
mainly due to the substances that expel the seeds. Seed washing is a cyclic process,
which is repeated approximately every 15 min. This process is detailed below with
the following steps described.
304 A. De La Cruz et al.

Table 2 Washing process specification

Washed Washed 1 Washed 2 Washed 3
Sample Sample 1 Sample 4 Sample 7
Sample 2 Sample 5 Sample 8
Sample 3 Sample 6 Sample 9
Time 15 min
RPM 400
Temperature 25 °C

Step 1. Income of raw material selected and weighed previously (200 g of seeds
and 1000 ml of water).
Step 2. Start of washing for a time interval of 15 min.
Step 3. After 15 min of washing, proceed to remove a sample of 50 ml where pH
and Electrical conductivity should be measured.
Step 4. Start of the second washing period for 15 min.
Step 5. After 15 min of washing, proceed to remove a sample of 50 ml where it is
necessary to measure pH and Electrical conductivity.
Step 6. The water in the mixture is changed to new (clean) water. With this change,
we will proceed to what we will call “Second Wash” and stages 2, 3, 4, 5,
and 6 should be repeated until the appropriate wash number is found, this
will be achieved with the precise measurement of the variables described
above. Table 2 shows the specifications of the washing process.

3 Results

The following graphs show the measured values of conductivity and pH obtained
in each sample during the washing tests, the behavior of the three pallets is also
observed at the time of washing (agitating) the seeds. Figure 7 shows that the level
of pH between the samples of different pallets is very similar, that is, there is no
significant change in the tests, despite the change of pallets. However, there is a
significant increase in pH every three samples, that is, after changing the water. This
allows us to know the number of washes in which our process is located.
In Fig. 8, it is observed that the conductivity measurement in each sample, of
different agitator, has a significant change, that the change of pallet allows obtaining a
higher conductivity value. The straight pallet is the one that allows obtaining a higher
conductivity value in the samples. Secondly, there is the inclined pallet and thirdly the
helical pallet. For the final design of the washing machine, it is recommended to use
a straight pallet agitator because it allows for greater agitation and the construction
is much simpler than the other two designs.
Additionally, it was observed that the color of the water changes during washing,
as the agitation of the seeds increases, and the color decreases each time the water
Development of a Machine to Control the Level of Washing … 305

Graph of the measurement of pH vs. Samples

according to the type of palette
8
Ph measurement 7
6
5
4 Straight pallets
3
Helical pallets
2
Inclined pallets
1
0
1 2 3 4 5 6 7 8 9
Samples

Fig. 7 Graph of the measurement of pH versus Samples according to the type of palette

Conductivity Chart vs. Samples according to the type

of palette
1200
Straight pallets
1000 Helical pallets
Inclined pallets
Conductivity

800

600

400

200

0
1 2 3 4 5 6 7 8 9
Samples

Fig. 8 Conductivity chart versus Samples according to the type of palette

change is made. The control of the level of color in the water could help us to control
the level of agitation of the seeds, in this way we could know when to stop stirring
the seeds. This can be determined in Fig. 9, since the samples obtained during the
washing process, can be observed.

4 Conclusions

The tests carried out determined that the best agitator to wash seeds is that of straight
blades, the performance of the blades is shown in Fig. 8. For this test, the time
306 A. De La Cruz et al.

Fig. 9 Images of the samples obtained in the washing process

established to agitate the seeds was 15 min, after this time a variation (increase)
of conductivity in the water is obtained. This increase was observed each time the
mixture was stirred more. However, the second variable we control (pH) does not
provide much information, since it only varies each time the water is changed. An
additional test was performed where hot water (40 °C) was used, as a result, the
number of washes decreased. This is mainly since the conductivity in the water
reaches, in less time of agitation, the values obtained in the first test. It is important
to mention that the water obtained during each wash had a reddish color and this
increased every time the mixture was stirred; in this way, the water could be analyzed
by image processing in order to detect the level of washing.

References

1. A. de E. ADEX: Perú puede mejorar su posicionamiento como proveedor mundial de

pimientos y ajíes. Asociación de exportadores. IOP Web. http://www.adexperu.org.pe/not
adeprensa/peru-puede-mejorar-su-posicionamiento-como-proveedor-mundial-de-pimientos-
y-ajies/. (2017) Accessed on 9 May 2019
2. Amazon Andes: Ají panca. Amazon Andes. IOP Web. https://peruvianchili.com/es/product-
item/aji-panca/. (2017). Accessed on 9 June 2019
Development of a Machine to Control the Level of Washing … 307

3. Comisión de Promoción del Perú para la Exportación y el Turismo: Súper ajíes nativos. Super
Foods Peru. IOP Web. https://peru.info/es-pe/superfoods/detalle/super-ajies-nativos. Accessed
on 17 June 2019
4. Cedrón, J.C.: La Capsaicina la molécula destacada. Revista de Química Pontificia Universidad
Católica del Perú 27, 816–824 (2013)
5. Yañez, P., Balseca, D., Rivadeneyra, L., Larenas, C.: Características morfológicas y de concen-
tración de capsaicina en cinco especies nativas del género Capsicum cultivadas en Ecuador. La
Granja: Revista de Ciencias de La Vida. Universidad Politécnica Salesiana, (2015)
6. Peralta, G.: Determinación del nivel de pungencia en unidades Scoville para Capsicum annum
var. aviculare procedente de regiones productoras de Guatemala. Informe de tésis, Universidad
de San Carlos de Guatemala, (2007)
7. Medina, R., Atencio, B., Romero, M., Castro, R.: Análisis estratégico del proceso productivo
en el sector industrial. Rev. Ciencias Sociales 8(1), 135–156 (2002)
8. Suntaxi, J.: Diseño y construcción de una máquina lavadora de ajonjolí para la empresa
productos Schullo con una capacidad de procesamiento de 100 Kg/h. Informe de Tésis
Universidad Politécnica Salesiana de Quito (2016)
9. Søpstad, S., Imenes, K., Johannessen, E.: Hybrid electrochemical sensor platform for capsaicin
determination using coarsely stepped cyclic squarewave voltammetry. Biosens. Bioelectron.
130, 374–381 (2019)
10. Doran, P.: Bioprocess Engineering Principles. Academic Press, (2013)
Bacteriological Contaminants Detection
in Some Consumed Raw Vegetables
Based on Spectrometry in Ayacucho
Town, Peru

José Yauri , Manuel Lagos , Rousell Vidalón , Julio Cárdenas ,

Kay Jerí , Lisber Arana , and Yuzo Iano

Abstract Humans need to eat foods for growing and getting energy to carry out your
daily activities. However, to find quality foods have become difficult because most of
them have been cultivated in contaminated environments. So, to detect contaminated
foods is a critical issue to care the human health. Being microorganisms the most
frequent sour of contamination, it is desirable to have a detector tool for discarding
unsanitary foods. In opposite to the traditional, expensive, slow, and invasive microbi-
ological analysis, approaches to quickly detect microorganisms have been proposed.
This work presents an approach to use spectrophotometry to detect bacteria in three
consumed raw vegetables and expended in Ayacucho tow, Peru. Cultures of lettuce,
peppermint, and cilantro irrigated with water with a high presence of Escherichia
coli and Thermotolerant coliforms were successfully detected using the measured

J. Yauri (B) · J. Cárdenas · K. Jerí

School of Technological Sciences and Engineering, University of Ayacucho Federico Froebel,
Ayacucho, Peru
e-mail: [email protected]
J. Cárdenas
e-mail: [email protected]
K. Jerí
e-mail: [email protected]
M. Lagos
Department of Mathematics and Physics, National University of San Cristobal de Huamanga,
Ayacucho, Peru
e-mail: [email protected]
R. Vidalón
Technical-Productive Education Center, Polytechnic of Ayacucho, Ayacucho, Peru
e-mail: [email protected]
J. Yauri · L. Arana · Y. Iano
School of Electrical and Computer Engineering, University of Campinas, Campinas, Brazil
e-mail: [email protected]
Y. Iano
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 309
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_30
310 J. Yauri et al.

spectrum from only their stem. Using an SVM classifier, we achieved a detection
accuracy rate of 91.89% and Kappa 89.87%, which demonstrates the feasibility of
our approach.

Keywords Bacteriological contaminants · Spectrometry · Consumed raw

vegetables · Escherichia coli · Thermotolerant coliforms

1 Introduction

Daily, humans need to eat food to obtain nutrients, proteins, and vitamins that allows
to have a healthy life and get the energy required to develop their activities [1]. But
having poor nutrition increases vulnerability to diseases alters the normal physical
and mental development, and diminishes productivity. Therefore, people need to eat
nutrient-rich foods, aiming to cover their dietary requirements [2].
Today, getting clean and healthy food is becoming hard –healthy foods are varied
in nutrient-rich foods such as fruits, vegetables, cereals, etc. which have been grown
preferably naturally, in accordance with the environment [3]. Because of the pollution
and bad agricultural practices, more crops are contaminated by pathogens [4], and
organic and inorganic residuals [5, 6]. Usually, farms are cultivated with contaminated
soil and streams, which contain residual waters from the population and livestock
[7–9], rest of chemical fertilizers [10] and pesticides [11]. The consumption of such
foods can cause health problems, usually infectious diseases of the stomach, which
are commonly observed in children and elderly people [12, 13].
In order to mitigate the action of pathogens, it is recommended a good hygiene
and washing practices for consumed raw fruits and vegetables, since recent studies
have concluded that the pathogen immigration on the leaf surface or pathogen inter-
nalization within the plant augment their environment survival [14]. Thus, many
studies have been conducted in favor to identify the type of contamination in soils
[9], water [6–8], fruits and leafy vegetables [11, 12, 14]. Identification and measure
of bacterial contamination are carried out by microbiological analysis. In such a
process, bacteria are identified, cultured and counted in a laboratory [15]. However,
the analysis for detecting pathogenic microorganisms of a single product is complex,
time-consuming and expensive. So, it is desirable to obtain a general contamination
profile, since some vegetables are more contaminated than others, depending on its
origin.
Recently, powered by new sensing technologies, novel methods for rapid detection
of bacteriological contamination in vegetables were presented [16]. A rapid detection
method has to be simple, rapid, and sensitive. Lately, rapid non-destructive detection
methods are based on bio-sensors and spectroscopy. The work [17] reported the
use of bio-sensors to detect effectively Pseudomonas aeruginosa and Salmonella
typhimurium, [18] use a laser ray to detect microorganisms, and in [19] used the
near-infrared (NIR) spectral as an indicator of contamination livestock products.
Bacteriological Contaminants Detection in Some Consumed … 311

This paper presents a method to detect the presence of bacteriological contam-

inants in three consumed raw vegetables (lettuce, cilantro, peppermint). To learn a
discriminative model, the vegetables come from two different places: the first, a field
irrigated with the river water that exceeds the permissible limits of biological organ-
isms for irrigation [20]; the second, a home garden irrigated with drinking water. No
fertilizer and no pesticides were applied. In harvest time, vegetables were collected
and measured their reflectance spectrum using a spectrophotometer [21]. By using
both of the visible and NIR measured wavelength as feature descriptors, trained a
linear Support Vector Machine (SVM), we achieved encouraging results as a fast
bacteriological contamination detector.
The remainder of this paper is organized as follows. Section 2 presents related
work. Section 3 explains the proposed method. Section 4 details the experimental
results and discussions. Finally, Sect. 5 exposes conclusions and future work.

2 Related Work

Humans need to eat foods to satisfy the nutritional requirements that allow them to
make activities. Among foods, vegetables and fruits are the most recommended,
because they content-rich carbohydrates, vitamins, and minerals [3]. However,
usually consumed raw vegetables or fruits can cause infectious diseases since they
were contaminated or handled hygienically [14]. The presence and degree of contam-
inants usually depend on the environmental conditions where the growth of the food
[22]. The contamination could come from the used water, soil, and fertilization type
employed during farming. Therefore, vegetables cultivated with the residual stream
are more contaminated than vegetables irrigated using water compliant with irrigation
water quality.
Because eating contaminated foods can produce undesirable diseases, the detec-
tion and discarding of contaminated foods can avoid them. Usually, bacteriological
contamination such as Escherichia coli is among the most common cause of stomach
and intestinal infections, recently even antibiotic-resistant [23]. Therefore, people
require methods to detect contaminated food.
In opposite to traditional detection methods based on biochemical analysis which
is lagging and costly, recent methods are faster and credible at low cost to detect
bacteriological contaminants [16–18]. In this sense, spectrometry seems a promising
method to detect microorganisms [19] and other properties of the object of interest.
Spectroscopy is concerned about the interaction between the light and the object
of interest and the spectrometry measures the electromagnetic radiation to obtain
information about the object [24]. Normally, a spectrometer provides a measurement
of the reflection of electromagnetic radiation, named reflectance [21], usually from
the visible and near-infrared spectrum–NIR (between 350 and 2500 nm).
Spectrometry has many applications. For instance in agriculture, instead of a soil
chemical test, the measurements of a spectrometer can provide data to determine
312 J. Yauri et al.

the contamination level and type of particles in the soil [25]; can be detected micro-
bial contamination in water [26] and pesticide residues in lettuces [27]; also plant
diseases can be detected as the yellowing leaves caused by the rhizospheric bacteria
in chrysanthemum [28], the leaf miner damage on tomato leaf [29], and fungal
pathogens on sugar beet leaves [30]. All these presented diseases are produced by
some microbiological pathogen, so is feasible to use a spectrometer to detect it.
Taking advantage of the spectrometry, in this work, we present an automatic
recognizer to detect bacteriological contaminants in three consumed raw vegetables
in Ayacucho town, Peru. Lettuces, cilantro, and peppermints were cultivated using
highly contaminated water with microbiological pathogens, such as Escherichia coli
and thermotolerant coliforms. According to the official report [20], the irrigation
water of the field contains 2.2E + 5 CFU/100 of Escherichia coli and 1.7E + 6
NMP/100 mL of thermotolerant coliforms, which is above the permissible 1000
CFUs for irrigation water.

3 Our Proposal

Ayacucho is a town located in the central part of Peru. There, it is common to eat
raw vegetables as an accompaniment to typical dishes. However, to eat that food is
becoming a health problem for citizens: most fruits and vegetables are irrigated with
wastewater of Ayacucho city. The residual treatment plant flows its residual waters
into the Alameda river, without proper treatment to make it suitable for irrigation, so
the river and the basin are contaminated.
Among contaminants, we highlight Escherichia coli and thermotolerant coliforms
reported in [20], because they cause digestive intoxication, and in a long term cancer
(recently, the Healthy Ministry of Peru communicated a high incidence of stomach
cancer in Ayacucho [31]). In the market, to detect and distinguish a vegetable from
a contaminated or non-contaminated field is hard, since microorganisms are imper-
ceptible to the human eye and because the sellers usually do not say the origin of the
product. Therefore, people are at risk of getting sick. For instance, Fig. 1 shows two
lettuce samples: one is from a field irrigated with contaminated water (Fig. 1a) and
then another one, from a field, irrigated with potable water (Fig. 1b), however, they
seem healthy.
To build an automatic detector, to discriminate between contaminated and non-
contaminated vegetables, the proposed framework is shown in Fig. 2. It works as
follows:
Vegetable dataset. The collected dataset consists of wavelengths of the elec-
tromagnetic spectrum measured with a spectrometer [21] for each vegetable
sample.
The used sensor provides measurements of the reflectance of the sunlight between
350–2500 nm, e.g., the visible and the near-infrared spectrum. There- fore, measures
were performed at 12 m–01 p.m., on different sunny days outdoor.
Vegetables consist of lettuce, cilantro, and peppermint and form two groups:
Bacteriological Contaminants Detection in Some Consumed … 313

Fig. 1 Lettuce samples. a From a contaminated field, b From a non-contaminated field

Fig. 2 The proposed detection framework

• The first group was cultivated in a field at the Latitude = − 13.136103 and Longi-
tude = − 74.193728, irrigated using residual water coming from the wastew-
ater treatment plant of Ayacucho city, located at the Latitude = − 13.140945,
Longitude = − 74.206654. Samples were labeled as contaminated.
• The second group was cultivated in a field at the Latitude = − 13.164441 and
Longitude = − 74.248020, irrigated using potable water. Samples were labeled
as non-contaminated.

Data pre-processing For this stage, each electromagnetic spectral measurement can
be thought of as a row vector. Because of the sun intensity, the moisture and the wind
at the time of measurement can produce noise, we applied two pre-processing steps:
• To remove that noisy water band, approximately 1800–2000 nm [21], so the vector
length is 1753–dim.
• To normalize the measurement according to Eq. 1, since the spectrum wavelengths
could have variations in their intensity, but nothing in their shape.

x − xmin
xn = (1)
xmax − xmin

Feature extraction Since the sensor provides a single row data for each spectral
measurement, each of one is treated as a row feature vector.
314 J. Yauri et al.

Classification To perform detection tasks, we trained an SVM classifier in one-

against-one fashion [32]. For classification, we have six classes lettuce-co, lettuce-
nc, cilantro-co, cilantro-nc, peppermint-co, and peppermint-nc, where co means
contaminated and nc, non-contaminated.

Detection This stage recognizes the given unknown vegetable and labels it as
contaminated or non-contaminated.

4 Results and Discussions

4.1 Spectral Wavelengths Dataset

We collected a dataset which contains spectral wavelengths from 20 samples of

lettuce, 10 of cilantro, and 10 of peppermint, both from contaminated and non-
contaminated fields, and both from the leaves and the stems. An illustration of the
measurement process to collect data is shown in Fig. 3.
Next, in order to remove noisy data, we discarded three segments from the
collected raw data: 1340–1350 nm, 1800–2000, and 2400–2500 nm. They are
produced by uncontrolled changes in moisture and wind during measurements. Next,
the data is normalized. Figure 4 illustrates the pre-processing step: the raw data
Fig. 4a, the raw data is cropped Fig. 4b and the raw data is normalized Fig. 4c.
We collected data from both the leaves and the stems of the vegetables. Figure 5
shows the plotted data as 1D signals for both the leaves and stem of lettuce,
peppermint, and cilantro.

Fig. 3 Electromagnetic spectrum measurement: a Lettuce, b Cilantro, and c permint

Bacteriological Contaminants Detection in Some Consumed … 315

Fig. 4 Pre-processing of the data: a Raw data signal, b Raw data after removing some useless
bands, and c Normalized data

4.2 Experiments

In order to assess the discriminative power of the measured electromagnetic spectral

data (both the visible and the near-infrared data), the classification was performed
changing the features in three ways: (I) Only measurements from the leaves, (II)
Only measurements from the stem, and (III) Concatenating of I and II. To reduce
the effect of a small dataset, experiments were conducted using the Leave-One-Out
Cross-Validation (LOOCV) scheme [32].

4.3 Results

Table 1 shows the obtained results using the Linear SVM (Cost = 1) under the
LOOCV experimental setup.
According to Table 1, the experimental setup II outperforms the other approaches,
achieving a detection accuracy rate of 91.89% and a Kappa statistic measure rate of
89.87%. Using only leaf features leads to lower performance, and even contrary to
be expected, combined features get the worst performance.
These results suggest that the classifier has detected microorganisms on the stem
of these vegetables but no on their leaf. It is reasonable since, during the irrigation
the residual water covers up to the stem of the plant, microorganisms remain on the
stem and over other parts that got wet. And few bacteria live on the leaves because
the water does not contact them and because of the bactericidal effect of the light
[33].

5 Conclusions and Future Work

This paper presented an approach to detect bacteriological contaminants in three

consumed raw vegetables, named lettuce, peppermint, and cilantro, in Ayacucho
316 J. Yauri et al.

tow, Peru. The proposal resides on the use of spectroscopy to measure the reflected
light spectrum of the object and since spectrophotometry is a low cost, fast, and
non-invasive tool of analysis. By measuring only the reflectance on the stem, the
detector achieved a recognition rate of 91.89% in a collected dataset from a highly
contaminated basin with bacteriological pathogens. It demonstrates his feasibility in
favor of food safety and people’s health in Peru.

Fig. 5 Illustration fo the spectral data both from the leaf and stem from a sample: a Lettuce leaf,
b Lettuce stem, c Peppermint leaf, d Peppermint stem, e Cilantro leaf, and f Cilantro stem. The
blue signal represents the contaminated data, while the red signal shows the non-contaminated one
Bacteriological Contaminants Detection in Some Consumed … 317

Table 1 Experimental results according to their accuracy and Kappa metrics

Protocol Accuracy (%) Kappa (%)
(I) Only leaf features 84.00 80.19
(II) Only stem features 91.89 89.87
(III) Both I and II 83.56 79.57

As future work, we intend to extend this study to other vegetables and to establish
the degree of contamination.

Acknowledgements This research was partially supported by the University of Ayacucho Federico
Froebel. The authors would like to thank Walter Solano, Renato Soca, and Noel Torres from the
Laboratory of Remote Sensing of the National University of San Cristobal de Huamanga, and the
anonymous reviewers for their valuable suggestions and helpful feedback.

References

1. World Health Organization: Health Topics: Nutrition (2014). https://www.who.int/topics/nut

rition. Accessed on 09 Sep 2019
2. Hernández Rodríguez, M., Sastre Gallego, A.: Tratado de Nutrición. Ediciones Díaz de Santos
(2000)
3. Mañas Almendros, M., Martínez de Victoria Muñoz, E., Yago Torregrosa, M.D.: Principios
Generales de Nutrición. Ediciones Díaz de Santos (2000)
4. Steele, M., Odumeru, J.: Irrigation water as source of foodborne pathogens on fruit and vegeta-
bles. J. Food Protect. 67(12), 2839–2849 (2004). http://jfoodprotection.org/doi/abs/10.4315/
0362-028X-67.12.2839
5. Kumar, V., Thakur, R.K., Kumar, P.: Assessment of heavy metals uptake by cauliflower (Bras-
sica oleracea var. botrytis) grown in integrated industrial effluent irrigated soils: a prediction
modeling study. Sci. Horticult. 257, 108682 (2019). https://www.sciencedirect.com/science/
article/pii/S0304423819305680
6. Weldegebriel, Y., Chandravanshi, B.S., Wondimu, T.: Concentration levels of metals in
vegetables grown in soils irrigated with river water in Addis Ababa, Ethiopia. Ecotoxicol.
Environ. Safe. 77, 57–63 (2012). https://www.sciencedirect.com/science/article/pii/S01476513
11003368
7. Okafo, C.N., Umoh, V.J., Galadima, M.: Occurrence of pathogens on vegetables harvested
from soils irrigated with contaminated streams. Science Total Environ. 311(1–3), 49–56 (2003).
https://www.sciencedirect.com/science/article/pii/S0048969703000573
8. Ensink, J.H.J., Mahmood, T., Dalsgaard, A.: Wastewater-irrigated vegetables: market handling
versus irrigation water quality. Trop. Med. Int. Health 12, 2–7 (2007). http://doi.wiley.com/10.
1111/j.1365-3156.2007.01935.x
9. Rostami, A., Ebrahimi, M., Mehravar, S., Fallah Omrani, V., Fallahi, S., Behniafar, H.: Contam-
ination of commonly consumed raw vegetables with soil transmit- ted helminth eggs in Mazan-
daran province, northern Iran. Int. J. Food Microbiol. 225, 54–58 (2016). https://www.scienc
edirect.com/science/article/pii/S0168160516301180
10. Sultana, Z., Sultana, R., Rokonujjaman, M., Simol, H.A., Sultan, M.Z., Salam, M.A.: Impact
of gypsum and potash fertilizers on heavy metals and nutrients levels in some selected leafy
vegetables and assessment of potential health risk. Bangladesh Pharm. J. 22(1), 56–67 (2019).
https://www.banglajol.info/index.php/BPJ/article/view/40074
318 J. Yauri et al.

11. Ng, P.J., Fleet, G.H., Heard, G.M.: Pesticides as a source of microbial contamination of salad
vegetables. Int. J. Food Microbiol. 101(2), 237–250 (2005). https://www.sciencedirect.com/sci
ence/article/pii/S0168160505000097
12. Mercanoglu Taban, B., Halkman, A.K.: Do leafy green gegetables and their ready-to-eat [RTE]
salads carry a risk of foodborne pathogens?. Anaerobe 17(6), 286–287 (2011). https://www.
sciencedirect.com/science/article/pii/S1075996411000552
13. American Academy of Pediatrics: Food Poisoning and Food Contamination (2014). https://
www.healthychildren.org/english/health-issues/conditions/abdominal/pages/food-poisoning-
and-food-contamination.aspx. Accessed on 09 Sep 2019
14. Heaton, J., Jones, K.: Microbial contamination of fruit and vegetables and the behavior of
enteropathogens in the phyllosphere: a review. J. Appl. Microbiol. 104(3), 613–626 (2008).
http://doi.wiley.com/10.1111/j.1365-2672.2007.03587.x
15. Lightfoot, N.F., Maier, E.A.: Microbiological Analysis of Food and Water: Guidelines for
Quality Assurance. Elsevier (1999)
16. Law, J.W.F., Ab Mutalib, N.S., Chan, K.G., Lee, L.H.: Rapid methods for the detection of food-
borne bacterial pathogens: principles, applications, advantages and limitations. Front. Micro-
biol. 5, 770 (2014), http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=PMC429063
1http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=PMC4290631
17. Starodub, N.F., Novgorodova, O., Ogorodnijchuk, Y.: Biosensors and express control of bacte-
rial contamination of different environmental objects. Microbial Contam. Food Degradation,
367–394 (2018). https://www.sciencedirect.com/science/article/pii/B9780128115152000123
18. Yoon, J., Lee, K., Park, Y.: A simple and rapid method for detecting living microorganisms
in food using laser speckle decorrelation. https://arxiv.org/abs/1603.07343 (2016). http://arxiv.
org/abs/1603.07343
19. Atanassova, S., Veleva, P., Stoyanchev, T.: Near-infrared spectral informative indicators for
meat and dairy products, bacterial contamination, and freshness evaluation. Microbial Contam.
Food Degradation, 315–340 (2018). https://www.sciencedirect.com/science/article/pii/B97801
2811515200010X
20. ALA Ayacucho: Results of Participatory Monitoring of Surface Water Quality in the Mantaro
Basin—2017, Local Water Administration of Ayacucho, (2017). https://www.ana.gob.pe/org
anos-desconcentrados/aaa-mantaro/ala-ayacucho
21. ASD Inc.: ASD Technical Guide, Analytical Spectral Devices Inc., (1999)
22. Alemu, G., Mama, M., Siraj, M.: Bacterial contamination of vegetables sold in Arba Minch
town, Southern Ethiopia. BMC Res. Notes 11(1), 775 (2018). https://bmcresnotes.biomedcen
tral.com/articles/10.1186/s13104-018-3889-1
23. Lillie, P., Johnson, G., Ivan, M., Barlow, G., Moss, P.: Escherichia coli bloodstream infec-
tion outcomes and preventability: a six-month prospective observational study. J. Hosp.
Infect. 103(2), 128–133 (2019). https://www.sciencedirect.com/science/article/pii/S01956701
19302208
24. Ozaki, Y., McClure, W.F., Christy, A.A.: Near-Infrared Spectroscopy in Food Science and
Technology. Wiley, Hoboken, NJ, USA (2006). http://doi.wiley.com/10.1002/0470047704
25. Schwartz, G., Eshel, G., Ben-Dor, E.: Reflectance spectroscopy as a tool for moni-
toring contaminated soils (Chap. 4). In: Pascucci, S. (ed.) Soil Contamination. InTech
(2011). http://www.intechopen.com/books/soil-contamination/reflectance-spectroscopy-as-a-
tool-for-monitoring-contaminated-soils
26. Jung, A.V., Le Cann, P., Roig, B., Thomas, O., Baur‘es, E., Thomas, M.F.: Microbial contami-
nation detection in water resources: interest of current optical methods, trends and needs in the
context of climate change. Int. J. Environ. Res. Public Health 11(4), 4292–4310 (2014). http://
www.mdpi.com/1660-4601/11/4/4292
27. Sun, J., Ge, X., Wu, X., Dai, C., Yang, N.: Identification of pesticide residues in lettuce leaves
based on near-infrared transmission spectroscopy. J. Food Process Eng. 41(6), e12816 (2018).
http://doi.wiley.com/10.1111/jfpe.12816
28. Pérez Naranjo, J.C., Soler Arango, J., Arango Pulgarín, G., Meneses Ospina, E., Ruiz
Villadiego, O.S., Villadiego, O.S.R.: Espectroscopía NIR como Técnica Exploratoria Rápida
Bacteriological Contaminants Detection in Some Consumed … 319

para Detección de Amarillamiento de Hojas Crisantemo (Dendranthema grandiflora var.

Zembla). Revista Facultad Nacional de Agronomía Medellín 67(1), 7163–7168 (2014). https://
revistas.unal.edu.co/index.php/refame/article/view/42629
29. Xu, H., Ying, Y., Fu, X., Zhu, S.: Near-infrared spectroscopy in detecting leaf miner damage
on tomato leaf. Biosyst. Eng. 96(4), 447–454 (2007). https://www.sciencedirect.com/science/
article/pii/S1537511007000268
30. Mahlein, A.K., Steiner, U., Dehne, H.W., Oerke, E.C.: Spectral signatures of sugar beet leaves
for the detection and differentiation of diseases. Precis. Agricult. 11(4), 413–431 (2010). http://
link.springer.com/10.1007/s11119-010-9180-7
31. MINSA Peru: Analysis of the Cancer Situation in Peru 2013. Technical Report, Ministry of
Health of Peru—General Direction of Epidemiology, Lima (2013). http://www.dge.gob.pe
32. Murphy, K.P.: Machine Learning: A Probabilistic Perspective. The MIT Press, Cambridge,
MA (2012)
33. Lubart, R., Lipovski, A., Nitzan, Y., Friedmann, H.: A possible mechanism for the bactericidal
effect of visible light. Laser Ther. 20(1), 17 (2011). http://www.pubmedcentral.nih.gov/articl
erender.fcgi?artid=PMC3806074
Feasibility in Using Banana Flour
in Bread Production: Centesimal
and Sensory Analysis

Nathália Rafaela Santos de Carvalho ,

Maria Thereza de Moraes Gomes Rosa ,
Isabela Francabandiera Tornizielo ,
and Daniela Helena Pelegrine Guimarães

Abstract Brazil produces about six million tonnes per year of banana (Musa spp),
with 35 kg/inhabitant/year average consumption. Banana peel represents about 47–
50% (w/w) of ripe fruit but it doesn’t have industrial applications; sometimes it is
used in animal feed but in small-scale. On the other hand, banana flour already has
proved to be a very promising ingredient, is used in baking, diet products, children’s
food, and animal feed. Based on these benefits and considering the great banana
productivity in Brazil, the present work proposes the development of formulations of
bread from banana flour, as well as analyze the sensory parameters, where acceptance
tests were carried out, with a hedonic scale of 9 points; the attributes evaluated:
flavor, appearance, texture, aroma, and overall impression. The results showed that
the formulations, in general, have been accepted, except that produced by a mixture
composed of 70% banana flour and 30% wheat flour, which presented a significant
difference in relations with the other, for all attributes analyzed.

Keywords Banana · Flour · Baking · Sensorial analysis · Centesimal analysis ·

Acceptance

N. R. S. de Carvalho · D. H. P. Guimarães (B)

Lorena Engineering School (EEL-USP), São Paulo University, Campinho Highway, 12602-810
Lorena, SP, Brazil
e-mail: [email protected]
M. T. de Moraes Gomes Rosa · I. F. Tornizielo
Center of Science and Technology, Mackenzie Presbyterian University, Brazil Avenue 1220,
Campinas, SP 13073-148, Brazil
N. R. S. de Carvalho · M. T. de Moraes Gomes Rosa · D. H. P. Guimarães
The University of Taubaté (UNITAU), Taubaté, SP, Brazil

© The Editor(s) (if applicable) and The Author(s), under exclusive license 321
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_31
322 N. R. S. de Carvalho et al.

1 Introduction

Banana (Musa spp) is one of the most consumed fruits in the world, whose consump-
tion is increased every year, due to its high energy content (about 100 kcal per 100 g
of pulp), for presenting easily assimilable carbohydrates, and also for being rich in
vitamins, potassium, phosphorus, calcium, and iron, when compared to apples or
oranges [1].
In green bananas, the main component is the starch (55–93% of the total solids
content), and as the fruit matures, the starch is converted to sugars (glucose, fructose,
and sucrose), reducing the acidity and, consequently, increasing sweetness [2].
The concentration of starch in the green banana makes it of great industrial interest,
and beside the fact that its shell surpasses the pulp content of some nutrients, this
organic residue became an inexpensive source of carbohydrates and minerals [3]. In
this context, the flour of the green banana peel appears in the bakery industry, which
can be used as supplementation or in partial substitution of wheat flour. Moreover,
the fact that banana flour is not digested by the small intestine makes it very efficient
in diets to eliminate weight [4].
Bread is one of the most popular foods and one of the main caloric sources of diet
in many countries and it is, therefore, the subject of many types of research, in order
to add nutritional value to the product, either by enhancing the nutrient content or by
correcting its deficiencies [5].
The consumer market for bread has grown, varying from region to region. The
Southeast and South Regions of Brazil consume around 35 kg/inhabitant/year, while
the Northeast reaches only 10 kg/inhabitant/year. The consumption of bread in Brazil
is 27 kg per year per person, which represents half of the proportion recommended
by world organizations such as the WHO—60 kg/inhabitant/year—, and the FAO—
50 kg/inhabitant/year. Baked goods occupy third place in the Brazilian shopping list,
representing, on average, 12% of the family budget for food [6].
Considering the great interest in the industrialization of green banana peel,
together with the great popularity of bakery products, the present work1 aims to
analyze the viability of the use of banana flour in the development of form loaves
formulations, alternative nutritional enrichment of the product, as regards the fiber
content. The different formulations were tested by means of sensorial acceptance
tests.

2 Materials and Methods

In order to obtain banana flour, bananas of the variety nanica (Musa acuminata’Dwarf
Cavendish’) were used in the ¾ fat development stage, that is, still green. First, the
fruits were dehydrated in tray drier (Macanuda, MS-P), and then crushed in the
industrial crusher (Metvisa, LQ-25). The obtained flour was evaluated for moisture,

1 This research was carried out by the author at the University of Taubaté.
Feasibility in Using Banana Flour … 323

protein content, ashes, and fat. The bread loaf was prepared by adding the banana
flour, partially replacing the wheat flour, following different formulations, in relation
to the proportions of brown flour/flour.
The sensorial attributes of the different bread formulations (flavor, appearance,
texture, aroma and overall impression) were analyzed by means of acceptance tests
with a hedonic scale of 9 points of conducting a survey are the agility and economy
in obtaining data, as well as obtaining data grouped in tables, which allow a rich
statistical analysis.

2.1 Characterization of the Banana Flour

The processed banana flour was initially characterized, and physicochemical anal-
yses were performed to determine its centesimal composition, and the following
physicochemical analyses were performed:
Moisture [7—Method 16192];
Ashes [7—Method 16196];
Total Lipids [8];
Proteins [7—Method 38012].
The analyses concerning the characterization of the banana flour were carried out
in triplicate, to verify if there is a divergence of results in the repetitions.

2.2 Processing of Loaves

The preparation of this type of bread followed the basic formulation composed of
flour, biological yeast, eggs, salt, sugar, and milk. Instead of wheat flour, it was
used a mixture composed of BANANA FLOUR + WHEAT FLOUR in different
proportions, as follows:
F1: 100% wheat flour
F2: 50% wheat flour + 50% banana flour
F3: 70% wheat flour + 30% banana flour
F4: 30% wheat flour + 70% banana flour.
The flour, together with the biological yeast, was sieved to be added in a spiral
kneader (Paniz, AE-25) and homogenized to the mixture composed of eggs, sugar
salt, and milk. Thereafter, the dough was shaped and baked at 180 °C for 30 min.

2.3 Sensory Evaluation

In the present work, the different formulations for bread loaf were subjected to
sensory analysis for appearance, flavor, and aroma, with reference to the product
324 N. R. S. de Carvalho et al.

prepared in a conventional manner (i.e., from wheat flour), the samples served in
single portions, in white disposable plates and coded with three figures.
In the sensory analysis, the evaluations were made using 40 testers, as indicated
by Stone and Sidel [9]. Prior to sensory testing, each tester signed the free and closed
consent term. Because it was a work involving human beings, it had to be submitted
to the evaluation of the Research Ethics Committee of the University of Taubaté,
according to protocol nº 603/11. Along with the bread loaves samples, each taster
received a score sheet containing 9 faces corresponding, respectively to 9 (extremely
enjoyed), 8 (liked very much), 7 (liked moderately), 6 (liked slightly), 5 (displeased
slightly), 3 (displeased moderately), 2 (disliked a lot) and 1 (disgusted extremely).
For the intent to buy the test, it contained 5 faces, corresponding to 5 (would certainly
buy), 4 (possibly buy), 3 (maybe buy/not buy), 2 (possibly not buy), and 1 (certainly
would not buy).
In order to verify significant differences between the different product formu-
lations, Variance Analysis (ANOVA) was necessary, in relation to the attributes
analyzed.

3 Results and Discussion

3.1 Characterization of Products

Table 1 presents the results concerning the characterization of banana flour, where
each value represents the average of the three, and the values in parentheses, the
respective standard deviations.
In relation to the results expressed in Table 1, it can be verified that the batches
of the products that were used in the elaboration of the different bread formula-
tions presented centesimal composition characteristics of each product, according to
results obtained by Neto et al. [10].

Table 1 Physical and chemical properties of banana flour

Analysis (%) Content (standard deviation)
Moisture 6.63 (0.03)
Ash 0.048 (0.007)
Lipid 0.0071 (0.0002)
Protein 4.534 (0.080)
Feasibility in Using Banana Flour … 325

Table 2 Sensory analysis of the bread enriched with banana flour

Attribute F.1 F.2 F.3 F.4
Texture 6.90a 5.35b 5.13b 4.45b
Flavor 7.03a 5.16b 5.16b 3.81c
Appearance 7.19a 5.55b 5.23b 5.35c
Global impression 6.94a 5.03b 5.09b 4.13c
a and b correspond to statistical parameters; the same item, along the same line, means that there is

no significant difference with respect to a certain sensory parameter

3.2 Sensory Analysis

The results of the sensory analysis, referring to the four different formulations of
bread enriched with banana flour, are shown in Table 2.
The results presented in Table 2 showed that the formulations, in general, presented
a great acceptance, except that elaborated from a mixture composed of 70% of banana
flour and 30% of wheat flour, which presented significant difference before the others,
with respect to all attributes analyzed.
From the results of Table 2, it can be observed that banana flour-enriched bread
presented satisfactory acceptability since, except for the F4 formulation, the others
presented scores above 5 for all attributes, indicating good acceptance. Still, in rela-
tion to the results of Table 2, it can be observed that the different formulations of
bread form presented significant differences in relation to all the attributes.
For all the sensorial attributes analyzed, it was observed that the greatest rejection,
by the tasters, was for the formulation 4, which proves the impossibility of replacing
the wheat flour for banana flour by more than 50%, in the production of bread.
The satisfactory acceptability of formulations 2 and 3 demonstrate viability in the
nutritional enrichment of loaves, partially replacing wheat flour with banana flour to
a certain extent (50% banana flour and 50% wheat flour), although preference, by
the tasters, was by formulation 1, that elaborated with 100% of wheat flour.

4 Conclusion

From the results presented in the previous item, it can be concluded that, except for the
formulation F.4, the others presented good acceptance, being that they differed signif-
icantly from the one elaborated with 100% of wheat flour, in relation to all the sensory
attributes. In this way, it can be affirmed that the addition of the banana flour inter-
feres in the organoleptic qualities of the bread of form. The formulations, in general,
presented satisfactory acceptability, except that elaborated from a mixture composed
of 70% of banana flour, which proves the viability in the nutritional enrichment of
bread loaves with banana flour to some extent.
326 N. R. S. de Carvalho et al.

References

1. Castelo-Branco, V.N., Guimarães, J.N., Souza, L., Guedes, M.R., Borges, A.M., Pereira, J.,
Lucena, E.M.P.: Green banana flour characterization. Food Sci. Technol. Cambridge 2, 333–339
(2009)
2. Melo, M. T. P., Rocha Júnior, V. R., Caldeira, L. A., Pimentel, P. R. S., Reis, S. T., Jesus, D.
L. S.: Cheese and milk quality of F1 Holstein x Zebu cows fed different levels of banana peel.
Acta Scientiarum. Anim. Sci. 32(2), 181187 (2017)
3. Agama-Acevedo, E., Islas-Hernández, J.J., Pacheco-Vargas, G., Osorio-Díaz, P., Bello-Pérez,
L.A.: Starch digestibility and glycemic index of cookies partially substituted with unripe banana
flour. LWT—Food Sci. Technol. Amsterdam 46, 177–182 (2012)
4. Almeida, E. L.: Efeito da adição de fibra alimentar sobre a qualidade de pão préassado
congelado, 328f. Dissertação (Mestre em Tecnologia de Alimentos). Universidade Estadual
de Campinas (UNICAMP), Campinas (2006)
5. Tavares, T.S., Bastos, S.C., Sousa, M.E., Pimenta, G., Pinheiro, A.C.G., Fabrício, L.F.F., Leal,
L.S.: Perfil sensorial de pão de forma enriquecido com farinha de Matrinxã (brycon lundii).
Anais do XIX Congresso de Pós-Graduação da UFLA. UFLA, Lavras (2010)
6. Castelo-Branco, V.N., Guimarães, J.N., Souza, L., Guedes, M.R., Silva, P.M., Ferrão, L.L.,
Miyahira, R.F., Guimarães, R.R., Freitas, S.M.L., Reis, M.C., Zago, L.: Uso da farinha de
polpa e de casca de banana verde (Musa balbisiana) como ingrediente para a elaboração de
massa tipo talharim. Braz. J. Food Technol. 20, e2016119 (2017)
7. AOAC.: Official Methods of Analysis of the of AOAC International (1ª edn.) 17th edn.
Gaythersburg, MD. Adolfo Lutz, p. 1020 (2000/2008)
8. Bligh, E.G., Dyer, W.J.: A rapid method of total lipid extraction and purification. Canad. J.
Biochem. Physiol. 37, 911–917 (1959)
9. Stone, H., Sidel, J. L.: Sensory evaluation practices, p. 308. Elsevier Academic Press, San
Diego (1993)
10. Neto, J. M. M., Cirne, L. M. E. R., Pedroza, J. P. P., Silva, M. G.: Componentes químicos
da farinha de banana (Musa ssp.) obtida por meio de secagem natural. Revista Brasileira de
Engenharia Agrícola e Ambiental 2(3), 316–318 (1998)
Spectral Behavior of Maize, Rice, Soy,
and Oat Crops Using Multi-spectral
Images from Sentinel-2

Lady Angulo and Sergio Pamboukian

Abstract Monitoring the status of crops at regional, local and smallholder farming
levels is essential in the search for sustainable agricultural management. One of the
methodologies applied for this purpose is precision agriculture, which, together with
the technological tools and equipment (Global Navigation Satellite Systems, Remote
Sensing, Geographic Information Systems and drones), form an integrated agricul-
tural system based on information of crop production and evolution. Measures of the
interaction response from electromagnetic radiation (ER) with soil and vegetation
are a basis for the development of agricultural yields and harvest index predictors.
In this research, spectral signatures and the Normalized Difference Vegetation Index
(NDVI) were analyzed for maize, soybean, rice and oat crops in a growing area of
320 ha, located in the city of Tremembé (São Paulo-Brazil) for the summer harvest
period from November to March of the crop cycles of 2017/2018 and 2018/2019,
through the use of multi-spectral images of Sentinel-2 satellite platform (and one
image from Landsat-8 to cover the missing image for December 2017). The results
showed a higher spectral response (especially in the RedEdge and Near-InfraRed)
for soybean crops, followed by rice and corn around the 60–90 days of sowing, and
a decrease with the approach to the harvest. In addition, were found higher NDVI
values to the crop cycle 2017/2018 (0.962), around 55–80 days of sowing for the
soybean and rice crops. However, the development of crop cycle 2018/2019 shows
more homogeneous spectral results.

Keywords NDVI · Precision agriculture · Spectral signatures · Sentinel-2

L. Angulo (B) · S. Pamboukian

Center of Radio Astronomy and Astrophysics at Mackenzie (CRAAM), Mackenzie Presbyterian
University, São Paulo, SP 01302-907, Brazil
e-mail: [email protected]
S. Pamboukian
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 327
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_32
328 L. Angulo and S. Pamboukian

1 Introduction

Digital geographic information is now considered a basis for acquiring essential envi-
ronmental and crop monitoring information, adding the availability of acquiring data
from different times and places. The increase in the availability of digital geographic
information data sources, as well as satellite platforms with a free and unrestricted
imagery access policy and improvements in spatial, spectral, and radiometric reso-
lutions, have made the multi-spectral imagery very useful for detecting crop and
vegetation changes [1]. One of the methods for obtaining information from different
objects through instruments that are not in physical contact with those objects is
remote sensing, broadly known as the science that studied the responses of ER
interactions with terrestrial materials by developing ground surface imaging [2]. To
increase Earth monitoring possibilities, space agencies have created initiatives such
as the Copernicus Sentinel-2 (S2), and Landsat-8 satellite missions, to guarantee a
permanent supply of satellite images of mid-spatial resolution (with 10 and 30 m,
respectively) and a great potential for generating vegetation coverage information.
The European’s Space Agency (ESA) Copernicus mission Sentinel-2 is a polar-
orbiting, multi-spectral mid-resolution imaging (10 m pixel size and 10–5 days of
temporal resolution) with twin polar-orbiting satellites, Sentinel-2A (launched 23
June 2015) and 2B (launched 7 March 2017) [3]. In addition, other of the satellite
programs widely used for Earth observation applications is the joint NASA/U.S.
Geological Survey (USGS) Landsat series; this program is the longest-operating
Earth observation satellite enterprise providing imagery of the Earth from space.
The USGS opened the access of the Landsat satellite data since 2008, with the last
satellite launched in February 2013 (Landsat-8) [4].
Remote sensing technology applications on satellite platforms, aerial unmanned
vehicles, and portable radiometers together with the use of Global Navigation Satel-
lite System (GNSS), Geographic Information System (GIS), and farm equipment
form an integrated system based on information of crop production and evolution
known as precision agriculture. This system has as a principal goal the practice of
sustainable agriculture and seeks to improve the management of agricultural inputs
(fertilizers, herbicides, seeds, and fuel) by increasing crop productivity and reducing
adverse impacts on the environment. Remote sensors capture the interaction response
of ER with soil and vegetation as a measure of the ratio of reflected radiance flux to
incident radiance flux at a specific wavelength (spectral reflectance of the surface)
[5]. Analyzing the spectral signature response of different crop types and vegetation
index values as the NDVI during the development of the crop can provide technical
support in farming practices such as fertilizer application, water management, and a
tool to estimates the crop health and growth status [3].
In this respect, we aim to obtain the spectral signatures (spectral reflectance values)
for maize, soybean, rice and oat crops in a growing area of 320 ha, located in the city
of Tremembé (São Paulo-Brazil) for summer sowing cycles (November to March)
2017/2018 and 2018/2019 (oat culture is studied from April to June 2018), with the
purpose of analyzing the signature behavior in the different phases of the crops. In
Spectral Behavior of Maize … 329

addition, we seek to employ the NDVI index to determine the behavior of the crops
in each cycle, showing the advantage, useful, and utility of Sentinel-2 images in
precision agriculture.

2 Materials and Methods

2.1 Data

There are some differences in the image characteristics from satellite platforms, given
in the spatial, spectral, radiometric, and temporal resolutions. The Multi-Spectral
Imager (MSI) sensor of Sentinel-2 and the Operational Land Imager (OLI) sensor
of Landsat-8 are similar but have a different spectral and spatial resolution, and their
data have different levels of radiometric and atmospheric correction.
Landsat 8 bands have wavelengths analogous to Sentinel-2 bands [6], as Table 1
shows.
The MSI Sentinel-2 sensor covers 13 spectral bands (443–2190 nm), but only
10 are useful for agricultural monitoring, involving the regions of visible (B2–B4),
vegetation redEdge (B5–B7 and B8A), NIR (Near infra-red B8) and the SWIR (Short
wave infra-red B11 and B12). About the OLI sensor of Landsat-8, it covers 8 spectral
bands (443–2200 nm), capturing also image data in the regions of visible (B2–B4),
NIR (B5), SWIR 1 (B6) and SWIR 2 (B7), adding the panchromatic region band
(B8), being only 6 of this bands useful in vegetation studies.

Table 1 The Landsat-8 OLI and Sentinel-2 MSI equivalent spectral bands considered in this study
Landsat-8 OLI Sentinel-2 MSI
a B2 (Blue)-30-(452–512) B2-10-(458–523)
B3 (Green)-30-(533–590) B3-10-(543–578)
B4 (Red)-30-(636–673) B4-10-(650–680)
B5 (RedEdge 1)-20-(689.1–719.1)
B6 (RedEdge 2)-20-(725.5–755.5)
B7 (RedEdge 3)-20-(762.8–802.8)
B5 (NIR)-30-(851–879) B8-10-(785–900)
B8A (RedEdge 4)-20-(843.7–885.7)
B9 (WaterVapor)-60-(843.7–885.7)
B6 (SWIR1)-30-(1566–1651) B11-20-(1565–1655)
B7 (SWIR2)-30-(2107–2294) B12-20-(2100–2280)
B8 (PAN)-15-(500–680)
B9 (CIRRUS)-30-(1360–1390) B10-60-(1342,5–1404,5)
a Band number (name)-spatial resolution (m)-spectral band range (nm)
330 L. Angulo and S. Pamboukian

Fig. 1 Spectral response for the approximately equivalent Sentinel-2 MSI (solid lines) and Landsat-
8 OLI (dashed lines) [6]

The revisit time of Landsat-8 satellite and Sentinel-2 platforms are 16 and 5 days,
respectively, availing observations spread over time; different Sentinel-2 image data
were used to this study, involving one image per month to the summer sowing period
(November to March) for the crop cycles of 2017/2018 and 2018/2019, which allowed
to include five images for each of the cycles under study, except for the Sentinel-2
image of December 2017 that showed dense cloud cover over the area of interest, and
was replaced for a Landsat-8 image for this research, given the equivalent spectral
band characteristics (see Fig. 1) and the available tools for their corrections.

2.2 Study Area

The study area is a plantation (maize, rice, soybean and oat crops) of approximately
320 ha located in the city of Tremembé (São Paulo-Brazil) (see Fig. 2) at the Universal
Transversal Mercator (UTM) coordinates of 438,814.90 m E, 7,461,118.52 m N for
the reference system SIRGAS 2000 zone 23 S. Local measures in the nearest weather
station identify as the driest month to July, with precipitation about 21 mm, in contrast
to the highest precipitation in January with an average precipitation of 238 mm [7].

2.3 Image Processing

Sentinel-2 (Level-1C) products were downloaded from OpenHub, an ESA’s web plat-
form for the publicly available download of Copernicus mission products (available
at https://scihub.copernicus.eu/), those are delivered directly at top of atmosphere
Spectral Behavior of Maize … 331

Fig. 2 The study area (delimiting boundary in the orange right region) in the city of Tremembé
(São Paulo)

reflectance values (feature a TOA—top of atmosphere geometric and radiometric

correction), and using Qgis 3.2 version software semiautomatic classification (SCP)
plugin for atmospheric correction to obtain the reflectance values at the base of atmo-
sphere (BOA) applying the dark object subtraction (DOS1) correction method. One
of the sources of error that can completely alter the image radiometrically is the
atmosphere, it affects the radiance measured in the image in two ways: as a reflector,
by adding extra radiance to the signal and as an absorber, attenuating the intensity of
energy illuminating the target on the surface. DOS1 atmospheric correction method
uses (1), (2) and (3) for an image-based technique which uses a metadata file that
contains the information required to transfer the digital counts (DC) of the image
into physical reflectance values [8].

0.01 ESUNλ ∗ COSθ s ∗ T z + Edown ∗ T v

LDO1% = (1)
π ∗ d2

L p = Lmin − LDO1% (2)

ρ = [π ∗ (Lλ − L p) ∗ d 2 ]/ESUNλ ∗ cosθ s (3)

where: Lλ = Spectral radiation in the sensor (Radiance in the satellite), Lp = Path

radiance, Lmin = Radiance obtained with the digital count value (DNmin), LDO1%
332 L. Angulo and S. Pamboukian

= radiance of Dark Object, assumed to have a reflectance value of 0.01, d = Earth-

Sun Distance in Astronomical Units (AU), Tv = Transmittance of the atmosphere
in the direction of vision, Tz = Atmospheric transmittance in the lighting direction,
ESUNλ = Mean solar exo-atmospheric irradiances (W/(m2 µm)), θ s = Solar zenithal
angle in degrees, Edown = Diffuse irradiance descending, ρ = surface reflectance.
Several DOS techniques consider different values about Tv, Tz, and Edown. The
DOS1 technique gives for Tv and Tz the value of 1, and for Edown the value of 0 [9].
After applying the atmospheric correction, the digital counts of Sentinel-2 images
resulting are measures of land surface reflectance.
In the case of the Landsat-8 image, the Earth Explorer website provides free
access to level L2A processed products, which have geometric, radiometric, and
atmospheric correction by applying the Landsat Surface Reflectance Code (LaSRC)
physical correction method. This method uses OLI sensor information for capture
values and optical properties of the atmosphere to correct the images and present
digital image counts in reflectance values of the surface (BOA) [10].
The physical reflectance values for each of the cultures were obtained after the
atmospheric correction of the images, and the graphics of the spectral signatures
were made using phyton 3. Considering the date of every image and the records
of field books obtained from the farmers (separate zoning areas were considered to
identify crop differences), different spectral firms were obtained, corresponding to
the development of the crops.

2.4 Normalized Difference Vegetation Index (NDVI)

NDVI is the most commonly used vegetation index when analyzing vegetation vigor,
it is one of the oldest, best known and most frequently used vegetation index [11].
The NDVI approach is based on the fact that healthy vegetation has low reflectance
in the visible portion of the spectrum due to chlorophyll and absorption of other
pigments, and has high reflectance in NIR due to internal reflectance by the green
leaf tissue, the values of NDVI calculations for a given pixel always result in a number
ranging from −1 to +1, where, only positive values correspond to vegetation areas,
values close to zero do not mean vegetation, and negative values belong to clouds,
snow, water, uncovered soil areas and rocks. In the formula for NDVI (4), NIR is the
near-infrared range band and R is the red range band [11].

NDVI = NIR − R/NIR + R (4)

Spectral Behavior of Maize … 333

3 Results and Discussion

3.1 Spectral Signatures Analysis

Comparative experimental analysis of the spectral behavior graphs (spectral signa-

tures) of the corn, rice, and soybean crops for the 2017/2018 and 2018/2019 sowing
cycles present the following observation results:
• In general, there is a similarity to the specific assumptions of all the cultures studied
(oat, rice, soybeans, maize), with marked differences in the spectral response by
presenting inflection points of values at bands 3 (0.560 µm), 4 (0.665 µm), 5
(0.705 µm), 7 (0.783 µm), 8 (0.842 µm) and 8A (0.865 µm) of Sentinel-2.
• In crop zone, 10b (rice), Fig. 3a, b, showed a higher reflectance response for
the spectral signatures in the 2017/2018 cycle, with a peak value (circled in red)
around of 48% (0.865 µm) for 60 days of sowing in equivalent to the 32% peak
value (circled in red) for the 2018/2019 cycle for 85 days of sowing.
• Similarly, in crop zone 10d (rice) (Fig. 3c, d), the maximum peak of reflectance
value is 57% (pointed out in red), given in band 8A (0.865 µm) at 50 days of
sowing, for crop cycle 2018/2019, different from the spectral response in the
2017/2018 of 47% at 45 days of sowing (circled in red), which gives the idea that
there is a better development of this culture in the second cycle, being that the
common behavior in all cultivation areas.

Fig. 3 Spectral behavior results of cycles 2017/2018 (left) and 2018/2019 (right) for growing zones
10b-rice (a and b), 10d-rice (c and d), 6-maize (e and f)
334 L. Angulo and S. Pamboukian

Wavelength (µm)

Fig. 4 Spectral behavior of corn, rice, soybean and oat crops for around 60 days of sowing

• For the sowing area, 10a (maize), Fig. 3e, f shows a spectral response peak in the
band 8A in the 2017/2018 cycle at 40% with 46 days of sowing and a spectral
peak in the same band of 60% for 80 days of sowing in the 2018/2019 cycle. It is
also noted that the duration of the first cycle was shorter (77 days) than the 95 days
of the second cycle, assuming a problem (a nutrient deficiency, for example) in
this parcel for the 2017/2018 cycle.
• The soybean crop has higher reflectance values compared to other crops (rice, corn,
oats) (Fig. 4), with the highest response values in the NIR region and the lowest
in the green region, which would be related to the good percentage absorption for
the soybean of nitrogen in the visible range (focused on green 0.560 µm).
According to [12] the values of reflectance in the visible range (0.490–0.665 µm)
are related to the content of chlorophyll (carotenes and anthocyanins) and nitrogen
in crops. Thus, it can be identified, as in [13], that nitrogen content (also associated
with better crop yields) is related to lower visible reflectance and higher redEdge
reflectance (0.705–0.783 µm). As is shown in [1], the spectrum intervals where
peaks occur may have the potential to generate information related to some plant
characteristics. In the case of nitrogen (N) content in leaves, it presents a highly
significant coefficient with reflectance values in the wavelength range of 450 and
780 nm. This may be related to the higher reflectance values found in soybean crops,
given the increased consumption of nitrogen fertilizers over the period of the sowing.

3.2 NDVI Analysis

The calculation of the normalized vegetation index (NDVI) revealed the relationship
of the spectral response of the crops with the NDVI values, that is an index that
measures the vigor of the crops, begins to decrease with get close to the harvest
season, given the variation of the pigment in the leaves that tend to yellow in the
crops studied.
Spectral Behavior of Maize … 335

13/11/17 17/12/17 17/01/18 11/02/18

13/11/18 13/12/18 17/01/19 01/02/19

Fig. 5 NDVI variations obtained for the 2017/2018 and 2018/2019 sowing cycles (November–
March)

The maximum value obtained for NDVI as seen in Fig. 5 was in the 2017/2018
cycle (0.962) for the December 2017 image, with sowing zones 10d (rice 79 days)
and 2 (soybeans 54 days) showing the largest area with the maximum value.
In the 2018/2019 cycle, the maximum value obtained for NDVI was 0.924 for
the January 2019 image, with the sowing zones 2 (soybeans 55 days), 3 (soybeans
65 days), and 11 (Soybeans 80 days) showing the largest area with the maximum
value. For the growing zones 6, 8, and 10 a (sown with maize) also showed a better
response in the 2018/2019 cycle with larger areas of 0.86 NDVI values.

4 Conclusion

There is a marked tendency in the curvature of a signature; generally, the reflectance

curves showed their highest responses around 60–90 days of sowing. The spectral
responses decrease with the approach to harvest, which is given by nutrient loss and
leaf chlorosis.
Thus it was also found that according to what was manifested in [12] the spectral
responses decrease with the approach to harvest, which is given by the senescent dry
vegetation, nutrient loss, and leaf chlorosis (yellowing).
In addition, it was observed in most cases that the soybean crop has higher
reflectance values compared to other crops (corn, rice, oats), this can be related
to the nitrogen content in the sowing and during the development of the crop. It
336 L. Angulo and S. Pamboukian

can be seen that apparently there is better crop development in the second cycle
(2018/2019).
In general, spectral signatures can be influenced by various external factors, such
as sowing time, nutrient content, soil moisture, leaf density, and others. Therefore,
spectral response analysis is considered as a baseline for area discrimination of
interest in field campaigns. It is also evidenced that vegetation indices can be used
as a tool in the detection of possible affected areas.
Results can only be verified by joining field analyzes of crop nutrient content
(nitrogen, potassium, phosphorus, and others), which are usually obtained using
spectroradiometers or chemical methods by taking soil and leaf samples.

References

1. Martinez, L.J.: Relationship between crop nutritional status, spectral measurements and
Sentinel 2 images. Agron. C. 35(2), 205–215 (2017). https://doi.org/10.15446/gron.colomb.
v35n2.62875
2. Tanriverdi, C.: A Review of remote sensing and vegetation indices in precision farming. KSU,
J. Sci. Eng. 9(1), 69–76 (2006)
3. Pasqualotto, N., Delegido, J., Van Wittenberghe, S., Rinaldi, M., Moreno, J: Multi-crop green
LAI estimation with a new simple sentinel-2 LAI index (SeLI). Sensors (Basel, Switzerland)
19(4), 904 (2019). https://doi.org/10.3390/s19040904
4. Costa, S., Santos, V., Melo, D., Santos, P.: Evaluation of Landsat 8 and Sentinel-2A data on the
correlation between geological mapping and NDVI, pp. 1–4 (2017). https://doi.org/10.1109/
grss-chile.2017.7996006
5. Nicodemus, F., Richmond, J., Hsia, J.: Geometrical Considerations and Nomenclature for
Reflectance. National Bureau of Standards, US Department of Commerce: Washington, DC,
USA, (1977). URL: http://physics.nist.gov/Divisions/Div844/facilities/specphoto/pdf/geoCon
sid.pdf
6. Hankui, K., David, P., Lin Yan, P., Zhongbin Li, Haiyan Huang, Vermote, E., Skakun, S., Roger,
J.: Characterization of Sentinel-2A and Landsat-8 top of atmosphere, surface, and nadir BRDF
adjusted reflectance and NDVI differences. Remote Sensing Environ. 215, 482–494 (2018).
https://doi.org/10.1016/j.rse.2018.04.031
7. Matarazzo, A.: The Tremembé Green Areas, p. 3. São Paulo, Economic Brazil (2012)
8. Moran, M., Jackson, R., Slater, P., Teillet, P.: Evaluation of simplified procedures for retrieval
of land surface reflectance factors from satellite sensor output. Remote Sensing Environ. 41,
169–184 (1992)
9. Sobrino, J., Jiménez-Muñoz, J., Paolini, L.: Land surface temperature retrieval from LANDSAT
TM 5. Remote Sensing Environ. (Elsevier) 90, 434–440 (2004)
10. U. S. Geological Survey: Landsat 8 Surface Reflectance Code (LaSRC) Product Guide (2009).
URL: https://prd-wret.s3-us-west-2.amazonaws.com/assets/palladium/production/atoms/files/
LSDS-1368_L8_Surface_Reflectance_Code_LASRC_Product_Guide-v2.0.pdf
11. Pinter, P., Hatfield, J., Schepers, J., Barnes, M., Moran, C., Daughtry, et al.: Remote sensing
for crop management. Photogramm. Eng. Remote Sensing 69, 647e–664 (2003)
12. Blackburn, G.: Hyperspectral remote sensing of plant pigments. J. Exp. Bot. 58, 855e–867
(2007). URL: http://www.scielo.br/scielo.php?script=sci_arttext&pid=S1516-359820090009
00001
13. Schlemmer, M., Francis, D., Shanahan, J., Schepers, J.: Remotely measuring chlorophyll
content in corn leaves with differing nitrogen levels and relative water content. Agron. J.
97(1), 106–112 (2005)
Emerging Trends in Humans and Urban
Sciences
Production Process of Bioethanol Fuel
Using Supported Saccharomyces
Cerevisiae Cells

L. C. Fardelone , T. S. Bella de Jesus , G. P. Valença , J. R. Nunhez ,

J. A. R. Rodrigues , and P. J. S. Moran

Abstract A new immobilization process of Saccharomyces cerevisiae cells with

calcium alginate and coating with solubilized chitosan in citric acid solution was
developed. The biocatalyst is used in fermentation processes to produce bioethanol
fuel through the use of carbon sources such as glucose, sugarcane juice, and corn
hydrolysate, which provided very good yields of productivity (71.3–88.7%) and
shows excellent biocatalyst stability during all production processes.

Keywords Bioethanol fuel · Calcium alginate beads · Fermentation process

Author Contributions: Fardelone, L. C. and Bella de Jesus, T. S. contributed equally

L. C. Fardelone (B) · T. S. B. de Jesus · G. P. Valença · J. R. Nunhez

School of Chemical Engineering, University of Campinas - Unicamp, Campinas, Brazil
e-mail: [email protected]
T. S. B. de Jesus
e-mail: [email protected]
G. P. Valença
e-mail: [email protected]
J. R. Nunhez
e-mail: [email protected]
J. A. R. Rodrigues · P. J. S. Moran
Chemistry Institute, University of Campinas - Unicamp, Campinas, Brazil
e-mail: [email protected]
P. J. S. Moran
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 339
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_33
340 L. C. Fardelone et al.

1 Introduction

The research for more efficient biofuel processes is strategic for Brazil, due to climate
change, world economic policies and the growing demand for sustainability regarding
the issue of energy generation [1–6]. Thus, bioethanol fuel has gained worldwide
prominence because it is biodegradable, has clean combustion due to the presence of
oxygen in its composition, and reduces the emission of nitrogen oxides and particulate
matter, as well as reducing CO2 and SO2 emissions into the atmosphere [5].
The process of alcoholic fermentation using sugarcane juice S. cerevisiae cells is
a process well established in the Brazilian mills [1, 4], however, the search for cost
reduction, unit operations and process time, as well as productivity increase is a big
challenge, which motivates us in this development of bioethanol fuel production.
Also, the production of bioethanol fuel from corn has numerous advantages, such as
logistics and storage, shorter plant cultivation cycle time and depending on the variety
3–4 annual harvests, can also be obtained during the sugar cane off-season, and if
planted in the same area can bring improvements to the soil, decreased fertilization
and correction of it [7, 8]. Corn also has lower production costs and its planting is
established in different regions of the country [8].
Our research group has been working on fermentative processes for bioethanol
fuel production using cellular supports such as calcium alginate [9], Montmoril-
lonite K-10 [10], and other clays. The literature also describes supported processes
that can promote longer process time and consequently higher productivity, as
they simplify product separation and purification, and biocatalysts are reused in
consecutive production recycles [9–16].
However, the biggest problem is still the stabilization of the biocatalyst-support
and its reuse, as well as the costs, which hinder its implementation in certain industrial
processes [17, 18]. The use of sodium alginate is extremely important because of its
compatibility and non-toxicity [13–21], but this type of matrix support is generally
used in only one process cycle, as during this process the spheres rupture due to
calcium leaching, agitation and the formation of gases that cause the deterioration
of the support, thus requiring further technical development.
We are developing a new cell support processes, to overcome the alginate bead
disruption problem, that provided greater mechanical resistance strength and non-
leaching of calcium when used citric acid instead of acetic acid to recover the spheres
of alginates. The bioethanol fuel production process used glucose in a batch reactor
for 24 h with yields of 89–95% [18].
In this work, we report the bioethanol fuel production using this cell support in
a more complex media as sugarcane juice and corn hydrolysate as carbon sources,
which provide us with greater robustness to the process enabling technology to be
transferred to industry.
Production Process of Bioethanol Fuel Using Supported … 341

2 Methodology

2.1 General Process

Fermentations were conducted in the fed-batch regime, replacing the regular time
intervals of two-thirds of the reactor volume. This avoids product inhibition and
allows the beads to be kept in action for long periods of time. Regardless of the
medium used, 3 g/L CaCl2 was added, as it is used for avoiding the leaching Ca2+
ions, responsible for the rigidity of the alginate matrix. The fermentations were
conducted using 75 g of alginate beads and 0.150 L of media (which was replaced
in periods of 24 h) in 0.5 L lightly stirred bioreactor for 168 h.
Ethanol and glucose concentrations were determined by HPLC with refractive
index detector, coupled to an Aminex HPX-87H column (300 mm × 7.8 mm, Bio-
Rad) and using 5 mmol/l mobile phase H2 SO4 , with flow 0.6 mL/min at 50 °C.
Quantifications of products and sugars were performed from calibration curves using
an external standard.

Microorganism S. cerevisiae JAY270 is being used for this process, which has
high-temperature resistance and high substrate concentration, pH variation, oxidative
stress, high cell mass production, and bioethanol fuel [22]. This strain of S. cerevisiae
JAY270 is found in our laboratory and is stored in the form of BCM (Master Cell
Bank) and BCT (Master Cell Bank) and WCB (Working Cell Bank), respectively
and were prepared in accordance with international precepts for good production and
laboratory practices [23–25].

Preparation of culture media For cell expansion (inoculums and pre-inoculums)

of S. cerevisiae yeast the culture medium with industrial components previously
established with a composition of 3.0 g/L malt extract, 3.0 g/L of yeast extract,
5.0 g/L peptone and 10 g/L glucose at pH 6.0. To obtain a high cell density (OD)
baffled flask is being used, which promotes a better shaker mix for 18 h at 120 rpm.

Cell Immobilization After cell mass growth, the cells are centrifuged at 1844 g for
20 min, and under aseptic conditions, the obtained cells undergo immobilization in
calcium alginate, followed by chitosan coating [9, 18, 26].

The cell immobilization technique is porous matrix entrapment using a 3% sodium

alginate suspension and S. cerevisiae cells. The mixture is then dripped and remains
for one hour in 0.3 M calcium chloride solution to obtain spheres with a diameter of
approximately 3–4 mm. To have a better calcium distribution in the cell support
matrix gel, the beads are kept in 0.3 M CaCl2 solution for one hour, followed
by coating with 0.25% chitosan solubilized in 5% citric acid for 30 min, with
recalcification for another hour in 0.3 M CaCl2 solution.
The coating of spheres containing S. cerevisiae in calcium alginate is performed
with the use of citric acid, an unprecedented technique developed in our laboratory,
in such a way as to promote greater mechanical resistance against the fermentation
342 L. C. Fardelone et al.

process time, as well as to maintain homogeneity the possibility of calcium leaching

present in the polymeric network of the spheres. The use of citric acid used in the
coating also decreases the lag phase from 24 to 8 h of bioethanol fuel production due
to its toxicity to S. cerevisiae cells.

Analytical Methodology The final concentration of ethanol and residual glucose is

determined by High-Performance Liquid Chromatography (HPLC) with refractive
index detector coupled to an Aminex HPX-87H column (300 mm × 7.8 mm, Bio-
Rad) and uses 5 mmol/L mobile phase H2 SO4 , with 0.6 mL/min flow at 50 °C.
Product and residual carbohydrate quantifications are performed from calibration
curves using external standards for ethanol and sucrose, glucose, fructose, and other
carbohydrates.

3 Results and Discussion

Fermented batch processes were carried out for bioethanol fuel production using S.
cerevisiae JAY270 cells immobilized in calcium alginate and coated with chitosan.
This process requires the addition of a small amount of calcium in the medium
(0.03 g/1.0 g of beads) to recalcify the beads, allowing for greater durability and effi-
ciency of the bioethanol fuel production process as shown in Table 1. It is important
to emphasize that the final yield was defined through the ethanol produced in relation
to the maximum conversion of carbon source to ethanol.
Coating the spheres with chitosan using citric acid and adding calcium to the
fermentation broth provided a highly stable process with a small lag (8 h) phase
and high productivity, 81.0, 88.7, and 71.3% yield in bioethanol fuel for glucose,
sugar cane, and corn hydrolysate, respectively. It also promoted greater mechanical
resistance of the spheres, as no support rupture was observed, nor free S. cerevisiae
cells were detected in the fermentation broth. Thus, the developed methodology is
stable and robust for supported bioethanol fuel production processes and can be
extended to other processes, when compared to other processes that provide low
yields and beads disruptions [9, 27] (Fig. 1).

Table 1 Bioethanol fuel production using S. cerevisiae JAY270 cells immobilized in calcium
alginate and coated with chitosan and glucose, sugarcane and corn hydrolysate aqueous solutions
Process Consumption (g) Bioethanol Productivity Yield (%)
using Glucose Fructose Sucrose Maltose fuel (g) (g/(Lh))

Glucosea 185.64 – – – 56.91 1.21 60.0

Glucose 256.26 – – – 106.10 4.21 81.0
Sugarcane 71.34 20.94 11.06 – 48.50 1.92 88.7
Corn 78.29 – – 48.37 47.10 2.06 71.3
hydrolysate
a Process using acetic acid in chitosan coating of cell support up to 94 h of bioethanol fuel production

process, with 60% of broken spheres

Production Process of Bioethanol Fuel Using Supported … 343

Fig. 1 Carbohydrate mass profiles and bioethanol fuel production in batch fed reactors provided
with immobilized cells of S. cerevisiae in calcium alginate spheres coated with solubilized chitosan
in citric acid solution a glucose; b sugar cane juice; c corn hydrolysate
344 L. C. Fardelone et al.

In these fermentation processes, 75 g of beads were used, 150 ml of medium

containing ~200 g/L of carbon source (glucose, sugar cane or corn hydrolysate),
2 g/L of ammonium phosphate and 8 g of CaCl2 . During the fermentation process,
2/3 of the medium volume was exchanged by a new medium every 8 h to maintain
bioethanol fuel production by a total period of 168 h. The medium exchange was
made by simple suction of the fermented liquid and replaced by the same volume of
new medium.
The biocatalyst supported on calcium alginate beads coated with citric acid-
soluble chitosan after 21 cycles in a total period of 168 h of fermentation shows
any disruption of the cell support spheres as shown in Fig. 2c in contrast to Fig. 2a,
where the calcium alginate beads were coated with acetic acid solubilized chitosan.
During the production process, it is also possible to visually monitor the produc-
tion of bioethanol fuel by observing the spheres of biocatalyst, in the process the
spheres of biocatalyst are floating due to the presence of CO2 inside and when all
carbon source is consumed the spheres decant for the absence of CO2 within the
spheres, Fig. 3.

Fig. 2 a Biocatalyst supported on calcium alginate beads coated with acetic acid solubilized
chitosan after 94 h of fermentation process; b Biocatalyst supported on calcium alginate beads
coated with citric acid-soluble chitosan before use in fermentation process; c Biocatalyst supported
on calcium alginate beads coated with citric acid-soluble chitosan after 21 cycles in a total period
of time of 168 h of fermentation; d typical process fermentation
Production Process of Bioethanol Fuel Using Supported … 345

Fig. 3 a Biocatalyst in bioethanol fuel production process, with carbon source present in a
bioreactor; b Biocatalyst at the end of the presence of carbon source

The conversion of the used carbohydrates to combustible bioethanol was achieved

excellently (>88%) after 21 cycles, demonstrating that the process developed using
S. cerevisiae cells supported on solubilized chitosan-coated calcium alginate beads in
citric acid is extremely efficient and robust, since the supports are intact and reusable,
and fermentation processes can be used for extended periods.
Another extremely important fact is that we can use more complex substrates,
such as sugarcane juice, corn hydrolysate, among others, with great possibilities
for technology transfer for application of this process to an industrial scale. The
production of bioethanol fuel with corn hydrolysate requires gelatinization and starch
hydrolysis steps, followed by fermentation [28, 29]. Generally, industrial processes
with corn hydrolysate take a very long time, over 48 h, and we intend to reduce this
time and increase productivity. This same technology can be extended to support other
microorganisms (engineered or not), enzymes, and various production processes.

4 Conclusions

The developed immobilization process gave excellent mechanical resistance of the

cell support, is possibly more than 21 consecutive reuse cycles in longer processes,
providing high productivity of bioethanol fuel using glucose, sugarcane juice, and
corn hydrolysate as carbon sources. Also, the easy of the beads separation from the
fermentative medium shows the possibility of process optimization to a larger scale.
Scaling up study is being developed and the results will be reported in the future.

Acknowledgements The authors thank São Paulo State Research Support Foundation
(FAPESP, 2016/12074-7), Higher Education Personnel Improvement Coordination (CAPES,
346 L. C. Fardelone et al.

88,882.329694/2019-01), and National Council for Scientific and Technological Development

(CNPq) for financial support.

References

1. Amorim, H.V., Lopes, M.L., Oliveira, J.V.C., Buckeridge, M.S., Goldman, G.H.: Scientific
challenges of bioethanol production in Brazil. Appl. Microbiol. Biotechnol. 92, 1267–1275
(2011)
2. Kang, A., Lee, T.S.: Converting sugars to biofuels: ethanol and beyond. Bioengineering 2,
184–203 (2015)
3. Breisha, G.Z.: Production of 16% ethanol from 35% of sucrose. Biomass Bioenerg. 34, 1243–
1249 (2010)
4. Miranda, M., Batistote, M., Cilli Jr., E.M., Ernandes, J.R.: Sucrose fermentation by Brazilian
ethanol production yeasts in media containing structurally complex nitrogen sources. J. Inst.
Brewing 115(3), 191–197 (2009)
5. Gonçalves, F.A., Ruiz, H.A., dos Santos, E.S., Teixeira, J.A., de Macedo, G.R.: Bioethanol
production by Saccharomyces cerevisiae, Pichia stipitis and Zymomonas mobilis from delig-
nified coconut fibrer mature and lignin extraction according to biorefinery concept. Renew.
Energy 94, 353–365 (2016)
6. Rosales-Calderon, R., Arantes, V.: A review on commercial–scale high–value products that
can be produced alongside cellulosic ethanol. Biotechnol. Biofuel 12, 240 (2019)
7. Mohanaty, S.K., Swain, M.R.: Bioethanol production from corn and wheat: food, fuel and
future (Chapter 3). In: Ramesh, R.C., Ramachandran, S. (eds.) Bioethanol Production from
Food Crops, pp. 45–59 (2019)
8. Marques, S.J.P., da Cunha, M.E.L.: Production of fuel alcohol using corn. Unopar Científica
Ciências Exatas e Tecnológicas 7, 45–51 (2008)
9. Duarte, J.C., Rodrigues, J.A.R., Moran, P.J.S., Valença, G.P., Nunhez, J.R.: Effect of immo-
bilized cells in calcium alginate beads in alcoholic fermentation. ABM Express 31, 1–8
(2013)
10. Wendhausen, R., Fregonesi, A., Moran, P.J.S., Joekes, I., Rodrigures, J.A.R., Tonella, E.,
Althoff, K.: Continuous fermentation of sugar cane syrup using immobilized yeast cells. J.
Biosci. Bioeng. 91, 48–52 (2000)
11. Karagoz, P., Ozkan, M.: Ethanol production from wheat straw by Saccharomyces cerevisiae
and Scheffersomyces stipitis co-culture in batch and continuous system. Biores. Technol. 158,
286–293 (2014)
12. Singh, A., Sharma, P., Saran, A.K., Singh, N., Bishnoi, N.R.: Comparative study on ethanol
production from pretreated sugarcane bagasse using immobilized Saccharomyces cerevisiae
on various matrices. Renew. Energy 50, 488–493 (2013)
13. Lee, H., Kim, S., Yoon, J., Kim, K.H., Seo, J., Park, Y.: Evolutionary engineering of Saccha-
romyces cerevisiae for efficient conversion of red algal biosugars to bioethanol. Biores. Technol.
191, 445–451 (2015)
14. Duarte, J., Lourenço, V., Ribeiro, B., Saagua, M.C., Pereira, J., Baeta-Hall, L.: Ethanol produc-
tion from different substrates by a flocculent Saccharomyces cerevisiae strain. Int. J. Chem.
Reactor Eng. 7, Article A58 (2009)
15. Sebayang, F., Bulan, R., Hartanto, A., Huda, A.: Enhancing the efficiency of ethanol production
from molasses using immobilized commercial Saccharomyces cerevisiae in two-layer alginate-
chitosan beads. IOP Conf. Ser. Earth Environ. Sci. 305, 012014 (2019)
16. Ivanova, V., Petrova, P., Hristov, J.: Application in the ethanol fermentation of immobilized
yeast cells in matrix of alginate/magnetic nanoparticles, on chitosan-magnetite microparticles
and cellulose-coated magnetic nanoparticles. Int. Rev. Chem. Eng. (I.RE.CH.E.), 3(2), 289–299
(2011)
Production Process of Bioethanol Fuel Using Supported … 347

17. Ghorbani, F., Younesi, H., Sari, A.E., Najafpour, G.: Cane molasses fermentation for contin-
uous ethanol production in an immobilized cells reactor by Saccharomyces cerevisiae. Renew.
Energy 36, 503–509 (2011)
18. Fardelone, L.C., Bella de Jesus, T.S., Oura, L.A.K., Valença, G.P., Nunhez, J.R., Rodrigues,
J.A.R., Moran, P.J.S.: Produção de bioetanol utilizando células de Saccharomyces cerevisiae
imobilizadas em esferas de alginato de cálcio recobertas com quitosana. Editora Atena,
Impactos das Tecnologias na Engenharia Química (2019)
19. Călinescu, I., Chipurici, P., Trifan, A., Bădoiu, C.: Immobilization of Saccharomyces cerevisiae
for the production of bioethanol. U.P.B. Scientific Bulletin, Series B 74, 33–40 (2012)
20. Najafpour, G., Younesi, H., Ismail, K.S.K.: Ethanol fermentation in an immobilized cell reactor
using Saccharomyces cerevisiae. Biores. Technol. 92, 251–260 (2004)
21. Kubota, N., Tatsumoto, N., Sano, T., Toya, K.A.: Simple preparation of half N-acetylated
chitosan highly soluble in water and aqueous organic solvent. Carbohyd. Res. 324(4), 268–274
(2000)
22. Argueso, J.L., Carazzolle, M.F., Mieczkowski, P.A., Duarte, F.M., Netto, O.V.C., Missawa,
S.K., Galzerani, F., Costa, G.G.L., Vidal, R.O., Noronha, M.F., Dominska, M., Andrietta,
M.G.S., Andrietta, S.R., Cunha, A.F., Gomes, L.H., Tavares, F.C.A., Alcarde, A.R., Dietrich,
F.S., McCusker, J.H., Pereira, G.A.G.: Genome structure of a Saccharomyces cerevisiae strain
widely used in bioethanol production. Genome Res. 19, 2258–2270 (2009)
23. Stacey, G.: Fundamental issues for cell-line banks in biotechnology and regulatory affairs. In:
Fuller, B.J., Lane, N., Benson, E.E. (eds.) Life in the Frozen State, pp. 437–452. CRC Press
LLC, Boca Raton, Florida (2004)
24. Health products and food branch inspectorate—Annex 2 to the current edition of the good
manufacturing practices guidelines schedule drugs (biological drugs)—GUI-0027 (2002)
25. Coecke, S., Balls, M., Bowe, G., Davis, J., Gstraunthaler, G., Hartung, T., Hay, R., Merten, O.-
W., Price, A., Schechtman, L., Stacey, G., Stokes, W.: Guidance on good cell culture practice.
ATLA 33, 261–287 (2005)
26. Nagashima, M., Azuma, M., Nogucid, S., Inuzuka, K., Samejima, H.: Continuous fermentation
using immobilized yeast cells. Biotechnol. Bioeng. 26, 992–997 (1984)
27. Bellido, J.D., Lemos, F.R., de Oliveira, M.C.A., Prota, M., Magalhães, J.S.F.: Immobilization
study of saccharomyces cerevisiae on polyurethane foam for ethanol production. Int. J. Sci.
Eng. Invest. 7, 83 (2018)
28. Thiruvengadathan, T.N.: Bioethanol Production Using Saccharomyces cerevisiae Cultivated in
Sugarcorne Juice. Master of Engineering Science. Western University Ontario, Canada (2017)
29. Bharti, B., Chauhan, M.: Bioethanol production using Saccharomyces cerevisiae with different
perspectives: substrates, growth variables, inhibitor reduction and immobilization. Ferment.
Technol. 5(2), 1000131 (2016)
An Automatic Biodiesel Decanting
System for the Optimization of Glycerin
Separation Time by Applying Electric
Field and Temperature

Kevin Bulnes , Diego Paredes , and Leonardo Vinces

Abstract During biodiesel production, crude biodiesel and glycerin are separated
in resting tanks due to gravity and differences in density, glycerin accumulates at the
base of the contender; such operation is called decantation. The decantation stage,
within the production of biodiesel based on recycled oil, takes from 8 to 24 h to
complete. Therefore, the development of an automatic biodiesel decanting system is
presented in order to optimize the production time in the line of this bio-fuel. The
process consists of applying an electric field through two electrodes at 9 kV and
simultaneously applying temperature. The results of the implementation showed
that the production time was reduced by up to 99% without affecting the quality
of biodiesel, according to the parameters of the American Society for Testing and
Materials (ASTM).

Keywords Biodiesel · Purification · Electrocoagulation · Electric field ·

Decantation stage

1 Introduction

Currently, there are several types of processes for obtaining biodiesel from recycled
cooking oils. However, the transesterification process is the most researched and
used in recent years [1]. This process consists of the chemical reaction of vegetable
or animal oils and alcohol (methanol or ethanol), through a homogeneous or hetero-
geneous catalyst. This reaction generates fatty acid methyl esters (Biodiesel) and

K. Bulnes · D. Paredes · L. Vinces (B)

Universidad Peruana de Ciencias Aplicadas, Av. Prolongación Primavera 2390, Santiago de
Surco, Lima, Peru
e-mail: [email protected]
K. Bulnes
e-mail: [email protected]
D. Paredes
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 349
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_34
350 K. Bulnes et al.

glycerin, as the main product and by-product, respectively [2, 3]. Once said the
process is finished, due to the difference in densities between biodiesel and glyc-
erin (880 kg/m3 and 1050 kg/m3 , respectively), the latter sits at the bottom of the
transesterification reactor in a time of 8 to 24 h [4, 5]. The subsequent stage of
transesterification, within the production process, is refining. Biodiesel refining is a
process that is responsible for purifying biodiesel to bring it to the quality necessary
for its consumption, impurities such as loose glycerin particles, soaps, and water [2,
6]. Among the most common refining strategies are water washing and spinning [7,
8]. As can be evidenced, the process for the production of biodiesel has a demand for
resources and time. This could have a negative impact on production at all biodiesel
scales.
The application of the electric field has been proposed by the scientific literature
with the aim of overcoming these limitations. Z. E. Ismail et al. [4] and Ratanabuntha
et al. [9] have proposed the immersion of an electric field in the biphasic mixture of
biodiesel and glycerin to accelerate the decantation. This theory has served as the
initial framework for the present investigation due to the scientific explanation and
laboratory experimentation that was carried out in both studies.
However, in the research of Ratanabuntha et al., the type of electrode they applied
has a very complex geometry, which complicates the acquisition or manufacture,
while Z. E. Ismail et al. did not carry out a study of optimal materials for the tank
and electrodes during the decanting stage.
On the other hand, Ahmad et al. [10] studied the extent to which the shape and
distance of electrode separation are involved. This proposal finds another way to
reduce the time it takes to decant the mixture of biodiesel and glycerin. However,
this study specifies that it was expected 24 h for the mixture to decant.
Noureddin et al. [11] studied the effect of temperature at the stage of decanting
where it was found that increasing this parameter above 30°C accelerates the decanta-
tion. Despite the aforementioned, this study only focused on varying said parameters
and not the effect that it could have on the application of the electric field. On the
other hand, Moncada et al. [12] make an analysis of time, economy, and technologies
when the entire process for the production of biodiesel is automated. However, the
biodiesel obtained is based on seaweed and marine microalgae, not with recycled
cooking oil.
The present study proposes to accelerate the process in the decanting stage and
thus optimize the production of biodiesel. At the same time, it is sought that such a
proposal does not affect the quality of the product. The final results showed that the
production time was reduced by up to 99% and the quality of biodiesel meets the
parameters according to the ASTM [13].

2 Development of the Proposed System

Figure 1 shows a block diagram of the proposed system. The details are shown in
the following sections according to the order specified in this diagram.
An Automatic Biodiesel Decanting System for the Optimization … 351

Design of the Level

Temperature Study of Automatic
electric field Sensor
control materials Station
system Design

Fig. 1 Block diagram of the proposed system

2.1 Design of the Electric Field System

The objective of this stage is to accelerate the decantation of glycerin. For this, an
electric field is applied through two electrodes charged with 9 kV in the biodiesel-
glycerin mixture. The magnitude of the said electric field is directly proportional to
the speed of separation of the mixture [9]. Therefore, according to

V
E= (1)
d
where V is the voltage in volts; d is the distance between electrodes in meters and E
is the magnitude of the electric field in volts per meter. It is convenient to place the
electrodes as close as possible.
The shape of the electrodes also influences the speed of separation of the glycine
from biodiesel. According to [10], the second most effective type of electrode is flat
geometry.
The electric field lines, following the path of an elliptical cylinder, which perform
the coagulation effect of glycerin [4], must be distributed vertically to achieve
maximum contact volume with the mixture. For the present design, the distance
between the electrodes is 6 cm, whereby a magnitude of 150 k V/m of the electric
field is obtained. The average field application time is 1.50 min.

2.2 Temperature Control

For temperatures above 55°C, glycerin keeps its viscosity low, and therefore it is
easier to decant [11]. With a resistance of 1.5 kW and PID control, it is controlled
that the temperature does not fall below the optimum. In control, the temperature is
carried out according to

Q = m × cp × T (2)

where Q is the heat delivered to the mixture in Joules, m is the mass of the mixture
in kilos, cp in Joules over kilos by degrees Celsius and ΔT the variation of the
temperature of the mixture in degrees Celsius. The power delivered by the electric
resistor is recalculated in real-time according to the variation of the temperature
within the mixture.
352 K. Bulnes et al.

Fig. 2 Two stainless steel electrodes 51 cm long and 1-inch width each with thickness 3 mm

The temperature control prevents the temperature from exceeding 70 degrees,

thus preventing the biodiesel from degrading to modify the acids by less than 0.5 mg
KOH/g, also preventing the temperature from falling below 50 for the displacement
of glycerin in the mixture.

2.3 Study of Materials

This stage served to select the appropriate materials to be used for the decanting
station. Due to its high corrosivity, biodiesel tends to be corrosive to some materials.
The material used for the tank designed at the station was fiberglass. This material,
in addition to supporting liquids with high corrosion rates, functions as an insulator
against the high electrical voltage used in the station.
On the other hand, stainless steel was used for the electrodes (Fig. 2). Stainless
steel has the property of supporting liquids with high corrosivity but also being a
good current conductor. This property is important to reduce energy losses and the
magnitude of the electric field.
Devices in contact with the mixture must be kept isolated from the electric field.
The Pt100 [14], the temperature sensor used in this project, has a ceramic sleeve in
the thermowell. The calorie exchanger is made of stainless steel but is connected at
the ends to the pump and water storage with polymer hoses.

2.4 Level Census Design

Figure 3 shows the interface that forms between glycerin and biodiesel when the
decanting stage, with the proposed methodology, has ended. Through the control of
the level variable, at this stage, the state of separation of the mixture is monitored
in real-time. Thus, when the decantation has been completed, the application of the
electric field must also end.
An Automatic Biodiesel Decanting System for the Optimization … 353

Fig. 3 Interface line between Biodiesel and Glycerol when the mixture finishes settling

An 8 MP CMOS camera is used together with a Raspberry Compute Module 3 +

device capable of processing images. The algorithm is based on a basic segmentation
of the “CV2” library based on the Python language. The color range of the segmen-
tation according to the Hue, Saturation, and Value (HSV) color model is between
[358 78 39.2] and [355 89.9 62.4], with the maximum-minimum glycerin shades,
respectively.

2.5 Station Automation

Using a Raspberry Compute Module 3 + computer, complete control of the decanting

station is taken. The sensors and actuators that are immersed in the development of
this project are connected to the computer to be monitored in real-time and in this
way, the decanting station works continuously. The P&ID diagram of this work is
shown in Fig. 4.

3 Results

For the evaluation of the settling station proposed in this document, 5L of the
biodiesel-glycerin mixture was fed to the settling tank. Table 1 shows the results
of the six tests performed. From this table, we show that the time, taken for the
glycerin to decant, can be reduced by up to 99% using the proposed settling station.
Table 2 compares the quality of biodiesel obtained with the parameters established
by the ASTM according to the Glycerol (Total) and Glycerol (Free) analyzes. These
analyses were performed on the biodiesel obtained in test one that was applied
electric field (Table 1). The results showed that the parameters analyzed are within the
standard set by ASTM. This allows the proposed decanting station to be considered
a satisfactory solution.
354 K. Bulnes et al.

Fig. 4 P&ID diagram of the proposed system

Table 1 Results of 5L of the mixture at 9 kV of the applied voltage

Test Number Electric Field Application Decantation Time by Percent of glycerin (%)
Time (min) Gravity (hours)
1 2.00 6.00 23
2 1.32 6.30 25
3 1.54 6.50 25
4 2.10 7.20 20
5 1.44 5.50 28
6 0.30 2.00 15

Table 2 Results of the Biodiesel analysis of test number 1 of Table 1

Analysis Type Norm/Method Limits Limits Results
Min. value Max. value
Glycerol (Total) AOCS Ca 14-15 – 0.240 0.182
Glycerol (free) AOCS Ca 14-15 – 0.020 0.002

4 Conclusion

In conclusion, the results in this investigation showed that the glycerin-biodiesel

decanting-time was reduced by up to 99% and the glycerin percentage in the biodiesel
An Automatic Biodiesel Decanting System for the Optimization … 355

produced. This result regards the parameters proposed by the American Society for
Testing and Materials (ASTM). As future work, the level of impact and the affection
to the biodiesel decantation process will be studied when higher o lower voltages
frequencies are considered.

References

1. Verma P., Sharma M.P.: Review of process parameters for biodiesel production from different
feedstocks. Renew Sustain Energy Rev 62:1063–71. doi: https://doi.org/10.1016/j.rser.2016.
04.054 (2016)
2. Shirazi M.M.A., Kargari A., Tabatabaei M. et al.: Acceleration of biodiesel–glycerol decanta-
tion through NaCl assisted gravitational settling: a strategy to economize biodiesel production.
Bioresour Technol. 134:401–406. https://doi.org/10.1016/j.biortech.2013.02.026 (2013)
3. Chouhan A., Sarma A.: Modern heterogeneous catalysts for biodiesel production: a compre-
hensive review. Renew Sustain Energy Rev 15(9):4378–99. https://doi.org/10.1016/j.rser.2011.
07.112 (2011)
4. Ismail Z.E., Moussa A.I., Deff M.M.: Utilization of High Voltage to Separate Glycerol during
Producing Biodiesel. J.Soil Sci. and Agric. Eng. Mansoura Univ. 9:329–332. https://doi.org/
10.21608/jssae.2018.35867 (2018)
5. Barbosa C., Ramírez L.N., Morales N.: Obtein of Biodiesel (ethil-ester) by Basic Catalysis to
Level Pilot Plant Oils Derived Food Industry. Revista Especializada en Ingeniería 8:99–116.
https://doi.org/10.22490/25394088.1293 (2014)
6. Ma, F., Hanna, M.A.: Biodiesel production: a review. Biores. Technol. 70, 1–15 (1999)
7. Sdrula N.: A study using classical or membrane separation in the biodiesel process. Desalination
250:1070–1072. https://doi.org/10.1016/j.desal.2009.09.110 (2010)
8. Atadashi I.M., Aroua M.K., Abdul A.: Biodiesel separation and purification: a review. Renew
Sustain Energy Rev 36:437–443. https://doi.org/10.1016/j.renene.2010.07.019 (2011)
9. Ratanabuntha T., Tonmitr K., Suksri A.: Acceleration in biodiesel production from palm oil
process by high voltage electric field. International Journal of Smart Grid and Clean Energy
7:225–230. https://doi.org/10.12720/sgce.7.3.225-230 (2018)
10. Abbaszadeh, A., Ghobadian, B., Najafi, G.: Electrostatic Coagulation for Separation of Crude
Glycerin from Biodiesel. Advances in Environmental Biology 8(1), 321–324 (2014)
11. Noureddin A., Shirazi M.M.A., Tofeily J., Kazemi P et al.: Accelerated decantation of
biodiesel–glycerol mixtures: Optimization of a critical stage in biodiesel biorefinery. Separa-
tion and Purification Technology 132:272–280. https://doi.org/10.1016/j.seppur.2014.05.011
(2014)
12. Naranjo, S.: Automation Of A Biodiesel Production Plant From Algae And Marine Microalgae.
Telematique 16, 81–105 (2017)
13. Knothe, G.: Biodiesel Fuel Quality And The Astm Standard. Palmas 31, 162–171 (2010)
14. JM Industrial Technology SA.: PT100. Iztapalapa, MX.: JM Industrial Technology SA de CV.
Retrieved from https://www.jmi.com.mx/ (2020)
Prototype Proposal for Control
and Inspection of Gas Stations

Lucas Eder de Araújo Sousa , Felipe Silveira Stopiglia ,

Luiz Vicente Figueira de Mello Filho ,
Daniela Helena Pelegrine Guimarães ,
and Maria Thereza de Moraes Gomes Rosa

Abstract The current fuel market in Brazil has high rates of the so-called “white
flag” posts, increasing the difficulty of enforcement due to the numerous types of
fraud present in this market. About 19% of the prohibitions made by the National
Agency of Petroleum, natural gas and biofuels (ANP) in 2017, relating to the scams of
minor bombs, which is applied to gas stations whose marketing is currently addicted
to quantity. Regarding inspection, the problem exposed here is that most posts that
practice this type of fraud use the technology to make it difficult to identify, either
by the user or by the inspection agency. The prototype developed in this work can
support the user complaint by allowing the monitoring of his vehicle supply from
a mobile application, connected to a flow meter, which is activated at the time of
supply, in case of disagreement between pump and application the user gives access
to the ANP complaints form, aiming at further investigation and records linked to
this type of fraud. This study shows that over 80% of surveyed consumers do not
feel safe about refueling your vehicle, and they are interested in the proposed device.
This device does not identify fuel chemical tampering.

Keywords Fraud · Gas station · Fuel · Control · Prototype

L. E. de Araújo Sousa (B) · F. S. Stopiglia · L. V. F. de Mello Filho ·

M. T. de Moraes Gomes Rosa
Center of Science and Technology, Mackenzie Presbyterian University, Brazil Avenue, 1220,
13073-148 Campinas, SP, Brazil
e-mail: [email protected]
D. H. P. Guimarães
Lorena Engineering School (EEL-USP), São Paulo University, Campinho Highwat, 12602-810
Lorena, SP, Brazil

© The Editor(s) (if applicable) and The Author(s), under exclusive license 357
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_35
358 L. E. de Araújo Sousa et al.

1 Introduction

According to data from EPE [1], liquid fuels, from renewable sources or not, represent
40% of the Brazilian energy matrix, and the major consumers of this energy processed
in Brazil are the transportation and industry sectors, demanding 61% and 6,90%
respectively.
The oil industry generates goods that are hard substitute inputs in the produc-
tion matrix of any country, which are of paramount importance for production
and consumption in modern society. This importance can be evidenced mainly in
petroleum derivatives used with liquid fuels for combustion engines (e.g. gasoline,
diesel). Without it, the productive and modern industry would not function [2].
In this energy industry of great importance to the national economy, there are also
fuel stations and distributors responsible for directing fuel to the final consumer. Data
from the 2017 National Petroleum, Natural Gas and Biofuels Agency (ANP) show
that there are currently 41,829 regulated gas stations in Brazil, which about 40% are
white flag gas stations, that is, they buy fuels freely without compulsory binding a
specific distributor [3].
The existence gas stations, despite contributing to increase competition and lower
prices, makes checks more difficult, since adultery can come not only from clan-
destine distributors but from the establishments themselves that, by not following
the regulations required by the supervisory bodies, can bring economic and environ-
mental impacts, given that their activity influences a whole supply chain and it is, of
course, extremely polluting [4–6].
Given the scenario above, it is possible to notice the difficulty that the ANP has
when it comes to inspecting and noting mainly the irregular gas stations. In 2017,
out of the 22,134 complaints received by the agency, only 67% were answered and
of these 12,910 were inspection actions performed at fuel dealers, where only 4064
were fined in some way. So, there is a large deficit between the complaints and gas
stations effectively fined, and this is due not only to the difficulty ANP has in carrying
out the inspections, but also to the doubt of the consumers, who make the complaint
without being sure if they are really being injured.
Enforcement action has a cost, and there are a lot of actions taken that did not reach
the assessments and enforcement measures, this waste of resources is immense. This
is mainly because the current system is flawed and it lacks technology that could
identify and highlight sectors that really need to be scrutinized on the spot, saving
resources and protecting the market from the economic inefficiency.
Understanding the importance of the fuel sector to the national economic scenario
and the problem exposed here with the difficulty of enforcement and regulatory
standards, this paper aims to develop a device that facilitates the identification of
minor pump fraud and mediate the way ANP receives the complaints too.
Prototype Proposal for Control and Inspection of Gas Stations 359

2 Methodology

Ethical approval for the experiment in this study was deemed unnecessary by the
Ethics Committee. As also the activities in the experiment do not pose risks greater
than those ordinarily encountered in daily life and all the subjects remained anony-
mous. As well as all participants were given consent and were also informed about the
academic objective of the research as well as the guarantee of anonymity giving the
choice of participation or not. Still considering that as it is an opinion poll, according
to Resolutions 466 and 510 of the National Health Council, there is no need to
go through the University’s Ethics Committee. This is following CNS Resolution
466/12 of the National Health Council, of December 12, 2012, (available at https://
conselho.saude.gov.br/resolucoes/2012/Reso466.pdf). Idem with Article 1 of Reso-
lution 510, of 7 April 2016, of Brazil’s National Health Council (available at http://
www.conselho.saude.gov.br/resolucoes/2016/Reso510.pdf) and also with item 8.0.5
of the Ethical Principles of Psychologists and Code of Conduct of the American
Psychological Association (available at https://www.apa.org/ethics/code/principles.
pdf).
In order to ascertain the real need for this control device and the gas stations
enforcement, and to support its development, the current study was elaborated in a
quantitative model that, according to Fonseca [7], can be characterized as an objective
study, which considers that reality can be understood based on the analysis of raw data
and through the aid of standardized and neutral instruments. A survey was conducted,
elaborated with closed questions in order to check the consumer’s confidence in
gas stations where they are consumers, and to measure the level of interest in the
device developed. The research was conducted in a structured questionnaire model,
developed in Google Forms. The model with the questions was disseminated through
e-mail and social media, being used for convenience by selecting individuals who are
available to participate in the study. According to Fonseca [7], among the benefits
of conducting a survey are the agility and economy in obtaining data, as well as
obtaining data grouped in tables, which allow a rich statistical analysis.
From the extraction of data through the tools available in the Google Forms
and Google Sheets platforms, it was found that from the sample of 30 responses
collected, about 88% of respondents have already questioned the amount supplied
and the amount that was presented by the pump and, of these, 86% are interested in
buying of the product. With this scenario established, the experimental procedures
were started, which, according to Gil [8], consists in determining the study object and
the variables capable of influencing it, defining the forms of control and observation
of the effects that the variable produces on the object. For this, it was proposed a new
product concept, which aims to detect the frauds in gas stations and make accessible
the submission of complaints by the user.
360 L. E. de Araújo Sousa et al.

2.1 Materials for Product Composition

According to Morrison [9], the low acceptance rate of new products is directly related
to the 21st-century consumer’s profile, which is richer, more demanding and with
high expectations of quality, service, and design, all offered at a low price.
Based on this precept, the activity of defining and selecting materials for the
composition of a product has a strong influence on its functional and/or marketing
success, since it is subject to aspects such as respect for the environment, usability,
and even economic viability. Because these are many variables to be considered,
measuring the efficiency of material selection and subsequent product development
becomes a difficult task, in which there are even divergences within the literature
[10].
After the devices available research for the construction of the project, it was
possible to obtain a list of materials with sufficient characteristics to compose it. In
view of the literature and the available materials found, the criterion for selecting
those that were used in this project was the cost/benefit. The components listed here
are presented in the functional prototype.
• DN20 pipe 3/4” water flow hall sensor Hall Flowmeter: consists of a plastic fairing,
coupled blade rotor, and Hall effect sensor. When the fluid begins to flow inside, the
paddles coupled to the rotor move, and when the paddles pass through the sensor,
the device sends a 5 V pulse through the communication cable. The sensor can
operate at a flow rate of up to 30 L per minute, pressure below 1.75 MPa, and has
3% accuracy in its results as indicated by the manufacturer. The Reading of the
input pulses will be converted to absolute values by the coupled microprocessor,
the Arduino.
• Arduino Board UNO R3: Arduino can be described basically like an opensource
single-board microcontroller that allows data input and output configurations and
is easily programmed using the C/C++ language. Widely used in academia, it
is possible to combine this tool with other devices, establishing communication,
reading and sending data through them. In the project, Arduino was crucial for
reading the data emitted by the flow sensor and sending it to the mobile appli-
cation through the Bluetooth module attached to the board. Receiving, sending
and processing the data were developed through a “Sketch” code that will be
approached later.
• Bluetooth Module—HC05: This module was coupled to the Arduino for the
purpose of sending information (output) from it. It was possible to estab-
lish communication between the microprocessor and the mobile application
responsible for displaying the data.
• Android App(MIT App Inventor): This is a visual and intuitive programming envi-
ronment capable of creating functional applications for smartphones and tablets.
The tool treats programming as blocks and facilitates the creation of complex
applications for users unfamiliar with the common software language.
Prototype Proposal for Control and Inspection of Gas Stations 361

Within the platform was developed an application capable of reading the data
emitted by the Arduino, controlling the devices attached to it, as well as facilitating
access to the ANP’s complaint panel.

3 Results and Discussion

After selecting the materials to compose the device, it was necessary to establish a
connection between them. The electrical scheme used is shown in Fig. 1.
The Arduino microprocessor board, after programming input, only needs a 12 V
power connection. If it is properly configured, it will operate constantly, and does
not require external interference. The flowmeter is connected to port 2, a digital pin
that allows it to work as an interrupt, thereby detecting the number of pulses emitted
by the Hall sensor in a given time frame. The sensor was also connected to power
ports 5 (V) and GND, responsible for power and grounding, respectively.
The Bluetooth was connected to Arduino’s digital ports 0 (RX) and 1 (TX). Also
known as SERIAL, these ports are responsible for communication, and in this case,
outputting information from the Arduino. The Bluetooth module was also connected
to the power supply port 5 (V) and the grounding GND port.
The mobile application has in its own layout an option to connect to a tablet. After
pairing with Bluetooth and connecting to the same device, it is ready for use.

3.1 Sketch Code Used for Microprocessor Programming

The language used is suitable for Arduino programming and closely resembles C++
programming (Fig. 2). To establish the connection between the three components,
some researches were conducted on similar projects that used the device as a whole
for other purposes, such as water flow reading or irrigation control.

Fig. 1 Device electrical diagram

362 L. E. de Araújo Sousa et al.

Fig. 2 Programming codes

3.2 Data Read Code

With the correct functioning of the device, it was necessary to find ways to constitute
the communication app. With the platform defined, all the code was written in blocks,
with the aid of research and studies in the sources mentioned below (Fig. 3).
Within the software, besides creating a box for automatic display of how much was
filled, four buttons have been configured with distinct functions that aim to make the
user’s life easier, according to Fig. 3. These buttons are: Connect/Disconnect buttons
to Bluetooth; button to reset the liters display after measurement; button to direct the
user to the ANP’s report page. Because it is a relatively simple platform, the model
contains some layout and operation constraints, but fully meets the proposal (Fig. 4).

3.3 Device Calibration

After the communication between device and application was occurring without fail
and with the assembly of the device properly completed (Fig. 5), it was necessary to
calibrate the device so that the displayed numbers corresponded to the value supplied
with the lowest possible error coefficient.
Calibration was performed using a 0.02 (g) precision scale, a 500 ml water
measurement and the device powered by a 3/4 hose. Through this procedure it was
Prototype Proposal for Control and Inspection of Gas Stations 363

Fig. 3 Application code in blocks

Fig. 4 Application layout

364 L. E. de Araújo Sousa et al.

Fig. 5 Device connected and turned on

Table 1 Experimental testing for vendor reported error

Measurements Observed value (ml) Error percentage
1 530 6
2 510 2
3 540 8
4 520 4
5 510 2
6 520 4
7 510 2
8 530 6
9 540 8
10 530 6
Average 524 5

possible to convert the pulses to liters, described in item 3.2 of this article, the
observed value corresponds to the hall effect meter manufacturer, where each liter is
equivalent to 390 pulses. The error aggregated to the device is 5%, information also
provided by the manufacturer and observed in tests presented in Table 1.

3.4 Improvements

Nevertheless, the prototype has points to be worked on and developed in further

research, such as the device’s adaptation to the vehicle and its accuracy when using
variable flow rates. Regarding these points, difficulties were found due to the need for
Prototype Proposal for Control and Inspection of Gas Stations 365

the adaptation to the tank to be universal and versatile, aiming to maintain the device’s
low cost and accessibility. Proposals that can be considered for this adaptation would
be to use a rubber hose fitted to the gas tank inlet and connected to the flow sensor
at the end of it. The other device components could be allocated externally, near the
tank or even under the passenger seat, as these are relatively small materials and it
would not be a problem. Another proposal that could be considered for this problem
would be the use of the sensor in the external region of the tank, being adapted with a
screw nozzle and manually coupled directly to the supply nozzle during the process
and removed later. But in this case, the use of the device would be dependent on the
attendant’s allocation at the time of fueling.

4 Conclusion

This study showed consumer insecurity regarding current supply control methods.
This insecurity becomes even greater when considering the evolution of fraudulent
bombs and the stagnation of supply detection methods of fraud, resulting in the
conversion of only 30% of complaints into cases. This research has brought to the
fuel market a new concept of a product to be worked on, which can help consumers
and the law enforcement agencies to detect and report bomb fraud at gas stations. The
device developed is functional, captures the actual volume supplied by the pump and
give to consumers the basis for formalizing his complaint to the ANP, by completing
the agency complaint form, which can be accessed through the application presented
here, developed throughout the project.

References

1. EPE.: Política Nacional de Biocombustíveis. Empresa de Pesquisa Energética, Rio de Janeiro,

RJ (2006)
2. Canelas, A.L.S.: Evolução da importância econômica da indústria de petróleo e gás natural
no Brasil: contribuição a variáveis macroeconômicas. Mestrado em Engenharia Ciências do
planejamento energético, COPPE/UFRJ (2007)
3. ANP—Agência nacional do petróleo, gás natural E Biocombustíveis.: Anuário estatístico
brasileiro do petróleo, gás natural e biocombustíveis. Rio de Janeiro: ANP (2017)
4. Rocha, J.A.: Padrões de concorrência e estratégias empresariais no setor de distribuição de
derivados de petróleo no Brasil. Instituto de Economia. Universidade Federal do Rio de Janeiro,
Rio de Janeiro, Rio de Janeiro (2002)
5. Dalmonech, L.F., Sant’anna, J.M.B., Teixera, A.C.C.: Análise dos fatores intervenientes
nas quebras de contrato no setor de combustíveis brasileiro, RAE eletrônica, Escola de
administração de empresas, São Paulo (2010)
6. Machado, F.H., Ferreira, O.M.: Postos de combustíveis: quantificação e qualificação da
atividade no município de Goiânia. Universidade Católica de Goiás, Goiás (2007)
7. Fonseca, J.J.S.: Metodologia da pesquisa científica. UEC, Fortaleza (2002)
8. Gil, A.C.: Como elaborar projetos de pesquisa, 4th edn. Atlas, São Paulo (2007)
9. Morrinson, I.A.: Segunda Curva. Campus, Rio de Janeiro (1997)
366 L. E. de Araújo Sousa et al.

10. Ferroli, P.C.M.: MAEM-6F (Método auxiliar para escolha de materiais em seis fatores): suporte
ao design de produtos, industriais. Doutorado em Engenharia de Produção. Universidade Santa
Catarina, Florianópolis (2004)
Modeling of Solid-Liquid Equilibrium
in Nitrobenzene and Para-Xylene
Mixtures

Michelli Fontana , Caroline dos Santos Ribas ,

Paulo Henrique Schuck , Rafael Lopes Turino , and André Zuber

Abstract Solid-liquid mixtures are very important for the study of molecular inter-
actions that describe the thermodynamic properties of the components. The objec-
tive of this work was to determine the behavior of the nitrobenzene and para-xylene
mixtures by means of two excess energy models, Margules and Non-Randon Two-
Liquid (NRTL), comparing the calculated results with experimental data from the
literature. The Margules model, whose parameters varied with temperature, was more
accurate than the model that did not consider the temperature variation. The NRTL
model was also accurate, presenting only a small deviation in the lower para-xylene
compositions, while the ideal case and Margues model with temperature indepen-
dence presented a large imprecision. Therefore, the two-parameter Margules model
with temperature dependence presented the best approximation of the equilibrium
system.

Keywords Nitrobenzene · Para-xylene · Solid-liquid · NRTL · Margules ·

Equilibrium

M. Fontana · C. dos Santos Ribas · P. H. Schuck · R. L. Turino · A. Zuber (B)

Academic Department of Engineering, Federal University of Technology—Paraná (UTFPR),
Francisco Beltrão, Brazil
e-mail: [email protected]
M. Fontana
e-mail: [email protected]
C. dos Santos Ribas
e-mail: [email protected]
P. H. Schuck
e-mail: [email protected]
R. L. Turino
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 367
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_36
368 M. Fontana et al.

1 Introduction

In solid-liquid mixtures, crystallization is a process where thermodynamic conditions

lead the molecules to approach and group in highly organized structures, reaching
a state of solid-liquid equilibrium. Among many ways to describe the solid-liquid
equilibrium of a binary system, establishing the parameters of an excess energy
model may be satisfactory to represent the behavior of the substances in the mixture.
In the phase equilibrium calculations, one of the most important steps is fitting
adjustable parameters of activity coefficient models to experimental data [1]. Each
model has a certain number of parameters, and their estimation is a common chal-
lenging problem. Considering the importance of the study of solid-liquid equilibria,
the mixture of xylene and nitrobenzene emerges as an interesting case.
Xylenes are light aromatic hydrocarbons derived from petroleum and processed in
petrochemical industries. Para-xylene is part of BTXs (benzene, toluene, and xylene),
which may be used to react to producing compounds as nitrobenzene [2]. The study
of the equilibrium of nitrobenzene and para-xylene may contribute to describing its
synthesis [3]. To the best of our knowledge, the solid-liquid equilibrium modeling
of nitrobenzene and para-xylene mixtures using the two-suffix Margules and the
Non-Random Two-Liquid (NRTL) models was not investigated in literature before.
Thus, the aim of this study was to determine the behavior of nitrobenzene and
para-xylene mixtures by means of two excess energy models, the two-suffix Margules
with dependence and independence of temperature and the NRTL, identifying the
eutectic point of the mixture and comparing the calculated results with those obtained
experimentally by Proust and Fernandez [4].

2 Methods

From the Gibbs excess energy definition of the two-suffix Margules model (Ge ) (1),
using the definition of Eq. (2), we obtained the equations that determine the activity
coefficients (γ) (3 and 4).

G e = x1 · x2 · [A + B · (x1 − x2 )] (1)

R · T · ln[γ1 ] = [∂(n · G e )/∂n 1 ] P,T,n 2 (2)

ln[γ1 ] = [(A + 3 · B) · x22 − 4 · B · x23 ]/(R · T ) (3)

ln[γ2 ] = [(A − 3 · B) · x12 + 4 · B · x13 ]/(R · T ) (4)

The activity coefficients were fitted using compositions obtained experimentally

by Proust and Fernandez [4] for different temperatures, considering molar fractions
in the liquid phase for para-xylene (x 1 ) and for nitrobenzene (x 2 ), and using an initial
Modeling of Solid-Liquid Equilibrium in Nitrobenzene … 369

estimation for parameters A and B. The calculation of the activity coefficients was
repeated considering parameters A and B temperature-dependent.
Theoretical compositions were calculated by means of the equilibrium condition
with the variation of activity coefficients for each temperature. Solid and liquid heat
capacity were considered the equal and the theoretical composition of the liquid
phase in equilibrium was calculated for each experimental temperature (5).

ln[xi · γi ] = [h f,T /R] · [1/T −1/Tm ] (5)

The deviation between theoretical and experimental compositions (6) of the calcu-
lated liquid phase (ARD %) was minimized by varying the parameters (A and B) of
activity coefficients (3 and 4). The deviations were evaluated neglecting experimental
or calculated values for molar fractions equal to zero or one (pure substances).
 exp  
ARD(%) = 100 ·  (x1calc − x1 ) /(x1 ) /(total points)
exp
(6)

The second model used was NRTL (7).

G e /(x1 · x2 · R · T ) = G 21 · τ21 /(x1 + x2 · G 21 )

+ G 12 · τ12 /(x2 + x1 · G 12 ) (7)

Taking the definition presented by Eq. (2), we obtain (8 and 9).

  
ln γ1 = x22 τ21 · (G 21 /(x1 + x2 · G 21 ))2

+(G 12 · τ12 /(x2 + x1 · G 12 )2 ) (8)

  
ln γ2 = x12 τ12 · (G 12 /(x2 + x1 · G 12 ))2
 
+ G 21 · τ21 /(x1 + x2 · G 21 )2 (9)

With the parameters defined as (10, 11, 12 and 13).

G 12 = exp(−α · τ12 ) (10)

τ12 = b12 /(R · T ) (11)

G 21 = exp(−α · τ21 ) (12)

τ21 = b21 /(R · T ) (13)

The same procedure used to obtain theoretical compositions in Margule’s excess

energy model was used to obtain the theoretical compositions of the mixture for the
NRTL model. Subsequently, the theoretical composition of the liquid phase of the
mixture was calculated for the ideal case. Table 1 shows the properties of the pure
components used in this study.
370 M. Fontana et al.

Table 1 Properties of nitrobenzene and para-xylene

Parameter Para-xylene Nitrobenzene
Melting point (K) 286.29 278.79
h273,15
fus
K
(J/mol) 16,810 11,590
Molecular mass (g/mol) 106.16 123.06

3 Results and Discussion

Both Figs. 1 and 2 show that different regions can be identified in the phase diagram
of the models. These regions separate the zones of the existence of different phases.
Eutectic line defines two regions: under it, there is a solid phase composed of para-
xylene and nitrobenzene, and over it, there are three other regions. The left region
is composed of a nitrobenzene solid phase and a liquid mixture of both compounds.
The right region is composed of para-xylene in the solid phase and a liquid mixture.
And the central upside region is composed of a liquid mixture of both compounds.
NRTL usually leads to reasonable phase equilibria predictions [5, 6]. NRTL model
(Fig. 1) provided eutectic point compatible with experimental data, both in temper-
ature (259.22 K) and composition ([x 1 = 0.38] and [x 2 = 0.62]). However, it was
evidenced that along the curve for low para-xylene molar fraction, there was a devi-
ation when compared to the experimental one (ARD = 3.56%). The parameters of
this model were: α equal to 0.015, b12 equal to 9.896 and b21 equal to 10.437. Also,
in order to define the behavior of all parameters, some correlations were obtained as

287

282

277
Temperature [K]

272

267

262

257

252
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
x1
NRTL Experimental* Eutectic line

Fig. 1 Phase diagram of the NRTL model to para-xylene (x 1 ) and nitrobenzene mixture
*Experimental data from Proust and Fernandez [4]
Modeling of Solid-Liquid Equilibrium in Nitrobenzene … 371

287

282

277
Temperature [K]

272

267

262

257

252
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
x1
Margules of two parameters dependent on temperature
Experimental*
Eutectic line

Fig. 2 Phase diagram of the Margules model of two parameters dependent on temperature to
para-xylene (x 1 ) and nitrobenzene mixture *Experimental data from Proust and Fernandez [4]

follows:

G 12 = −0.1691 · x15 + 0.4482 · x14 − 0.407 · x13 + 0.1348 · x12

 
− 0.008 · x1 + 0.0063; R 2 = 0.977 .
G 21 = −0.1372 · x15 + 0.3635 · x14 − 0.3301 · x13 + 0.1094 · x12
 
− 0.0065 · x1 + 0.0048; R 2 = 0.9766 .
τ12 = 1341.1 · x15 − 3554.1 · x14 + 3221.7 · x13
 
− 1046.4 · x12 + 49.27 · x1 + 330.03; R 2 = 0.984 .
τ12 = 1414.4 · x15 − 3748.4 · x14 + 3397.8 · x13
 
− 1103.6 · x12 + 51.962 · x1 + 348.06; R 2 = 0.984 .

In the two-suffix Margules model of Gibbs excess energy with temperature depen-
dence (Fig. 2), there was a better estimation of the composition as a function of
temperature. Thus, the calculated eutectic point was compatible with the experi-
mental one, both in temperature (259.22 K) and in molar composition ([x 1 = 0.38]
and [x 2 = 0.62]). The correlations of the molar fraction of nitrobenzene with both
parameters A and B are presented below. These correlations do not consider the
points x 1 = 0 and x 1 = 1, because these points did not adapt to any model proposed.

A = −27 · x16 + 47 · x15 − 57 · x14 + 37 · x13 − 86 · x12

372 M. Fontana et al.
 
+ 16 · x1 − 70450; R 2 = 0.983 .
B = 76 · x16 − 27 · x15 + 17 · x14 − 36 · x13 − 712407 · x12
 
+ 468146 · x1 − 60596; R 2 = 0.984 .

Despite its simplicity, the Margules equations showed a reasonable quality (ARD
= 0.01%) for the description of eutectic solid-liquid equilibrium profiles [7].
In the two-suffix Margules model of Gibbs excess energy with temperature inde-
pendence (Fig. 3), (ARD = 27.15%) the eutectic point was also at the same temper-
ature (259.22 K) of that obtained with experimental data. However, the composition
of molar fraction in the liquid phase was different (experimental [x 1 = 0.38] and [x 2
= 0.62], model [x 1 = 0.27], and [x 2 = 0.73]). The parameter A of this model was
equal to 10,833.546 and B were equal to 185.106.
In the ideal case (Fig. 4), both the range of the para-xylene composition from
0.2–0.8 as the eutectic point was different from experimental data (ARD = 7.93%).
This can be justified by the fact that in the ideal case, the interactions between distinct
molecules are disregarded.
Margules model, whose parameters varied with temperature, was more accurate
than the one without considering temperature changes. In fact, this result is in good
agreement with the observations presented by Rocha and Guirardello [8] that also
used the two-parameter Margules equation, with temperature dependence, to model
the liquid phase in solid-liquid equilibria for many other binary systems.

287

282

277
Temperature [K]

272

267

262

257

252
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
x1
Margules of two parameters independent on temperature
Experimental*
Eutectic line

Fig. 3 Phase diagram of the Margules model of two parameters independent on temperature to
para-xylene (x 1 ) and nitrobenzene mixture *Experimental data from Proust and Fernandez [4]
Modeling of Solid-Liquid Equilibrium in Nitrobenzene … 373

287

282

Temperature [K] 277

272

267

262

257

252

247
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
x1
Ideal Experimental* Eutectic line

Fig. 4 Phase diagram of the ideal case to para-xylene (x 1 ) and nitrobenzene mixture *Experimental
data from Proust and Fernandez [4]

Furthermore, NRTL was also accurate, showing a small deviation in the lower
molar composition of para-xylene. NRTL model is also efficient to provide a great
approximation of solid-liquid equilibrium to other systems [9]. Literature also brings
the use of Margules and NRTL models to represent other solid-liquid equilibrium
systems. Pelaquim et al. [10] used NRTL and three-suffix Margules models to
describe the properties of the liquid of many solid-liquid mixtures of fatty acids.
Likewise, Matos et al. [11] evaluated binary solid-liquid equilibrium of fatty acids,
fatty alcohols, and triolein, in which the liquid lines of these systems were adequately
described. Three-suffix Margules and NRTL models were also accurate in repre-
senting the solid-liquid equilibrium of octacosanol in three pure solvents 1-pentanol,
1-hexanol, and toluene [12].
Margules model with temperature independence, however, showed great devia-
tion when compared to experimental data, even bigger than the ideal case deviation.
This is justified by the lack of information that the absence of temperature depen-
dence causes. This mixture shows an increase in the intensity of attractive molecular
interactions with increasing system temperature [3]. The literature also highlights
the need to develop unified interaction parameters to represent all possible phase
equilibria [5].
374 M. Fontana et al.

4 Conclusion

This study presented an analysis of Gibbs’s excess energy models to describe the
equilibrium system of the solid-liquid mixture of para-xylene and nitrobenzene. It
was found that the Margules model of two parameters with temperature depen-
dence was the best to represent the equilibrium of the system. NRTL model also
provided a good prediction of the experimental data. However, Margule’s model of
two parameters independent of temperature did not represent the mixture effectively.
The representation of the ideal case was also not accurate.
Although some models are known to represent well mixtures, there are some cases
that they can diverge, so it is very important to compare experimental data with the
modeling results. The agreement between the predicted results by means of Margules
and NRTL models with the experimental phase equilibrium data contributes to the
modeling of solid-liquid mixtures with nitrobenzene and para-xylene.

Acknowledgements To the Federal University of Technology—Paraná (UTFPR).

References

1. Vatani, M., Asghari, M., Vakili-Nezhaad, G.: Application of Genetic Algorithm to the calcu-
lation of parameters for NRTL and Two-Suffix Margules models in ternary extraction ionic
liquid systems. J. Indus. Eng. Chem. 18, 1715–1720 (2012). https://doi.org/10.1016/j.jiec.2012.
03.008
2. Franck, H.G., Stadelhofer, J.W.: Production of benzene, toluene and xylenes. In: Industrial
Aromatic Chemistry. Springer, Berlin, Heidelberg. pp. 99–131 (1988). doi: 10.1007/978–3-
642-73432-8_4
3. Rafiee, H.R., Frouzesh, F.: The study of partial and excess molar volumes for binary mixtures
of nitrobenzene and bensenaldehyde with xylene isomers from T = (298.15 to 318.15) K and
P = 0.087 MPa. J. Adv. Res. 7(5), 769–780 (2016). https://doi.org/10.1016/j.jare.2015.11.003
4. Proust, P.C., Fernandez, J.C.: Experimental solid-liquid equilibria of binary mixtures of organic
compounds. Fluid Phase Equilib. 29, 265–272 (1986)
5. Moudjari, Y., Louaer, W., Meniai, A.: Modeling of solid-liquid equilibrium using a modified
group contribution based NRTL model. Fluid Phase Equilib. 492, 118–136 (2019). https://doi.
org/10.1016/j.fluid.2019.03.021
6. Zhang, X., Zhao, Y., Liu, Y., Tang, W.: The effects of solvent properties on solid-liquid phase
equilibrium of ethylene thiourea. J. Mol. Liq. 285, 459–467 (2019). https://doi.org/10.1016/j.
molliq.2019.03.178
7. Maximo, G.J., Carareto, N.D.D., Costa, M.: Chapter 8—Solid-liquid equilibrium in food
processes. In: Thermodynamics of Phase Equilibria in Food Engineering, pp. 335–384 (2019).
https://doi.org/10.1016/b978-0-12-811556-5.00008-9
8. Rocha, S.A., Guirardello, R.: An approach to calculate solid-liquid phase equilibrium for binary
mixtures. Fluid Phase Equilib. 281, 12–21 (2009). https://doi.org/10.1016/j.fluid.2009.03.020
9. Tan, Q., Leng, Y., Wang, J., Huang, C., Yuan, Y.: Correlation and prediction of the solubility
of the racemic tartaric acid–ethanol-water system with the NRTL model. J. Mol. Liq. 216,
476–483 (2016). https://doi.org/10.1016/j.molliq.2016.01.080
10. Pelaquim, F.P., Matos, F.C., Cardoso, L.P., Batista, E.A.C., Meirelles, A.J.A., Costa, M.C.:
Solid-liquid phase equilibrium diagrams of binary mixtures containing fatty acids, fatty alcohol
Modeling of Solid-Liquid Equilibrium in Nitrobenzene … 375

compounds and tripalmitin using differential scanning calorimetry. Fluid Phase Equilib. 497,
19–32 (2019). https://doi.org/10.1016/j.fluid.2019.05.020
11. Matos, F.C., Costa, M.C., Meirelles, A.J.A., Batista, E.A.C.: 12. Binary solid-liquid equilibrium
systems containing fatty acids, fatty alcohols and triolein by differential scanning calorimetry.
Fluid Phase Equilib. 404, 1–8 (2015). https://doi.org/10.1016/j.fluid.2015.06.015
12. Cuevas, M.S., Crevelin, E.J., Moraes, L.A.B., Oliveira, A.L., Rodrigues, C.E.C., Meirelles,
A.J.A.: Solubility of commercial octacosanol in organic solvents and their correlation by ther-
modynamic models at different temperatures. J. Chem. Thermodynamics 110, 186–192 (2017).
https://doi.org/10.1016/j.jct.2017.02.025
X-Ray Characterization of TiO2
Obtained by Chemical Route
for Photocatalysis

Tiago Siqueira Lima , Wesley Rafael Penteado ,

André Luis de Castro Peixoto ,
and Ademir Geraldo Cavallari Costalonga

Abstract The need for freshwater in our time has become a problem of important
concern. Recovered water for new uses makes purification an essential process to
achieve a desired degree of quality. This need depends on each different purpose,
such as purification of drinking water or reuse water for industrial needs. Due to the
toxic characteristics of non-biodegradable organic pollutants, new treatment tech-
nologies must be studied and advanced oxidation processes (AOPs) come up as
suitable for degradation of a wide range of organic contaminants in polluted water.
For catalyst effectiveness, the crystalline phase of TiO2 present contributes to photo-
catalytic activity in the material. This work involved the synthesis of titanium dioxide
from Pechini’s method using citric acid, titanium isopropoxide, and ethylene glycol,
obtaining samples of titanium dioxide in calcination temperatures from 500, 600, and
700 °C. From the X-ray diffraction results, it is possible to check the TiO2 crystalline
phases present in the samples synthesized at different temperatures and identified as
T500, T600, and T700. For sample T500, there was a mixture of phases between
anatase and rutile, and for samples T600 and T700, the predominant phase was rutile.
With this, we can say that the calcination process influenced the polymorphic compo-
sition of the material, increasing the rutile phase with increasing temperature, which
brings undesirable characteristics for application in photocatalytic processes.

Keywords Photocatalysis · Titanium dioxide · Rietveld’s method

1 Introduction

The need for freshwater in our time has become a problem of important concern. The
water is not only for industrial purposes but also for domestic uses, and many sources
of freshwater have been polluted due to industrial activity. Many different chemicals

T. S. Lima · W. R. Penteado · A. L. de Castro Peixoto · A. G. C. Costalonga (B)

IFSP—Instituto Federal de Educação Ciência e Tecnologia de São Paulo Câmpus Capivari,
Avenida Doutor Ênio Pires de Camargo, 2971, Capivari, SP-CEP 13360-000, Brazil
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 377
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_37
378 T. S. Lima et al.

are discharged into the aquatic environment. Some of them are biodegradable and
nontoxic, however, others not are.
Due to the toxic characteristics of non-biodegradable organic pollutants, new
treatment technologies must be studied and advanced oxidation processes (AOPs)
come up as suitable for degradation of a wide range of organic contaminants in
polluted water.
In the previous work,1 we have focused on optimizing photocatalysis, as many
factors interfere with photocatalytic activity, finding the best conditions for the degra-
dation of methyl orange [1], working on a variation of the amount of TiO2 and the
temperature of the solution. We also focus on optimizing photocatalysis as an impor-
tant issue to make the process viable and present experimental results using a mixture
of HNO3 and NaNO3 in aqueous solutions to optimize photocatalytic activity [2].
Heterogeneous photocatalysis is an example of AOP that is showing good results
for the oxidation of these pollutants. TiO2 is potentially the best catalyst that
researchers chose for this reaction, due to its high efficiency, physical and chemical
stability, low cost, and low toxicity [3].
For this catalyst, the crystalline phase of TiO2 present in the material contributes
to the efficiency of photocatalytic activity in the material, P25 Degussa is the one
that presents the best results, one of the aspects that contributes to its efficiency
is the phase mixture (75% anatase and 25% rutile) [4]. The photocatalyst can be
synthesized when one wants to evaluate some variables in the synthesis process and
effect on photocatalytic activity.
The most commonly used TiO2 synthesis routes [5–8] can be classified according
to the reaction medium (liquid or gas). In the gas phase, the most commonly used
methods are chemical vapor deposition (CVD) and liquid phase flame oxidation,
and the most reported methods are sol–gel, microemulsion, polymeric precursors,
and homogeneous precipitation. Synthetic routes performed in the liquid phase are
preferable when obtaining powders and films, as they allow greater stoichiometric
control of reagents which gives greater morphological control [6].
Work is developed [5, 6] in the synthesis of titanium dioxide from the Pechini
method (1967) using citric acid, titanium isopropoxide, and ethylene glycol,
obtaining five samples of titanium dioxide in 1:1, 2:1 stoichiometric proportion,
3:1, 4:1, and 5:1. First, a titanium citrate solution was prepared, and then, polyester-
ification reaction with ethylene glycol was added until the resin was formed which
was calcined at 400 °C for 1 h. This resin was macerated and sieved for 1 h calci-
nation at a temperature of 500 °C, and later, the compound was characterized by
X-ray diffraction. In diffraction, it was possible to determine that according to the
variations of citric acid concentrations, 78.99–94.70% were obtained in the anatase
crystalline phase.
Synthesis and crystallization methods that offer the best cost-effectiveness and
speed in terms of improving the properties that qualify these materials for techno-
logical applications are of interest. Some parameters such as the crystalline phase,

1 Noteof transparency that indicates that it was based on earlier work by one of the co-authors,
according to Refs. [1, 2].
X-Ray Characterization of TiO2 Obtained … 379

surface area, morphology, crystallinity, particle size, and porosity are influenced by
the heat treatment employed in TiO2 synthesis [4].

2 Materials and Methods

2.1 Synthesis and Characterization of TiO2

To perform the synthesis, the adapted Pechini (1967) method was used. Initially, a
citric acid solution was prepared by dissolving 21 g of the reagent in approximately
30 mL of distilled water. The addition of isopropanol (20 mL) was continued under
heating to 60 °C. For the formation of titanium citrate, it was necessary to prepare
a 30 mL 1:2 ratio of titanium isopropoxide: isopropanol, which was slowly poured
into the citric acid solution while stirring until slightly liquid formed yellowish and
clear.
The solution was boiled, and upon reaching approximately 50, 10 mL of glycerin
was added. The mixture was stirred and heated until a clear, viscous gel formed.
Subsequently, 4 mL of concentrated H2 O2 was poured, thereby developing an intense
orange tinge. The gel was muffled and pyrolyzed at 300 °C. Finally, the synthesis
product was fractionated in three to study the effect of calcination temperature on
the photochemical properties of the material. The calcination temperatures adopted
were 500, 600, and 700 °C, and the obtained materials were coded as T500, T600,
and T700, respectively, besides the commercial one, coded as TC.
The materials were characterized to identify the crystalline phases of the
compound, applying the X-ray diffraction (XRD) technique by the powder method.
Using Panalytical diffractometer and Empyrean model, in which the samples were
scanned in the range of 2θ from 15° to 85°, with copper anode tube, applying elec-
tromagnetic waves wavelength, λ = 1.5418 Å, with a voltage of 45 kV acceleration
and 40 mA applied current.

3 Results

3.1 Characterization by X-Ray Diffraction

All TiO2 samples obtained with different heat treatments at different temperatures
were characterized by X-ray powder diffraction to identify and quantify the possible
phases present in the obtained material. For TiO2 , the possible phases present are
the rutile and anatase in the sample; using the X-ray diffraction technique, it was
possible to identify the crystalline phases present.
380 T. S. Lima et al.

Fig. 1 X-ray diffractogram of TiO2 synthesized by the Pecchini method with heat treatment
at 500 °C. Analysis performed with scanning range between 15° and 85°, with copper anode
tube, applying electromagnetic waves wavelength, λ = 154.18 pm. Diffractogram calculated
by the Rietveld method, experimental diffractogram and the difference between calculated and
experimental diffractograms

Table 1 Quantitative results from characterization by X-ray diffraction

Phase TC T500 T600 T700
Anatase 99.42 42.43 0.56 0.04
Rutile 0.58 57.57 99.44 99.96

To identify and quantify these phases, the Rietveld method was applied, which
uses the result of X-ray diffraction to quantify each phase. Figure 1 shows the result
of applying the method with the X-ray diffractogram calculated by the Rietveld
method, the experimental diffractogram and the calculated and experimental differ-
ence for the treated sample at a temperature of 500 °C. In Fig. 1, we can observe an
excellent agreement between the calculated result and the experimental result, with
all the diffractions, identified, indicating the presence of only the two phases before
proposed.
Using the same methodology for analyzing diffractograms, the sample TC
presents the crystalline anatase phase in greater quantity, for the T500 sample there
was a mixture of phases between anatase and rutile, and for the samples, T600 and
T700 were observed to be the rutile phase and predominant as shown in Table 1.

4 Conclusion

From the X-ray diffraction analysis, it is possible to evaluate the TiO2 crystalline
phases present in the samples synthesized at different temperatures and identified
X-Ray Characterization of TiO2 Obtained … 381

as T500, T600, and T700. For the T500 sample, there is a mixture of anatase and
rutile phases, and for the T600 and T700 samples, it was observed that the rutile
phase and the predominant phase. With this, we can say that the calcination process
influenced the polymorphic composition of the material, increasing the rutile phase
with increasing temperature, which brings undesirable characteristics for application
in photocatalytic processes.

Acknowledgements This work was partly supported by Grants 2018/17913-2 and 2018/05698-0,
São Paulo Research Foundation (FAPESP).

References

1. Iga, G. D., de Castro Peixoto, A. L., Costalonga, A. G. C.: Optimization of photocatalytic

degradation of methyl orange using TiO2 . In: Iano, Y., Arthur, R., Saotome, O., Vieira Estrela,
V., Loschi, H. (eds.) Proceedings of the 3rd Brazilian Technology Symposium. BTSym 2017.
Springer, Cham (2019)
2. Iga, G. D., de Castro Peixoto, A. L., Costalonga, A. G. C. (2019) Efficient optimization of
photocatalytic process for Azo Dyes. In: Iano, Y., Arthur, R., Saotome, O., Vieira Estrela, V.,
Loschi, H. (eds.) Proceedings of the 4th Brazilian Technology Symposium (BTSym’18). BTSym
2018. Smart Innovation, Systems and Technologies, vol. 140. Springer, Cham (2018)
3. Kanakaraju, D., Glass, B., Oelgemöller, M.: Titanium dioxide photocatalysis for pharmaceutical
wastewater treatment. Environ. Chem. Lett. 12(1), 27–47 (2014)
4. SANTOS: Lidiaine Maria dos. Síntese e caracterização de TiO2 com modificações superficiais
para aplicação em fotocatálise heterogênea. 2017. 135 f. Tese (Doutorado em Química) - Univer-
sidade Federal de Uberlândia, Uberlândia, Disponível em (2017). http://dx.doi.org/10.14393/
ufu.te.2017.48
5. Araujo, dos Santos, D., et al.: Avaliação gap óptico do TiO2 obtido pelo método Pechini:
influência da variação das fases anatásio-rutilo. Matéria (Rio J.), Rio de Janeiro 23(1), e-11949
(2018)
6. Brito, S. L. M.: Síntese, caracterização e modificação superficial de nano partículas de titanato
de bário produzidas pelo método Pechini. Tese de Doutorado, Escola Politécnica. Universidade
de São Paulo, São Paulo (2009). https://doi.org/10.11606/t.3.2009.tde-28052009-141416
7. Costa, A.C.F.M., et al.: Síntese e caracterização de nanopartículas de TiO2. Cerâmica, São Paulo
52(324), 255–259 (2006)
8. Mutuma, B.K., Shao, G.N., Kim, W.D., Kim, H.T.: Sol-gel synthesis of mesoporous anatase-
brookite and anatase-brookite-rutile TiO2 nanoparticles and their photocatalytic properties. J.
Colloid Interface Sci. 442, 1–7 (2015)
Co-authors Networks in Adsorption
Refrigeration and Air-Conditioning
with Solar Energy

Carla Abregú Marcos , Abraham Sopla Maslucán ,

Miguel Cano Lengua , and José C. Alvarez

Abstract The refrigeration and air conditioning systems by adsorption with solar
energy are a very interesting option to replace the conventional compression systems
because they save energy, are noiseless and use non-aggressive ozone layer refriger-
ants; however, there are scarcities of studies at this topic. The acquisition of knowl-
edge and positioning in research on this topic by new research groups could be facil-
itated by knowing the context, trends, and collaborations that emerge in the subject.
A map of collaborative co-authors between the greatest authors on the subject was
drawn up, the same one that is done using Gephi software. From the analysis of these
maps, the centrality degree and collaboration between the authors were determined.

Keywords Refrigeration · Air-conditioning · Adsorption · Solar energy ·

Co-authoring

1 Introduction

Adsorption cooling systems, which use a refrigerant/adsorbent pair, are an ecological

alternative to conventional refrigeration systems because they do not use refrigerants
that damage the ozone layer and use recovered heat or solar thermal energy. Dieng and
Wang [1], through a literature review, pointed out the importance of the applicability

C. A. Marcos · A. S. Maslucán · M. C. Lengua

Universidad Nacional Mayor de San Marcos, Lima, Peru
e-mail: [email protected]
A. S. Maslucán
e-mail: [email protected]
M. C. Lengua
e-mail: [email protected]
J. C. Alvarez (B)
Universidad Peruana de Ciencias Aplicadas, Lima, Peru
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 383
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_38
384 C. A. Marcos et al.

of solar energy in both air conditioning and refrigeration by adsorption, for the reason
of performance coefficient improvement, air conditioning and ice making are quiet,
non-corrosive and environmentally friendly.
According to Hassan and Mohamad [2], refrigeration and air conditioning
processes are considered essential needs for all human beings in today’s world,
but it is necessary providing refrigeration using green energy such as solar energy
as the key solution to electricity and pollution problems that traditional refrigeration
systems present.
Therefore, new research is necessary to improve efficiency, through development,
test and simulate working prototypes. In this sense, using technology surveillance,
business intelligence, and data mining, particularly co-authorship it is possible to
search and identify new research topics.
Wang et al. [3] performed a co-authorship analysis among Chinese researchers in
the field of management, in the period 2006–2015, finding a low network density that
means a weak collaboration between authors on this topic, and the opportunity for
new collaborations. These authors [4] also encountered that: similar work, academic
authority, and geographical closeness, helping to increase the collaboration network.
Yeo and Lewis [4], through a qualitative case study of the process, the benefits, the
motivations, and the difficulties of co-authorship in the field of linguistics, these
authors [4] concluded that the co-authorship generates benefits for both co-authors
and it is a trend.
Zang [5] makes a quantitative analysis of the network of 511 co-authors of
Paul Erdö, using Grey relationship analysis, for which previously, based on the
UCINET software analysis network, this author [5] evaluate five measures of the
significant influence of co-authors: degree, betweenness, closeness, eigenvector, and
publications.
Thus, there is a great possibility, applying co-authorship qualitative or quantita-
tive analysis, to determinate relations between authors and until the form that the
knowledge is constructed by the researchers. The contribution of this work is the
determination of the research co-authors groups around the topic adsorption refrig-
eration and air-conditioning with solar energy, and the measure of the centrality
degree.

2 Methodology

Firstly, the keywords are defined: “refrigeration” and “air conditioning” and “adsorp-
tion” and “solar energy”. After that, a search was made in the Science Direct database
[6] using search strings. Identifying the authors with more publications in the period
2000–2019.
Co-Authors Networks in Adsorption … 385

2.1 Search for Co-authors and Publications

Technological surveillance was carried out based on authors and journals for the
following chain: “refrigeration” and “air conditioning” and “adsorption” and “solar
energy”, for this purpose we searched the Science Direct Database, obtained 17
papers from 2001 to 2019, as appear in Table 1.
Table 2 was also implemented with co-authors who had more publications from
2000 to 2019.

Table 1 Co-authors of the papers

Keywords Papers Main co-authors
“Refrigeration” and “air conditioning” and “adsorption” and “solar 17 [1, 2, 7–21]
energy

Table 2 Authors with more publications in that period

Keywords Papers Main authors Source title No
“Refrigeration” and “air 17 R. Z. Wang Renewable and 2
conditioning” and “adsorption” sustainable energy
and “solar energy reviews
Energy conversion and 2
management
Applied energy 1
Progress in energy and 1
combustion science
Energy and buildings 1
Applied thermal 1
engineering
Renewable and 1
sustainable energy
reviews
T. S. Ge Energy conversion and 1
management
Energy 2
Y. J. Dai Renewable and 1
sustainable energy
reviews
Energy conversion and 1
management
Applied energy 1
386 C. A. Marcos et al.

Table 3 A representative part of the co-authoring matrix

F. Xu Z. F. Bian T. S. Ge Y. J. Dai
F. Xu 0 1 1 1
Z. F. Bian 0 0 0 0
T. S. Ge 0 0 0 1
Y. J. Dai 0 0 0 0

2.2 Construction of Author-Networks

The networks between the authors are mapped using Gephi software to determine
the groups of authors that have generated the greatest collaboration in the study of
the subject. The nodes of each group are analyzed separately looking for determining
the co-authorly inside each group.

3 Results

3.1 Relations Between Authors

Firstly, the collaboration network between authors was determined using the Gephi
network analysis tool. For this, the co-authoring matrix was generated (Table 3); in
that matrix, the names of the authors are written in a column, these same authors
are written in a header row, then for 17 articles is identified the co-authoring “1”
or not co-authoring “0”. The network allows determining which authors have more
co-authoring, as well as the authors who have less co-authoring. This determination
is through the observation of links quantity. In Fig. 1, the results obtained were
evidenced.
In Fig. 1 the most prominent authors through the size of the tags represent greater
co-authorship. Next, the modularity of the graph is established by means of the Gephi
modularity command, obtaining Fig. 2.
The modularity of the networks is a measure of the structure of the networks or
graphs, so high-modularity networks have strong connections inside the groups but
weak between nodes of different groups.
In Fig. 2 the labels are removed, and a 90° rotation is made, thus obtaining a new
distribution of the nodes in the network. In Fig. 3 the subgraph with the largest number
of nodes is separated, that is, the subgroup with the highest number of authors.
Co-Authors Networks in Adsorption … 387

Fig. 1 Network of authors

Fig. 2 Group of authors

388 C. A. Marcos et al.

Fig. 3 Extraction of the

main node

4 Analysis of the Results

According to Fig. 1, the font size of the author’s name indicates that the author has
more co-authorship. In this case R. Z. Wang has most co-authors, followed by Zhai,
Fong, Choudhury, Sarbu, Ma, Xu, Ge, Lu and Day.
In Fig. 2, eight nodes are identified (R. Z. Wang, I. Sarbu, R. Z. Wang, K. F. Fong,
K. Sumathy, H. Z. Hassan, B. Choudhury, and N. Qadir as main authors), of which
the nodes of R. Z. Wang are almost at center the image and are more scattered than
the authors of the other node of this same author on the left of Fig. 2, which means
that these authors are conducting further research.
It is also evident that four nodes (K. F. Fong, K. Sumathy, N. Qadir and H. Z.
Hassan) haven´t bridge with any another group, indicating that no author of these sub-
groups collaborates outside their group; however, a bridge is observed in common
with two nodes (both of R. Z. Wang). In the largest node of R. Z. Wang, it’s also
seen that five main authors are Song, Lu, Zhai, Ge, and Xu, having all of them as the
principal author of R. Z. Wang. Also is observed that the nodes of Choudhury and
Sarbu have a tenuous connection by Chatterjee author.
By the dispersion of nodes in the network of Fig. 2, it is evidenced that it is a
multigraph with eight groups, which also is visualized that in the central part the
nodes are more connected these are two groups (both of R. Z. Wang, one minor and
one major). In Fig. 3, the connectivity inside the major group of R. Z. Wang, which is
linked to the minor group of R. Z. Wang evidenced the preponderance of this author.
The centrality degree, that is measured by the number of connections of each node
with another node, was determined as 7 for R. Z. Wang, 5 for K. F. Fong, 4 for X.
Q. Zhai and 3 for B. Choudhry and I. Sarbu.
Co-Authors Networks in Adsorption … 389

5 Conclusions

In the matter of refrigeration and air conditioning by adsorption with solar energy, it
is found that the research from 2001 to 2019 did not provide many bibliographical
references, but it was possible to identify seventeen documents that were analyzed by
Gephi software, the result was that the main author of this topic is R. Z. Wang, who
is referenced as the main author in two nodes, having as major co-authors Song, Lu,
Zhai, Ge, and Xu. However, there are other authors such as Fong, Qadir, Sumathy,
and Hassan who do not have any bridge in common but are still investigating.
Inside these identified research nodes and their respective links, knowledge is
codified participatory and the understanding of this construction results important to
enter and participate in these knowledge spaces.
The challenge is to continue with these measurements of software tools in much
more numerous environments and in contexts in which social networks are part of
the daily life of each author, comparing performance and results according to the
case study proposed.
New researches about the publication’s contents of determined groups (node and
links) with tools as Lingua Kit, for a revision and classification of contents without
a previously reading, will be the continuity of this research.

References

1. Dieng, A.O., Wang, R.Z.: Literature review on solar adsorption technologies for ice-making
and air-conditioning purposes and recent developments in solar technology. Renew. Sustain.
Energy Rev. 5(4), 313–342 (2001). https://doi.org/10.1016/S1364-0321(01)00004-1
2. Hassan, H.Z., Mohamad, A.A.: A review on solar-powered closed physisorption cooling
systems. Renew. Sustain. Energy Rev. 16(5), 2516–2538 (2012). https://doi.org/10.1016/j.rser.
2012.02.068
3. Wang, C., Cheng, Z., Huang, Z.: Analysis on the co-authoring in the field of management in
china: based on social network analysis. Int. J. Emerg. Technol. Learn. 12(6) (2017). https://
doi.org/10.3991/ijet.v12i06.7091
4. Yeo, M., Lewis, M.: Co-Authoring in action: practice, problems and possibilities. Iran. J. Lang.
Teach. Res. 7(3), 109–123 (2019)
5. Zang, X.L.: The data mining based on the co-author network. Adv. Mater. Res. 926–930,
3402–3405 (2014). https://doi.org/10.4028/www.scientific.net/AMR.926-930.3402
6. Science Direct Database. https://www.sciencedirect.com/
7. Choudhury, B., Chatterjee, P.K., Sarkar, J.P.: Review paper on solar-powered air-conditioning
through adsorption route. Renew. Sustain. Energy Rev. 14(8), 2189–2195 (2010). https://doi.
org/10.1016/j.rser.2010.03.025
8. Xu, F., Bian, Z.F., Ge, T.S., Daia, Y.J., Wang, C.H., Kawi, S.: Analysis on solar energy powered
cooling system based on desiccant coated heat exchanger using metal-organic framework.
Energy 177, 211–221 (2019). https://doi.org/10.1016/j.energy.2019.04.090
9. Sarbu, I., Sebarchievici, C.: General review of solar-powered closed sorption refrigeration
systems. Energy Convers. Manag. 105, 403–422 (2015). https://doi.org/10.1016/j.enconman.
2015.07.084
390 C. A. Marcos et al.

10. Fong, K.F., Lee, C.K., Chow, T.T., Lin, Z., Chan, L.S.: Solar hybrid air-conditioning system
for high temperature cooling in subtropical city. Renew. Energy 35(11), 2439–2451 (2010).
https://doi.org/10.1016/j.renene.2010.02.024
11. Fong, K.F., Lee, C.K., Chow, T.T., Lin, Z., Chan, L.S.: Simulation-optimization of solar—
thermal refrigeration systems for office use in subtropical Hong Kong. Energy 36(11), 6298–
6307 (2011). https://doi.org/10.1016/j.energy.2011.10.002
12. Zhai, X.Q., Wang, R.Z., Wu, J.Y., Dai, Y.J., Ma, Q.: Design and performance of a solar-powered
air-conditioning system in a green building. Appl. Energy 85(5), 297–311 (2008). https://doi.
org/10.1016/j.apenergy.2007.07.016
13. Ul Qadir, N., Said, S. A. M., Ben Mansour, R.: Modeling the performance of a two-bed solar
adsorption chiller using a multi-walled carbon nanotube/MIL-100 (Fe) composite adsorbent.
Renew. Energy 109(C), 602–612 (2017). https://doi.org/10.1016/j.renene.2017.03.077
14. Choudhury, B., Saha, B.B., Chatterjee, P.K., Sarkar, J.P.: An overview of developments in
adsorption refrigeration systems towards a sustainable way of cooling. Appl. Energy 104,
554–567 (2013). https://doi.org/10.1016/j.apenergy.2012.11.042
15. Sumathy, K., Yeung, K.H., Yong, L.: Technology development in the solar adsorption refrig-
eration systems. Prog. Energy Combust. Sci. 29(4), 301–327 (2003). https://doi.org/10.1016/
S0360-1285(03)00028-5
16. Lu, Y.Z., Wang, R.Z., Zhang, M., Jiangzhou, S.: Adsorption cold storage system with zeolite–
water working pair used for locomotive air conditioning. Energy Convers. Manag. 44(10),
1733–1743 (2003). https://doi.org/10.1016/S0196-8904(02)00169-3
17. Ma, Q., Wang, R.Z., Dai, Y.J., Zhai, X.Q.: Performance analysis on an air-conditioning system
of a green building. Energy Build. 38(5), 447–453 (2006). https://doi.org/10.1016/j.enbuild.
2005.08.004
18. Ge, T.S., Day, Y.J., Wang, R.Z.: Review on powered rotary desiccant cooling system. Renew.
Sustain. Energy Rev. 39, 476–497 (2014). https://doi.org/10.1016/j.rser.2014.07.121
19. Song, F.P., Gong, L.X., Wang, L.W., Wang, R.Z.: Study on gradient thermal driven adsorption
cycle with freezing and cooling output for food storage. Appl. Therm. Eng. 70(1), 231–239
(2014). https://doi.org/10.1016/j.applthermaleng.2014.04.066
20. Wang, R.Z., Oliveira, R.G.: Adsorption refrigeration - An efficient way to make good use of
waste heat and solar energy. Prog. Energy Combust. Sci. 32(4), 424–458 (2006). https://doi.
org/10.1016/j.pecs.2006.01.002
21. Ge, T.S., Dai, Y.J., Wang, R.Z.: Performance study of silica gel coated fin-tube heat exchanger
cooling system based on a developed mathematical model. Energy Convers. Manag. 52(6),
2329–2338 (2011). https://doi.org/10.1016/j.enconman.2010.12.047
Microsimulation of Public Transport
Stops for the Optimization of Waiting
Times for Users Using the Social Force
Model

Francis Mendoza , Mayling Tong , Manuel Silvera ,

and Fernando Campos

Abstract Cities in the world aim to ensure the mobility of people, through the
implementation of efficient Integrated Transportation Systems (ITS). This aims to
improve the transport of people, which guarantees that they can be mobilized safely
and without delays in the terminals and bus stops of the public transport system. The
present article proposes a design of public transport stops aimed at optimizing the
waiting time of users when transferring from one bus to another. For the validity
of the proposal, the social force model of the Vissim program was used, where the
behavior of the users within the bus stops was reflected. The results showed that the
waiting times in the calibrated and validated microsimulation model were optimized
by approximately 20%, which generates an improvement in the efficiency of the
public transport system.

Keywords Bus stops · Transshipment · Public transport · Social force model

1 Introduction

Lima is a city composed of 43 districts and a population of approximately 9.32 million

inhabitants, most of which are mobilized through the public transport system. Within
this large number of users who have multiple trips, there are those who travel for work

F. Mendoza (B) · M. Tong · M. Silvera · F. Campos

Universidad Peruana de Ciencias Aplicadas, Av. Prolongación Primavera 2390, Santiago de
Surco, Lima, Peru
e-mail: [email protected]
M. Tong
e-mail: [email protected]
M. Silvera
e-mail: [email protected]
F. Campos
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 391
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_39
392 F. Mendoza et al.

and study reasons. These represent the largest number of trips in the public transport
system, especially during peak hours [1]. The municipality of Lima, in an attempt to
provide a service that facilitates interurban transport, has implemented an Integrated
Transportation System (ITS) which has failed to meet the needs of intermodal users.
Taking these limitations into account, people have spontaneously identified, between
the different public transport lines, interconnections where transshipment is made for
the purpose of reducing travel time. These intersections can be identified within the
city of Lima as specific points where a large number of these intermodal operations
are carried out. An example where this type of activity is presented is the intersections
located on main avenues. However, in general, these intersections present problems
in the geometric design that prevent detecting the transshipment of people in an
optimal and safe way, thus denying conflicts in the vehicle–turbo interaction and
increases in the waiting time in the users. Considering that this problem is common
in the entire public transport system of the city of Lima, the present research selects
as an area of study an intersection located in a commercial area that has bus stops
where this type of interconnections is located. This intersection will be simulated in
the program Vissim1 that uses the social force model2 with the aim to optimize the
current transshipment times seeking the efficiency of the public transport system.

2 State of the Art

The study of “urban transshipment” derived from the term “modal exchange,” has
been successfully developed in several cities of the world such as Melbourne, Toronto,
London, Sao Paulo, New Jersey, among others, as a clear concept used for the trans-
port system management [2]. The main objective of these studies is to design an
intermodal bus stop that guarantees the integrated and efficient transfer of passen-
gers between routes and different modes of transport [3]. In this sense, the bus stops
infrastructure must be designed taking into account the distribution of spaces to facil-
itate the route of the pedestrians, thus providing easy access to two or more modes
of transport at the same time [4].
On the other hand, the safety of people during transshipment is one of the main
elements in the design of an efficient intermodal bus stop. The security takes into
account the protection of passengers against climatic conditions, the analysis of
passenger travel paths, and the operating conditions of different modes of transport
[5]. In addition, the waiting time and the user spacing when walking are considered
when assessing the comfort level of a modal interchange stop [6]. In this way, the
satisfaction and stress levels of the users are measured.
There is little research related to the use of microsimulation models to measure
the residence time of buses at a bus stop. This is due to a large amount of data to be
collected and the high costs that are generated in this kind of study. It is for this reason

1 PTV Vissim—Viswalk.
2 Helbing and Molnár [9].
Microsimulation of Public Transport Stops … 393

that existing investigations take into account for the calculation of transshipment time
only small field samples where a single route is analyzed, in order to identify the
reasons for delays in the arrival of buses [7].
With regard to the Social Force Model3 (SFM), there are several studies that apply
it from different points of view. For example, extensions of the model have been made
to analyze how the increase in the walking speed of some pedestrians generates an
aggressive interaction in pedestrian behavior [8]. On the other hand, SFM allows
the modeling of attractive and repulsive forces, to simulate the concentration and
dispersion of pedestrians in specific places within the environment, such as artistic
attractions and places where people who share social bonds gather [9]. In addition,
SFM is also used to simulate people waiting for the arrival of a vehicle or bus.
These waiting times influence the level of service of intermodal stations because
they generate various paths that represent the pedestrian flow. Thus, Johansson et al.
[10]. conclude that the analysis of people in the intermodal bus stop waiting for a
vehicle, interacting with pedestrians walking through the same area, has a significant
impact on the predictions provided by pedestrian microsimulation models.

3 Methodology

For the development of the new design of the intersection looking for the optimization
of the waiting times of the buses, first, the information survey of the study area was
carried out. Here, data related to intersection geometry, vehicular and pedestrian
volumes, and pedestrian travel times were collected. Then, the different trajectories
and maneuvers of public and private transport vehicles were identified.
Taking into account the scope of the investigation, four main vehicular intercon-
nection patterns present at the intersection were considered within the analysis. With
this information, the different routes and trajectories described by the users of the
public transport system were plotted as shown in Fig. 1. In this way, it can be seen
that each transshipment pattern has two different routes at most. These are the most
used by the people present in the study area.
One of the particular problems of the study area was the identification of two trans-
shipment routes described in unofficial pedestrian crossings, which puts the safety
of pedestrians at risk. Another of the difficulties encountered in the transshipment is
in the pedestrian–pedestrian interaction due to their density in zebra-type pedestrian
cruises.
394 F. Mendoza et al.

Fig. 1 Routes used by people for transshipment at the intersection studied

3.1 Technical Considerations

More than two transshipment paths within the study area overlap each other and,
as a consequence, conflicts between pedestrians paths are generated. Moreover, the
current state of the study area exposes the users to danger while they make an unoffi-
cial crossing between refuges islands. In order to increase the safety of the users and
to reduce the necessary time to make a complete transshipment, some modifications
aimed to facilitate the flow of pedestrians were proposed:

3 Helbing and Molnár [9].

Microsimulation of Public Transport Stops … 395

Fig. 2 Pedestrian crossing design proposed in the National Roads Authority (NRA)

Fig. 3 Design proposal for pedestrian cruises within the intersection studied

1. Two-way pedestrian cruises to facilitate the flow of pedestrians by normalizing

the direction of the crossings and channeling the flows as shown in Fig. 2. In
order to perform the simulation of this new pedestrian behavior, it was necessary
to iterate the social force parameters (Tau, Lambda, and Noise);
2. Considering the different transshipment routes generated by pedestrians, new
pedestrian crossings were generated as shown in Fig. 3, which are made between
the four refuge islands within the intersection. This facilitated the implementation
of two transshipment routes (transshipment 2 and 3);
3. The Tau parameter was iteratively decreased until it reached a value that guar-
antees the optimum use of the spaces designated for pedestrians. Particularly
396 F. Mendoza et al.

the calibration of this parameter in the corners of the intersection was consid-
ered, where the turning radius of the vehicles had to be reduced to increase the
acceleration of the pedestrians.
4. The Lambda parameter was reduced due to the volume of pedestrians that make
the transshipment. As this number of users is high, they are not very influenced
by the people present in their environment; therefore, they can move together
optimizing the space designated for pedestrians within the intersection.
5. The noise parameter of randomness was modified to ensure the optimum speed
of pedestrians at the intersection. In this way, an efficient walking path of the
pedestrian flow is guaranteed when the transshipment is made.

4 Results

When evaluating the behavior of the system with the modifications proposed at the
intersection, reductions were obtained in the four transshipment times.
In the current configuration, the average time for transshipment 1 ranges between
90.92 and 99.53 s. However, when our proposal was implemented regards the simu-
lation scenario, this average time decays to a range of 72.81–80.25 s as shown in
Fig. 4.
In the case of transshipment 2, it can be seen that for the current state, the waiting
times range between 120.12 and 142.23 s. This particular transshipment shows quite
large data ranges compared to the other three transshipments, reaching up to 22.11 s
in amplitude. However, through implementation, the average time could be reduced
by 21.1% as shown in Fig. 5.
In the case of transshipment 3, the average times range between 134.56 and
147.82 s. These values were reduced to a new range of 107.61 and 116.54 s. In this
case, the difference in amplitudes is 4.33 s and the average transshipment time for
the 30 runs was reduced by 20.3% as shown in Fig. 6.

Fig. 4 Comparative chart of current delays versus delays generated with the proposal for
transshipment 1
Microsimulation of Public Transport Stops … 397

Fig. 5 Comparative chart of current delays versus delays generated with the proposal for
transshipment 2

Fig. 6 Comparative chart of current delays versus delays generated with the proposal for
transshipment 3

In the case of transshipment 4 shows average time values between 46.35 and
49.35 s, which were optimized to achieve a new range with limits of 35.47 and
41.05 s. In this transshipment, the amplitude of the results of the proposal is greater
than those of the current state, however, a decrease in the average transshipment time
of 20.2% was obtained, as shown in Fig. 7.
Finally, Table 1 summarized the results obtained in this research highlighting the
difference regards the average time for each transshipment.
398 F. Mendoza et al.

Fig. 7 Comparative chart of current delays versus delays generated with the proposal for
transshipment 4

Table 1 Evaluation of maximum, minimum, and average transshipment times between the current
situation (field data) of the intersection and our proposal
Transshipment 1 Transshipment 2 Transshipment 3 Transshipment 4
Current Proposal Current Proposal Current Proposal Current Proposal
Min (s) 90.92 72.78 120.12 96.07 134.56 107.61 46.35 35.47
Max (s) 99.53 80.25 142.23 112.61 147.82 116.54 49.35 41.05
Average 95.91 76.32 131.84 104.00 141.19 112.57 48.34 38.56
(s)
Dif % 20.42 21.09 20.26 20.24
average

5 Conclusions

It is of high importance the preferences that pedestrians on some routes for transship-
ment. This helped our research to optimize the transshipment time and to facilitate
decision making from the pedestrians. With the new design and redistribution of the
spaces designated for pedestrians and vehicles, the time of each transshipment was
approximately reduced by 20% compared to the times measured in the field. This
result shows that studies involved in the optimal use of the current infrastructure
are desired. As future work, the existing circumstances created by the users will be
studied in order to enhance the current system considering other scenarios.
Microsimulation of Public Transport Stops … 399

References

1. Vamos, L.C.: IX Informe de percepción sobre calidad de vida en Lima y Callao. PUCP, Lima
(2018)
2. Durán Bernal, L.M.: Basic parameters for the design of intermodal public. Transport. Res.
Procedia 14, 499–508 (2016)
3. Pitsiava-Latinopoulou, M., Iordanopoulos, P.: Intermodal passengers terminals: design stan-
dards for better level of service. Procedia Soc. Behav. Sci. 48, 3297–3306 (2012)
4. Steinsiek, D.: On the Way to an Intermodal Transportation System-Improving the Cooperation
in the Public Transport Sector of the Stadsregio, Utrecht (2014)
5. Litman, T.: Introduction of Multi-Modal Transportation Planning. Victoria Transport Policy
Institute, Victoria (2009)
6. Oostendorp, R., Gebhardt, L.: Combining means of transport as a users’ strategy to optimize
traveling in an urban context: empirical results on intermodal travel behavior from a survey in
Berlin. J. Transp. Geogr. 71, 72–83 (2018)
7. Milkovits, M.N.: Modeling the factors affecting bus stop dwell time: use of automatic passenger
counting, automatic fare counting, and automatic vehicle location data. Transp. Res. Rec. 2072,
125–130 (2008)
8. Parisi, D., Dorso, C.: Microscopic dynamics of pedestrian evacuation. Phys. A 354, 606–618
(2005)
9. Helbing, D., Molnár, P.: Social force model for pedestrian dynamics. Phys. Rev. 51, 4282–4286
(1995)
10. Johansson, F.: Pedestrian Traffic Simulation Platform. Swedish National Road and Transport
Research Institute (2013)
Mathematical Model
for the Improvement of Vertical Traffic
at Universities

Julio C. Lazo Lozano and Grimaldo Wilfredo Quispe Santivañez

Abstract This article seeks to develop a methodology to reduce the travel time of
the elevators and also to increase the number of users served by the elevators of
the university, according to the schedule, said time saved achieves an improvement
of the elevator service. The methodology based on a mathematical algorithm will
determine that by changing the configuration of the elevators in attention capacity
per floor and the round and attention times they establish a better travel time of the
elevators. It is concluded that by applying the methodology developed in this article,
a reduction of up to 37.5% of the travel time of the elevators is achieved, as well as
an increase of the users attended by the elevators, being the increase of 10.66%.

Keywords Elevators · Mathematical model UpPeak · Vertical transport · Timing

1 Introduction

Multistory buildings have faced vertical elevator traffic problems during the past few
years. Even though elevators solve transportation issues in high-rise buildings, they
can also cause problems. For example, if an elevator is not correctly programmed for
a building several stories high and with many different types of inhabitants (workers,
students, etc.), it will probably experience an excess of demand and therefore fail to
properly accomplish its task.

J. C. Lazo Lozano
Universidad Tecnológica del Perú, Lima, Peru
e-mail: [email protected]
G. W. Quispe Santivañez (B)
Universidad Nacional San Marcos, Lima, Peru
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 401
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_40
402 J. C. Lazo Lozano and G. W. Quispe Santivañez

Vertical elevator traffic is not properly managed at universities. Although there

have been several changes to the existing elevator stopping points at the university,
these changes only provide a temporary solution to the problem. Assessments such
as the waiting time approach allow recording the times and rhythms of the work
corresponding to the elements of a given activity, either a specific or a general task,
with the aim to measure exactly the entire workflow [1]. The queues that are formed in
a vertical traffic system have irregular sizes [2] because many times this type of system
is designed according to the criteria and experience of the designer without objective
parameters [3], and cannot be maximized the service provided [4]. The problem
of vertical traffic is centered on buildings for work purposes, which have a large
number of floors. Traffic patterns can be divided into five different traffic patterns,
which are: maximum ascending traffic, maximum descending traffic, noon traffic,
traffic with balanced plans, and traffic between unbalanced plans [5]. To carry out
an adequate study [6], the simulation will compare different designs [7], which will
be analyzed in discrete events [8], which are generated in the vertical upward traffic
pattern, because preliminary evidence shows that demand of the users comes mainly
from the main floor to the other floors of the building and are the busiest hours of
traffic In this context, this study assesses the possibility of an optimal selection of the
elevator stop points to avoid vertical traffic. Consequently, an UpPeak mathematical
model and a study of times in the elevators of the Technological University of Peru
(UTP) were developed and validated.

2 Mathematical Model for Upward Traffic (UpPeak)

The items below are the most relevant for the mathematical vertical traffic model.
They will be listed with their corresponding mathematical formulas:
1. The average number of passengers departing from the main floor:

80
P= CC (1)
100
CC maximum capacity of each car,
P average number of passengers per second departing from the main floor.
2. The average number of stops for a single elevator in each trip:

  P 
N −1
S = N 1− (2)
N

N total number of floors served by the elevator without counting the main floor.
Mathematical Model for the Improvement … 403

3. The average number of floors served at higher altitude:

N −1
 
i
H=N− P (3)
i−1
N

4. The Average time per cycle, RTT :

RTT = Tr + Td + T p + Tl (4)

Tl = 0.1 ∗ Td ∗ TP (5)

Td = td ∗ S (6)

TP = Tpo + Tpi (7)

Tpi = P (8)

1
Tpo = k ∗ P ∗ f 3 (9)

Tl Calculation of time lost while the elevator is stopped

Td Calculation of time when the door opens until it closes.
Tr = tLigth Time taken by a stopped elevator to reach the next adjacent floor at
nominal speed.
TP Calculation of passenger loading and unloading times.
5. Meantime of arrival at the ground floor for each elevator:

RTT
INT = (10)
L
L The total number of elevators in the group was considered.
6. System solvency capacity: It is defined as the number of passengers, measured
as a percentage of the total population of the building, that the system is capable
of transporting to its destination in 5 min (300 s) at maximum demand

300
INT = PL (11)
RTT
404 J. C. Lazo Lozano and G. W. Quispe Santivañez

PL Number of people measured as a percentage of the total population of the

building in 300 s.
7. Average waiting time per passenger:

 2 
P
AWT = 0.4 + 1.8 0.77 INT (12)
CC

8. Average elevator trip time:

ATT = 0.5 INT + 0.5 RTT (13)

9. Average journey time:

AJT = AWT + ATT. (14)

3 Methodology

First, we assessed the current UTP elevator service by conducting a user survey to
better understand the degree of user dissatisfaction. Once the current service was
assessed, a time study was performed to measure the time the elevator takes to travel
to each floor, the time the elevator takes to go from the main floor to the last floor,
the time the elevator takes to respond to a call from the main floor (Floor No. 01),
and passenger waiting times from the moment they call the elevator to their arrival
at their destination.
After the necessary information was gathered, an analysis is carried out [9], for the
introduction of data to the software (Arena), once the current situation was simulated
to observe the number of users served in more detail. Finally, we proposed a new
distribution of the elevator stop point to increase the number of users served while
traveling times are reduced, and the speed of service provided is increased [10].

4 Design

For the purposes hereof, the population is represented by the total number of
users, approximately 15,160, during the night shift elevator service from 18:30 until
Mathematical Model for the Improvement … 405

20:00 pm). The sample will be determined by a statistical calculation with a 95%
confidence interval and an estimated error of 5%, which gives us 375 users.

4.1 Time Study

A time study was performed for all elevators, with the exception of the elevator that
must stop on all floors according to regulation A.120, Sec.4 of the Accessibility
for Disabled and Elderly People Act, which establishes that elevators must be set
up to allow disabled people to move freely throughout the building. Therefore, this
elevator, which is required by law to make a stop at every floor, was not taken into
consideration.
The study began by taking the travel time from the main floor (Floor 1) to different
floors and the number of users exiting the elevator on these floors.

4.2 Simulations

The following data were required to prepare the simulation [11].

• Passenger arrival time: 6 s
• Travel time to each floor
• Number of passengers that fit inside the elevator
• Number of passengers exiting the elevator on each floor.
The simulation timeframe was from 6:30 to 7:00 pm (30 min), as the highest
demand occurs within this period. Passengers arrive every 6 s, pass through a counter,
and then make a queue to use the elevator. The elevator departs once it is loaded with
17 passengers, which represents the mean capacity of the elevator. Two passengers
exit on the 4th floor, six passengers exit on the 8th floor, eight passengers exit on the
12th floor, and one passenger exits on the 15th floor. Similarly, the time taken by the
elevator to go up to the fourth floor and unload passengers is 12.91 s, from the 4th
to the 8th floor is 24.64 s, from the 8th to the 12th floor is 29.14 s, from the 12th
to the 15th floor is 26.48 s, from the 1st to the 15th floor is 88.97 s, from the 1st to
the 12th floor is 60 s, and from the 1st to the 8th floor is 38.81 s. Therefore, current
traffic flow at the Arequipa tower elevator service is as follows:
In the first simulation performed in Hall A, a total of 229 passengers were served
from 6:30 to 7:00 pm. As two elevators are being used, we may consider 458
passengers served in total.
In the second simulation conducted in Hall B, a total of 169 passengers were
served from 6:30 to 7:00 pm. As two elevators are being used, the total would be 338
considering both.
406 J. C. Lazo Lozano and G. W. Quispe Santivañez

4.3 Mathematical Model Application

The mathematical model was applied for data validation. A small margin of error
can be found between the field study and the mathematical model.
Table 1 presents the data obtained by applying the mathematical model, it can be
concluded that current elevator passenger travel times can be reduced.
Regarding the time summary table for elevators in Halls A and B, there is a
4% difference between the mathematical model and the field study as presented in
Table 2.

4.4 New Floor Selection

Considering the results from the simulation above, the new stopping points proposed
are as follows:
• Tower A:

– S2-1-4-8-12
– 1-8.
• Tower B:

– S2-1-4-8-12-15
– 1-12.

The stopping floor assignment will be changed only for two elevators, one for each
tower, choosing the 8th and 12th floors as stopping points because there are many
classrooms on them. Therefore, by applying the new stopping points, we obtain the
following flow:

Table 1 Arequipa tower summary (UpPeak Math.Mod.)

RTT (seg.) AWT (seg.) ATT (seg.) AJTT (seg.)
1.2 100.61 83.45 75.45 158.90
3.4 107.07 88.81 80.31 169.12

Table 2 Time summary—arequipa tower

Field study (seg.) M. M. UpPeak (seg.) Delta (seg.)
A 133 158.90 25.90
B 182.14 169.12 13
Average 315.14 328.02 12.88 <> 4%
Mathematical Model for the Improvement … 407

Table 3 Journey times

Current Proposed Improvement (seg.)
Assignment Journey time (seg.) Assignment Journey time (seg.)
s2-1-4-8-12 133 1-8 77.62 55.38
s2-1-4-8-12-15 182.14 1-12 120 62.14
Total 315 Total 197 117.52 <> 37.5%.

• In the simulation of the new floor proposal for the tower A elevator, a total of 255
users were served from 6:30 to 7:00 pm.
• In the simulation of the new floor proposal for the tower B elevator, a total of 238
users were served from 6:30 to 7:00 pm.

5 Results and Discussion

A time study was performed for all elevators, with the exception of elevators that
are required to stop at all floors according to government regulation [12], which
establishes that elevators must be set up to allow disabled people to move freely
throughout the building.
Based on the results obtained, the stop points defined for the elevator represent
the main factor increasing journey times; the more the number of stop points, the
longer the trip will take. Hence, a new stop point distribution was realized to reduce
the number of elevator stops.
After acquiring the journey times, it was concluded that the elevator does not need
to stop at the 4th and 15th floors because their passenger flow was between 1 and 2
people, which does not compensate for the time this journey lasts (Table 3).
As a result. the floors were only redistributed for two elevators, one from each
tower. Comparing the current design with the new proposed design, total journey
times could be reduced by 117.52 s, which translates into a 37.5% reduction.

6 Conclusion

This works found that reducing the points of stops of the elevators allows improving
the efficiency and effectiveness of their service by 37.5%. In addition, the number of
passengers served increased by 10.66%, thus reducing the waiting time. The main
factor that generates an increase in travel time, are the stop points, the more stop
points, the longer the trip will take.
As future work, a study of route frequencies may be carried out in order to calculate
performance measures and propose improvements in the vertical traffic.
408 J. C. Lazo Lozano and G. W. Quispe Santivañez

References

1. Wong, P.: Propuesta de mejora del proceso de admisión en una empresa privada que brinda
servicios de salud ambulatorios (2009)
2. Singer, M.: Una introducción a la teoría de colas aplicada a la gestión de servicios. En
revista Abante (2008). http://www.abante.cl/files/ABT/Contenidos/Vol-11-N2/Singer.pdf. Last
accessed 29 Apr 2019
3. Al-Sharif, L., Al Sukkar, G., Hakouz, A., Al-Shamayleh, N.A.: Rule-based calculation and
simulation design of elevator traffic systems for high-rise office buildings. Build. Serv. Eng.
Res. Technol. 38(5), 536–562 (2017). https://doi.org/10.1177/0143624417705070
4. Ruiz, C.: Análisis de mejora de eficiencia energética en ascensores (2017). https://dialnet.uni
rioja.es/servlet/tesis?codigo=122704. Last accessed 07 Apr 2019
5. Iglesias, G.: Diseño y simulación de un algoritmo basado en lógica difusa para el despacho de
ascensores en un edificio (2013). http://docplayer.es/41380371-Diseno-y-simulacion-de-un-
algoritmo-basado-en-logica-difusa-para-el-despacho-de-ascensores-en-un-edificio.html. Last
accessed 27 Apr 2019
6. Fernández, F.: El análisis de contenido como ayuda metodológica para la investigación (2002).
https://www.revistacienciassociales.ucr.ac.cr/images/revistas/RCS96/03.pdf. Last accessed 07
May 2019
7. Urquía, A.: Modelo orientado a objetos y simulación de sistemas hibridos en el ambito del
control de procesos quimicos (2000) (Consulted on April 29, 2019)
8. Silva, C. A., Oliveira, M. N., Calixto, W. P., Santos, L. E. B., Barbosa, J. L. F., Silva, D.
F. A.: Simulation and minimization of waiting time in rows of elevators of public build-
ings. In: 2017 CHILEAN Conference on Electrical, Electronics Engineering, Information and
Communication Technologies (CHILECON), Pucon, pp. 1–6 (2017)
9. Suntaxi, C.: Diseño e implementación de un sistema de control y monitoreo para ascensores
residenciales (2009). http://dspace.ups.edu.ec/handle/123456789/6734. Last accessed 29 Apr
2019
10. Hurtado, D.: Eficiencia en el proceso de instalación de ascensores en edificios residenciales
para lograr un servicio de calidad (2016). http://repositorio.usil.edu.pe/handle/USIL/2584. Last
accessed 29 Apr 2019
11. Systems Modeling: Arena (software) (2000). https://www.arenasimulation.com
12. Norma A.120 Accesibilidad para personas con discapacidad (2 de marzo de 2019) diario oficial
del bicentenario el peruano
A Novel Dataset for the Transport Sector
in a Province of Peru

Miguel Arango Guerrero and Pedro Shiguihara Juárez

Abstract Problems related to public transport and private transport in Peru are
persistent. New proposals to solve them arise, currently the world of data analysis
is starting in Peru, there are not many open datasets useful that allow proposing
solutions in each environment. In this paper, we will collect relevant data of the
transport located in a province of Peru with more than 1000 users involved, restricted
by a delimited geographic area and with 2 years of operations and more than 3000
transport services tracked. In this way, we highlight the importance of the data, the
possible potential uses within the transport, and a case of use of the collected dataset.

Keywords Transport planning · Transport dataset · Public and private transport

1 Introduction

In this paper, we study a forecast model simulation for dropoff locations of the
users in transport, using the concept of collaborative filtering and knowledge-based
approaches. We construct a dataset from information provided by an urban transport
system. The dataset contains records of completed transport services covering one
year from December 2016 until October 2018 restricted by the geographic location
in the province of Pisco, Perú. The generated data is quite faithful about the current
behavior of users in that location. In Peru, urban transport systems are evolving at
the digital level, regarding both public and private transports for different segments
of socioeconomic groups, with the sole objective of improving the service provided
to their customers. The advantage of digitizing transport systems is the ability to
improve management, operations, reduce response times in processes, and especially

M. A. Guerrero (B) · P. S. Juárez

Universidad Peruana de Ciencias Aplicadas, Av. Prolongación Primavera 2390, Santiago de
Surco, Lima, Peru
e-mail: [email protected]
P. S. Juárez
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 409
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_41
410 M. A. Guerrero and P. S. Juárez

the new way of storing and managing the data that is recorded through interactions
with users [1]. Due to the interactions of users and operators of urban transport
systems, it is possible to record all operations performed, regarding a request for a
service to a rating based on the performance of the service provided. This data can be
used to process them and get valuable information for the business in transport. The
obtained dataset can be used to predict the destination location of a user. To do that,
it is necessary to use algorithms that can detect patterns of behavior counting the
variables that we provide and that have determined your destination location within
a service.
To carry out this project, a prediction model of the user’s destination location was
built and allows to forecast a model for dropoff locations of users.
In Sect. 2, the related work is described. In Sect. 3, a brief theoretical foundation
is depicted. In Sect. 4, the process for the generation of the dataset is detailed. In
Sect. 5, the conclusions of this work are described.

2 Related Work

In [2], the destination or next pickup location prediction problem was explored and
the authors proposed four new incremental learning methods. They found that multi-
variate multiple regression (MMR) is the best method in terms of accuracy when the
training data sizes are large. When the training data sizes are small or moderate, then
both the random forest (RF) and support vector regression (SVR) methods are good
choices considering prediction accuracy and computational time.
Also, when the next pickup prediction problem is considered, the MMR method
and the ensemble method are good performers in terms of prediction accuracy.
Finally, in [2], the MMR method is seen to be generally better than the incremental
artificial neural network (ANN) method in terms of prediction accuracy.
In [3], the authors used neural networks, obtaining a range of possible outputs
for any given instance. Also, their work is focused on the personalized driving
experience, helping to prevent accidents by reducing driver distraction. It can be
trained to recognize who is in the car, learning their specific preferences and driving
style and then apply this learning by using a range of variables, event history in the
neighborhood can be added.
In [4], the authors experimented with models for predicting taxi pickups in New
York City, observing the decision tree regression model performed best, likely due
to its unique ability to capture complex feature dependencies. The decision tree
regression model achieved a value of 33.47 for root-mean-square deviation (RMSD)
and 0.9858 for R2—a significant improvement upon the baseline values of 145.78
for RMSD and 0.7318 for R2. The model proposed in [4] could be useful to city
planners and taxi dispatchers in determining where to position taxicabs and studying
patterns in ridership.
In [5], the authors proposed k-means clustering in order to estimate the most
likely pickup points at any given hour and also predict the top nightlife hotspots by
A Novel Dataset for the Transport … 411

learning trends from past pickups. In [6], the authors highlighted special events with
peak travel times such as holidays, concerts, inclement weather, and sporting events.
In this sense, the authors asseverate that to calculate demand time series forecasting
during extreme events is a critical component of anomaly detection, optimal resource
allocation, and budgeting that need to be covered.
In [7], it is specified that the travel demand can be measured and analyzed by
different ways depending on the metrics like the entire trip route, the area of the
end trip, traffic volume, person trip and vehicle trip, passenger vehicle and freight
vehicle, person-mile traveled and vehicle-mile traveled. On the other hand, the travel
demand can be measured and analyzed by different ways depending on metrics such
as the entire trip route, the area of the end trip, traffic volume, person trip and vehicle
trip, passenger vehicle and freight vehicle, person-mile traveled, and vehicle-mile
traveled [8].

3 Theoretical Foundation

A forecast model allows us to estimate predictions from historical data and live
data. It can be used regularly for the trending analysis. The prediction of time series
or predictions of time series are a research area of great interest that has been in
development for several decades.
In particular, in the last three decades, the interest in this area has increased
especially because of the highest processing capacities by the computers, which
allows realizing complex calculations in a few minutes or even in a few seconds. This
increase in interrelation is reflected in the diversity of its applications in different
disciplines, ranging from economics to engineering, where the prediction of time
series is a currently active field of research.
With the information collected by urban transport systems, a dataset can be made
available sequentially through time with the use of Bayesian networks and automatic
learning is possible to find patterns or abstract behaviors that allow us to make
predictions and make decisions based on the results.
The collaborative filtering approach uses the identities of users and items. Implicit
descriptions of the user and items are obtained from a sparse matrix of ratings of
items by users [9]. It is possible to learn about a user by the items they have previously
rated and the users who have rated items in common with them by using cutting-edge
technologies related to the machine learning modules provided by different software
frameworks [10].

4 Datasets

In this paper, we provide two datasets, one for users and other for trips. Dataset related
to users contains 1383 user records, collected from December 1, 2016, to October
412 M. A. Guerrero and P. S. Juárez

1, 2018. The quality of the user dataset is consistent because they were recorded by
real people located in the area delimited by the project and were manually verified
by the data entered through the mobile application.
Between the months of August and September of 2017, greater user activity is
seen within the application, since marketing strategies were applied in order to attract
more users and promote the use of the application as shown in Fig. 1.
In Fig. 2, we can observe a heatmap of the final locations of all trips made by users,
the green areas represent the locations geographically and the green areas represent
locations in high attendance. A heat map was made of the routes taken by each driver
to complete the trips assigned to the dataset. The place that has been taken by more
drivers is ‘Abraham Valdelomar’ Avenue, which is one of the most transited because
connects almost the entire city.
Figure 3 shows the heat map of all the routes of the trips made by the drivers, the
colors represent the number of times a driver has passed through that area, where
then the green color represents a frequency of few times, and the red color represents
a frequency of many times. For the elapsed time in the trips, we have categorized the
time of all registered trips, in such a way:
1. High time trips: trips that last from 301 up to the maximum allowed number of
seconds elapsed.
2. Medium time trips: trips that last from 121 to 300 s of duration.
3. Low time trips: trips that last from 0 to 120 s of duration.
The largest number of trips has been recorded with a lapse of time between 121 and
300 s. For the distances traveled in the trips, we have categorized the distance of all
the records such as long-distance: trips that locomote from 2.51 km to the maximum
allowed amount of distance traveled, medium-distance: trips that locomote from
0.51 to 2.5 km of travel and low-distance: trips that locomote from 0 to 0.5 km away.
Further analysis corroborates that the largest number of registered trips is represented
by low-distance trips.
As a compilation, Fig. 4 shows the categories of the trips by time and the quantity

Fig. 1 User activity—from December 1, 2016, to October 1, 2018

A Novel Dataset for the Transport … 413

Fig. 2 Heatmap of dropoff locations for users

Fig. 3 Heatmap of drivers routes for trips

414 M. A. Guerrero and P. S. Juárez

Fig. 4 Frequency of validated trips according to time

Fig. 5 Frequency of validated trips according to distance

of validated trips and Fig. 5 shows the categories of the trips by distance and quantity
of validated trips.
These preliminary results indicate all the potentials this database might able to
have. Potential uses of the dataset could be (1) forecasting of service demand: get
to predict the demand of trips within a certain period, (2) forecasting of the shortest
route to complete a trip: get to choose the shortest route to get as quickly as possible
to the user’s destination location, (3) forecasting of the estimated time for the journey
of the trip: get to predict the time it will take the trip to be completed, (4) forecasting
of the estimated distance for the journey of the trip: get to predict the distance that the
driver will travel on the trip to be completed, (5) recommendation system for dropoff
locations oriented to drivers: recommend dropoff locations to the driver, after having
accepted a trip without a defined destination.

5 Conclusions

This work obtains information from a transportation business with a moderate effort.
As a result, heat maps characterizing the behavior of users from private transport
regards dropoff location in the city of Pisco, Perú, was successfully unveiled and
A Novel Dataset for the Transport … 415

with more than three thousand records, it is possible to predictive models with an
acceptable likelihood for final destinations from users.
As future work, there are some issues that can be approached such as forecasting
of service demand, forecasting of the shortest route to complete a trip and forecasting
of the estimated time for the journey.

References

1. Westerman, G., Tannou, M., Bonnet, D., Ferraris, P., McAfee, A.: The Digital Advantage:
How Digital Leaders Outperform Their Peers in Every Industry (1 June, 2019). https://www.
capgemini.com/resources/the-digital-advantage-how-digital-leaders-outperform-their-peers-
in-every-industry/
2. Laha, A. K, Putatunda, S.: Real Time Location Prediction with Taxi-GPS Data Streams. India
(2017)
3. Gandhi, S.: Analysing and Predicting Driver’s Next Location (2015)
4. Grinberg, J., Jain, A., Choksi, V.: Predicting Taxi Pickups in New York City. Autunmn (2014)
5. Kumar, J., Surana, J. Kapoor, M., Nahar, P. A.: Perfecting Passenger Pickups: An Uber Case
Study
6. Nikolay, L.: Engineering Uncertainty Estimation in Neural Networks for Time Series Prediction
at Uber (1 June, 2019). https://eng.uber.com/neural-networks-uncertainty-estimation/
7. Chatterjee, A., Venigalla, M.: Travel Demand Forecasting for Urban Transportation Planning
(2004)
8. Studio 3T, GUI for MongoDB. Available in: https://studio3t.com/
9. Ben, J., Dan Frankowski, S., Jon, H., Shilad, S.: Collaborative Filtering Recommender Systems
10. Machine Learning Studio. Available in: https://azure.microsoft.com/es-es/services/machine-
learning-studio/
Development of a Mobile Application
for English Language Learning Through
Corrective Feedback

Adriana Guanuche , Gustavo Caiza , Osana Eiriz , and Roberto Espí

Abstract Learning English as a foreign language is committed to using learning

models that are linked to error correction by analyzing comments provided through
corrective feedback to correct the error or to confirm the correct response. Having
the chance to receive immediate corrective feedback allows students to be in contact
directly with their learning process. This article implements an augmented reality
application for learning English as a foreign language using corrective feedback
and thus creating an accessible teaching tool that allows students to have additional
teaching material to interact with the real world with virtual objects and animations.
The application was designed by using Vuforia, Unity, Visual and Android Studio
software. The objective of this article is to teach the student in an interactive way
and consequently create a user-friendly link. The results show a great application in
functionality using augmented reality for English language learning.

Keywords Corrective feedback · Augmented reality · Vuforia · Mobile application

1 Introduction

Knowledge and proficiency in the English language have become a fundamental

necessity to meet the challenges of the twenty-first century. 1.5 billion people out of

A. Guanuche · G. Caiza (B)

Universidad Politécnica Salesiana, UPS, 170146 Quito, Ecuador
e-mail: [email protected]
A. Guanuche
e-mail: [email protected]
A. Guanuche · O. Eiriz · R. Espí
Universidad de La Habana, 10200 La Habana, Cuba
e-mail: [email protected]
R. Espí
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 417
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_42
418 A. Guanuche et al.

7.53 billion people who exist in the world speak English, according to [1], that is,
approximately one in five people are dominant by English, and so if English is their
second language, any professional will be given better opportunities in the personal
and professional world.
Due to the above and with the advance in Information and Communication Tech-
nologies (ICT), new techniques and applications for learning English have been
developed. As for the techniques [2], mentions four models for language improve-
ment which focus on learning-acquisition, self-monitoring, entry and affective filter.
The first model focuses on language learning in a similar way to language maternal
naturally and unconsciously, the second one has to do with conscious learning which
is exposed to the correction of errors, the third model connects the student with their
next level of language proficiency and the fourth takes into account the motivation of
the student. In relation to the second and fourth models both take into account correc-
tive feedback and motivation through apps. Corrective feedback consists of making
known the errors present in evaluation so that the student does not feel dissatisfaction
or annoyance with the mistake made, on the contrary, this serves to motivate him
or her not to commit it again [3]. The main advantages of corrective feedback are:
It provides comments on the wrong answers. It keeps students informed about their
academic performance. It fosters in students the interest in self-correction. It is timely
because it intervenes at the right time. It is flexible since it is adapted to the level of
the student. It uses technological resources to speed up their intervention and, above
all, it contributes to the improvement of learning [4–8]. The Augmented Reality (AR)
aims to combine images generated by electronic devices, on the real-world vision
captured by one or several cameras of that device [9, 10].
In several educational and design applications, information that is associated with
the real world is generated; it can be mentioned the 4D Anatomy application, which
allows visualizing on a class book the anatomical parts of the human body with their
names and main characteristics in a three-dimensional way [11], a design application
that has been very popular in North America is IKEA Place, which allows placing
furniture from your catalog in the different home spaces having a three-dimensional
idea of the appearance of this furniture before being placed [12].
This article implements an application of augmented reality for learning the
English language using corrective feedback, the application was designed with
vuforia, unity, visual studio, and android studio software. The objective is to reach
the student in an interactive way and thus to create a friendly link with the user,
leaving aside the traditionalism of education. The article is organized as follows:
Sect. 2 describes the materials and methods used, Sect. 3 shows the design and
implementation, Sect. 4 shows the results and finally Sect. 5 provides the conclusions.
Development of a Mobile Application … 419

2 Materials and Methods

2.1 Augmented Reality

Augmented reality increases the images of reality, from its capture by the camera of
computer equipment or mobile device that adds virtual elements for the creation of a
mixed reality to which computer data has been added. There are four levels at which
levels can be understood as a way of measuring the complexity of the technologies
involved in the development of RA systems [13, 14].

2.2 Vuforia and Unity

Vuforia is an application for the development of augmented reality applications, it is

a web environment where the users can create and administrate their markers. This
app is the leader thanks to its exceptional technology of artificial vision, efficient
tracking benefits and its compatibility with multiple platforms. It is the most used
software for digital devices and sunglasses currently available in the market [15].
Unity is a 3D figure reader that uses C, C++ and C Sharp languages. In order to
execute the projects from Unity to mobile devices, it is indispensable to have the
appropriate software development kits (SDK) that enable their reading [15].

2.3 English Language Learning with M-Learning

Mobile learning can be defined as the educational modality that facilitates knowl-
edge construction, learning problem solving and the development of different skills
in an autonomous way through the interaction with portable mobile devices. The
development of M-learning has resulted in the creation of multiple applications for
English language learning, being the most important ones: Duolingo, Wibbu English,
uSpeak.

2.4 Corrective Feedback

It is defined as the information given to the students about their academic perfor-
mance, it can be used to correct errors or congratulate on a correct answer [16].
The corrective feedback should be coherent, dynamic, clear, frequent and proactive
[17–19]. The corrective feedback is classified as follows.
Positive and negative: Named positive to comments based on the correct answers,
and the negative is granted to the wrong answers [8].
420 A. Guanuche et al.

Oral, written and mixed: The corrective feedback could be through the oral or
written word, and the mixed one with the combined use of the previous two [20].
Explicit and implicit: The first occurs when the teacher identifies the error and
directly provides the correct answer, and the second occurs when the teacher invites
the student to find the correct answer [21].
Self-correction and peer correction: The self-correction provides the opportu-
nity for students to address the correct answer for themselves [22].

3 Development of the Mobile Application

The following tools were used to perform the English test application with AR:
vuforia analyzes the recognition of images, unity allows to develop the graphic
environment, and android installs the application previously generated in visual as
shown in Fig. 1.

3.1 Images Selection

For image recognition, the vuforia target image tool was implemented, this tool
recognizes full detail images. The selected numbers obtained a result of 4 stars so,
they will have easy recognition when executing the augmented reality application.
The numbers selected to perform the application are those shown in Fig. 2.
Once the images are selected, each of the numbers is imported into vuforia, to
generate the points that allow recognizing the characteristic patterns of the image,
this analysis is carried out in grayscale, in this instance the image qualification is
generated obtaining 4 stars as shown in Fig. 3.

Fig. 1 Implementation of the application

Development of a Mobile Application … 421

Fig. 2 Numbers selected for the augmented reality application

Fig. 3 Image recognition points

3.2 Application Development

The images must be imported to Unity where the application screen is configured
with a resolution of 2200 × 1400 pixels because it is the standard size of the screens,
in addition, this measure allows us to run the application on higher resolution screens.
Each image was assigned a scene of its own, each question has its own interface,
where the texts of the application, theme, question number, answers, option buttons
for correct or incorrect answers and feedback animations are generated as shown in
Fig. 4.
Through Visual Studio, the feedback animations for correct or incorrect responses
were programmed, this algorithm remains active for 2 s and then becomes invisible
so that the user can practice several times until the topic is understood.
422 A. Guanuche et al.

Fig. 4 Development in the unity program

3.3 Simulation of the Application

Before installing the app on the Smartphone, simulation in Unity is performed. Once
the correct answer is chosen the app displays the message “Correct” within the check
icon and a happy face as shown in Fig. 5.

Fig. 5 Correct answer feedback

Development of a Mobile Application … 423

4 Results

Next, the apk file, English Test.apk, is generated. Finally, the application is built thor-
ough the Android Studio platform and now is ready to be installed in a Smartphone.
The application contains 4 types of questions: Fill in the blanks, Multiple Choice,
Structure Sentences and True or False. The type of questions used on the application
responds to the advantage of using accurate content on learning, an appropriate tech-
nological medium, and proper monitoring of the academic progress. Figure 6 shows
how the application works for Multiple Choice questions. The first step is to run the
application on the smartphone, the next step is to focus on any number belonging to
the database. Finally, the question is shown.
In the last figure the option “ANSWER ME” is displayed, in which you click on
and the different options appear depending on the type of question. Then the student
chooses whatever option and up to his answer a message is showed if the choice was
correct or incorrect with the corresponding feedback as it is seen in (Fig. 7).

Fig. 6 Development of the unity program application

Fig. 7 Feedback for answers

424 A. Guanuche et al.

The functioning application results show that it is very friendly with the user and
considers advantages of corrective feedback for learning of the english language.

5 Conclusions

The English learning application as a foreign language using augmented reality is an

approachable and didactic tool that allows students to have extra didactic material
in order to interact with the real world through virtual objects and animations. An
important fact of the application is that it can be installed in whatever mobile phone
without being limited to the operative system so it can be set up for multiplatform
systems.
The corrective feedback represents support for the English learning process since
it allows students and teachers to have the possibility of being informed about the
academic advance as well as it encourages students to be more participative through
the error correction or the verification of the correct answer. If the answer is correct
feedback allows students to confirm their response or if it is incorrect error correction
motivates students to look for the right response.

References

1. Robin, L.: ¿Podrá el inglés seguir siendo el ‘idioma favorito’ del mundo? (2018). Available:
https://www.bbc.com/mundo/noticias-44232687. Accessed: 22 July, 2019
2. Cerdas, G., Ramírez, J.: La enseñanza de lenguas extranjeras: historia, teoría y práctica. Rev.
las Lenguas Mod. 297–316 (2015)
3. Aumentada, R.: Aplicaciones De Los Dispositivos Móviles Augmented Reality, an Evolution
of Theapplication. Rev. Medios Y Educ. 41, 197–210 (2012)
4. Daneshvar, E., Rahimi, A.: Written corrective feedback and teaching grammar. Procedia Soc.
Behav. Sci. 136, 217–221 (2014)
5. Aranguiz, M. F., Espinoza, A. Q.: Estudios de lingüística inglesa aplicada, 103–131 (2016)
6. Chen, S., Nassaji, H., Liu, Q.: EFL learners’ perceptions and preferences of written corrective
feedback: a case study of university students from Mainland China. Asian-Pacific J. Second
Foreign Lang. Educ. 1–17 (2016)
7. Lima, S.: Enriquecer la realimentación para consolidar aprendizajes 14, 9–26 (2017)
8. Ene, E., Upton, T.A.: Journal of second language writing synchronous and asynchronous teacher
electronic feedback and learner uptake in ESL composition. J. Second Lang. Writ. 41(May),
1–13 (2018)
9. Prendes Espinosa, C.: Realidad aumentada y educación: análisis de experiencias prácticas.
Píxel-Bit, Rev. Medios y Educ. (46), 187–203 (2014)
10. Basogain, X., Olabe, M., Espinosa, K., Rouèche, C., Olabe, J. C.: Realidad Aumentada en la
Educación: una tecnología emergente (2012)
11. Interactive Anatomy Zrt: 4D Anatomy (2018). Available: https://www.4danatomy.com/.
Accessed: 23 July, 2019
12. Inter IKEA Systems: IKEA Mobile App (2019). Available: https://www.ikea.com/gb/en/cus
tomer-service/mobile-apps/. Accessed: 23 July, 2019
13. Fombona Cadavieco, J., Vázquez-Cano, E.: Posibilidades de utilización de la geolocalización
y realidad aumentada en el ámbito educativo 319–342 (2017)
Development of a Mobile Application … 425

14. Méndez, M., Carracedo, J. P.: Realidad Aumentada: Una Alternativa Metodológica en la
Educación Primaria Nicaragüense. IEEE Rita 7, 102–108 (2012)
15. Estefanía, Paspuel, T.: Estudio del motor de videojuego unity con sdk vuforia para el desarrollo
de aplicaciones móviles de realidad aumentada, aplicación de tarjetas ilustradas en pares para
niños (2014)
16. Kerr, P.: Giving feedback on speaking. Giv. Feed. Speak. Part Cambridge 92, (December, 2017)
17. Sánchez, S.: retroalimentación oral en la adquisición de segundas lenguas, 283–311 (2014)
18. Hernández-Islas, M.: La retroalimentación automatizada en la enseñanza del inglés como
segunda lengua. Rev. Iberoam. Ciencias 3, 156 (2016)
19. Núñez, L. M.: La retroalimentación en la enseñanza de la expresión escrita en inglés: un
proyecto de desarrollo profesional Feedback on teaching writing in English (2016)
20. Ortiz, M.: Uso de la retroalimentación correctiva focalizada indirecta con claves metalingüís-
ticas en la adquisición del sufijo -s en la tercera persona del singular en inglés, en estudiantes
de un programa de formación pedagógica en efl de una universidad chilena. Folios 44, 127–136
(2016)
21. Hernández, M.: La retroalimentación automatizada en la enseñanza del inglés como segunda
lengua. Rev. Iberoam. Ciencias 3 (2016)
22. Silva, M.: La retroalimentación en la corrección de la escritura. Rev. Lingüística 15 (2015)
Influence of the Chinese Culture
on the Negotiation Process Peru-China

Nicole Alexandra Acevedo Cueva

Abstract This research is an intercultural study conducted with a deductive

approach by generating theory-based questions derived from the socio-cultural
framework of the Chinese business negotiation process, to be proved by the applica-
tion of a self-managed open questionnaire on a sample of seven experienced Peruvian
negotiators in negotiating with Chinese in order to determine how the Chinese nego-
tiating culture influences the negotiation process. The results proved that the Chinese
negotiating culture does influence the negotiating process according to the perception
of 100% of the participating Peruvian negotiators and derives the following charac-
teristics that a Peruvian negotiator could visualize and for which should be prepared
during negotiations with Chinese to achieve success: the importance of knowing
about the negotiating culture of the Chinese counterpart, the relevance of presenting
exact information, the presence of an intermediary or mediator, the significance
of socializing with the Chinese, the difficulty of negotiating the price, the slow and
hierarchical Chinese decision-making process, the reluctant Chinese attitude to make
concessions, the importance of signing a legal contract, and the Chinese tendency to
renegotiate.

Keywords Chinese negotiating culture · Negotiation process · Peruvian negotiator

1 Introduction

China has held a position of being an economic power worldwide that made Chinese
culture and its negotiation process the focus of attention for western countries [1]
interested in establishing agreements to access the desired Chinese market. However,
the negotiation with a Chinese counterpart is complicated [2], since foreign and
Chinese negotiation process often seems incompatible because of their differences

N. A. Acevedo Cueva (B)

Universidad Privada del Norte, Trujillo 13006, Peru
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 427
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_43
428 N. A. Acevedo Cueva

with deep cultural origins [3], which indicates that the way people negotiate is influ-
enced by culture. At the same time, culture is defined as a framework in which
international negotiators adjust their strategies, tactics, and styles in order to obtain
success in negotiations; because when intercultural negotiations fail, the common
reason is related to cultural misunderstandings. So, this fact determines that a Peru-
vian negotiator will have to adapt to the culture and, especially, to the negotiation
process of the Chinese counterpart in order to achieve success [4].

1.1 The Influence of Chinese Negotiating Culture

on the Negotiation Process

Ghauri and Fang [2] designed a framework that lends a way of studying the influ-
ence of Chinese negotiating culture in three interlinked dimensions such as Confu-
cianism, Sun Tzu’s stratagems, and sociopolitical and cultural condition; on the stages
of negotiation process: pre-negotiation (search for influence, presentation, informal
discussion, and trust-building), formal negotiation (task-related exchange of infor-
mation, persuasion, concession, and agreement), and post-negotiation (new rounds
of negotiation). Therefore, the Chinese cultural bases are determining factors in the
negotiation process with a Chinese counterpart, because the success or failure will
depend on the adaptation and comprehension that the negotiator has about Chinese
culture and its way of doing business [5].

1.2 The Motivation of the Work

China has positioned for four consecutive years as Peru’s first strategic business
partner with almost 170 Chinese companies operating in Peru’s territory with an
investment of more than 18 billion dollars [6]. A reality that proves the existence of
efficiently developed negotiations.
The purpose of this investigation was to benefit the negotiations between Chinese
and Peruvian parties by providing a number of characteristics derive from the influ-
ence of Chinese negotiating culture on the negotiation process that a Peruvian nego-
tiator could face and therefore adapt in order to establish agreements, which is highly
feasible due to the Peruvian tendency to accommodate their negotiation process
according to the counterpart’s culture in order to avoid conflicts and obtain success
[7]. Despite being able to collect numerous documents about the commercial trajec-
tory between Peru and China, almost none of them describes the actual negotia-
tion; and regardless the numerous theory about Chinese negotiation or culture that
describes how the Chinese act during international negotiations, there is a possibility
that could be certain variations while negotiating with Peruvians. Because of this,
Influence of the Chinese Culture … 429

the importance of this exploratory investigation is to be an initial point for further

investigations about the intercultural negotiations.
This research sought to determine how the Chinese negotiating culture influ-
ences the negotiation process according to the perception of a Peruvian negotiator.
Therefore, the first objective is to determine the manifestation of the Chinese negoti-
ating culture regarding the dimensions of Confucianism, Sun Tzu’s stratagems, and
sociopolitical and cultural condition, according to the perception of a Peruvian nego-
tiator. The second objective is to determine the existence of influence of the Chinese
negotiating culture’s dimensions on the stages of the negotiation process according
to the perception of a Peruvian negotiator. The third objective is to determine the
characteristics of the negotiation process influenced by the Chinese negotiating
culture’s dimensions regarding to the stages of pre-negotiation, formal negotiation,
and post-negotiation according to the perception of a Peruvian negotiator.

2 Methodology

This research is based on a qualitative methodology, a triangulation technique of

three types: data triangulation, methodological triangulation, and theoretical trian-
gulation [8] was applied to derive a procedure of three phases: (1) applying data
triangulation by designing the self-administered open questionnaire in line with the
indicators of each of the dimensions of the respective study variables; (2) applying
the methodological triangulation by executing the questionnaire; and (3) applying
theoretical triangulation by compiled the participants responses and classifying the
qualitative data, involved that the data collected had to be read and reread, where
codes were assigned to the participants to identify their responses in each category
and how often they alluded to the same topic or word—to finally be contrasted with
previous research, theories, or findings.
The study was carried out considering a type of non-probability sampling for
convenience, obtaining a sample of seven Peruvian international negotiators with
experience in the role of negotiator with a counterpart of Chinese nationality. The
self-managed open questionnaire data collection method was applied and declared
qualified with a corrected range validity coefficient of 1.00. As a qualitative research
that required consent by implying the identification, dissemination, and interpre-
tation of the participants judgments and perceptions, this research was conducted
in an ethical manner by delivering to the participants an informed consent letter
that detailed the reasons and conditions of the investigation, while recognizing the
commitment to ensure the integrity and confidentiality of the information provided
during the completion of questionnaire, and certified the consent of the participants
through their personal data and digital signature.
The investigation was conducted and took place on 2019, where all the participants
signed an Act of Informed Recognition. At this act, they acknowledge the objective
of the study and that all the information provided would be only used for this research
430 N. A. Acevedo Cueva

as well as the guarantee of anonymity giving the choice of participation or not, with
no need for appreciation by ethics committee.

3 Results

3.1 Chinese Negotiating Culture’s Manifestation During

the Negotiation

Figure 1 shows the results of the questionnaire about the Confucianism dimension.
The 100% of Peruvian negotiators determined the importance for Chinese to build
relationships and trust to access to concessions, the relevance of understanding non-
verbal communication, and the tendency of showing off power to be more demanding.
The 85.7% pointed out that Chinese delegations are numerous, while 14.3% alluded
that depends on the extent of the topic and could be taken as a signal of inexperience.
Another 85.7% mentioned the relevance for Chinese to have a formal behavior;
however, 14.3% declared that it depends on the context because there is not formalism
while negotiating with Chinese merchants or in the public sector. The 74.4% classified
Chinese communication as direct.
Figure 2 shows the results of the questionnaire about Sun Tzu’s stratagems dimen-
sion. The 100% of Peruvian negotiators declared the use of tactics from the Chinese
as manipulating the counterpart by pointing out weakness, intimidating by being
aggressive or mocking, and showing off power to make the others give in and the
71.4% of them cataloged the Chinese way of doing business as mixed (cooperative
and competitive).

Fig. 1 Chinese confucianism dimension manifestation during a negotiation

Influence of the Chinese Culture … 431

Fig. 2 Chinese Sun Tzu’s stratagems dimension manifestation during a negotiation

Fig. 3 Chinese sociopolitical and cultural dimension manifestation during a negotiation

Figure 3 shows the results of the questionnaire about the sociopolitical and cultural
condition dimension. The 100% of Peruvian negotiators declared that the Chinese
ostentation of being an economic power is used as a negotiation mechanism and
the alignment between the plans of Chinese companies with the interests of their
government since this is the major shareholder of most of the companies.

3.2 Chinese Negotiating Culture’s Influence

on the Negotiation Process

Figure 4 shows the results of the questionnaire about the influence of Chinese culture
when negotiating. The 100% of Peruvian negotiators affirmed that the Chinese nego-
tiating culture does influence the negotiation process. The 57.14% of them considered
that this derived the need of adaptation that should be only in a cultural way to gain
more receptiveness from the Chinese part to ensure a good negotiation.
432 N. A. Acevedo Cueva

Fig. 4 Influence of the Chinese negotiating culture on the negotiation process

3.3 Stages of the Negotiation Process Influenced

by the Chinese Negotiating Culture

Figure 5 shows the results of the questionnaire about the pre-negotiation stage. The
100% of Peruvian negotiators described the importance of knowing about the Chinese
negotiating culture to identify Chinese strategies and the relevance of presenting exact
information to ensure credibility. The 71.4% described the significance of socializing
with Chinese to gain more fluidity of the negotiation even when 28.6% cataloged
it as not mandatory until the negotiations have significative advances. The 57.1%
considered relevant the presence of an intermediary or mediator with experience
or member of Chinese culture to help in building trust; however, 42.9% described
it as unnecessary when both parties are officials or representatives in international
organizations.

Fig. 5 Characteristics of the pre-negotiation stage influenced by the Chinese negotiating culture
Influence of the Chinese Culture … 433

Fig. 6 Characteristics of the formal negotiation stage influenced by the Chinese negotiating culture

Fig. 7 Characteristics of the post-negotiation stage influenced by the Chinese negotiating culture

Figure 6 shows the results of the questionnaire about the formal negotiation stage.
The 100% of Peruvian negotiators described the existence of slow and hierarchical
Chinese decision-making process, the difficulty of negotiating the price due to the
Chinese custom of bargaining, the reluctant Chinese attitude to make concessions,
and the importance of signing a legal contract where the Chinese expect that any
formal verbal declaration during the negotiation will be considered in the contract.
Figure 7 shows the results of the questionnaire about the post-negotiation stage.
The 100% of Peruvian negotiators described the Chinese tendency to renegotiate
where the Chinese consider feasible to change or adapt some terms on the already
signed contract that could bring them complications.

4 Discussion

Considering the accomplish of the objectives, there are main differences and similar-
ities between the recollected theory about Chinese negotiating culture that describes
how the Chinese act during international negotiations, and the actual findings of this
paper about the Chinese-Peruvian negotiation’s reality.
434 N. A. Acevedo Cueva

About the main similarities, the first and most important is when 100% of Peru-
vian negotiators affirmed that the Chinese negotiating culture does influence in the
negotiating process and leads to the need of adaptation; something aligned to the find-
ings of Rodriguez [5] that declared that Chinese cultural bases are decisive during
negotiation because the success or failure depends on the negotiator’s adaptation
to Chinese culture and its way of doing business, which is highly feasible due to
the fact explained by Ogliastri and Salcedo [7] that a peruvian negotiator tends to
adapt or accommodate the negotiating process according to the counterpart’s culture
as much as possible by the negative conception of conflicts and by the continuous
search for success. The second is when 74.4% of Peruvian negotiators described
that Chinese negotiation style as cooperative and competitive, agreeing with Torres
[4] that considered that the Chinese negotiation style is nurtured by stratagems of
Sun Tzu, and that a “seeming” cooperative Chinese attitude during negotiation could
also be stratagems of Sun Tzu, as stratagem number 17 which consists in “giving the
adversary something to obtain greater benefit,” or stratagem number 11 which refers
to “making concessions to achieve the main objective.”
About the main differences, the first lies in the explanation of Rodriguez [5] that
described that Chinese prioritize formalism or etiquette in all negotiations due to
the Confucianist value of respect, while 14.3% of Peruvian negotiators declared that
a formal behavior during negotiation actually depends of the context—especially
while negotiating with Chinese merchants and at the public sector where there is not
formalism at all. The second resides in the declaration of Torres [4] that explained that
a large number of Chinese delegations derives from Chinese position of feeling more
confidence negotiating in group; however, 14.3% of Peruvian negotiators alluded that
the number of members on Chinese delegations actually depends on the extent of
the topic and could be also taken as a signal of inexperience from the Chinese coun-
terpart. The third is when Hall and Reed Hall [9] classified Chinese communication
as indirect due to the definition of high context culture. Nevertheless, this is at the
same time a similarity, because 74.4% of Peruvian negotiators cataloged Chinese
communication as direct, matching the declarations of McColl et al. [1] that Chinese
have adapted western forms of negotiation and communication. The fourth is found
on the statement of Miles [10] that proposed socializing with the Chinese counterpart
as mechanism to gain more fluidity at the beginning of the negotiation, a fact world-
wide endorsed by other authors, while 25.6% of Peruvian negotiators cataloged it
as “not mandatory” and declared it as something that can be postponed until nego-
tiations have advance between parties. The fifth lies in the declaration of Graham
and Lam [11] that highlighted the need of a permanent intermediary or mediator
during negotiations with a Chinese counterpart as a way of building relationships
and trust; however, 42.9% of Peruvian negotiators cataloged as “unnecessary” when
both parties are officials or representatives in international organizations, or in other
words, share groups or affiliation in common.
Influence of the Chinese Culture … 435

5 Conclusions

This paper aimed to determine how the Chinese negotiating culture influence the
negotiation process according to the perception of the Peruvian negotiator, by
providing a list of characteristics derive from this influence, that a Peruvian negotiator
could face and therefore adapt in order to establish agreements.
The main findings were the importance of knowing about the negotiating culture
of the chinese counterpart; the relevance of presenting exact information; the pres-
ence of an intermediary or mediator that 42.9% of peruvian negotiators cataloged as
“unnecessary” when both parties are officials or representatives in international orga-
nizations, or in other words, share groups or affiliation in common; the significance
of socializing with the chinese that according 25.6% of peruvian negotiators is “not
mandatory” and declared it as something that can be postpone until negotiations have
advance between the parties; the difficulty of negotiating the price due to the chinese
custom of bargaining; the slow and hierarchical chinese decision-making process;
the reluctant chinese attitude to make concessions derived from the chinese disgust
of yielding that implies having to show them the mutual benefit of a concession; the
importance of signing a legal contract where the chinese counterpart expect that any
formal verbal declaration during the negotiation will be considered; and the chinese
tendency to renegotiate where the chinese counterpart consider feasible to change or
adapt some terms on the already signed contract that could bring them complications.

References

1. McColl, R., Descubes, I., Elahee, M.: How the Chinese really negotiate: observations from an
Australian-Chinese trade negotiation. J. Bus. Strateg. 38(6), 38–46 (2017)
2. Ghauri, P., Fang, T.: Negotiating with the Chinese: a socio-cultural analysis. J. World Bus.
36(3), 303–325 (2001)
3. Messmann, S.: Management by Sun Zu and confucius—doing business with Chinese partners.
In: 2nd International OFEL Conference on Governance, Management and Entrepreneurship,
pp. 44–45. Ciru, Dubrovnik (2014)
4. Torres, J.: Chinese Negotiation Styles in International Business Negotiation, pp. 9–65, Argosy
University, Sarasota (2011)
5. Rodriguez, C.: Key Aspects of the Current Business Negotiations in China. Universitat Oberta
de Catalunya, Catalunya (2013)
6. UDEP Homepage. http://udep.edu.pe/hoy/2018/relaciones-entre-china-y-peru-estan-en-su-
mejor-momento/. Last accessed 21 November 2019
7. Ogliastri, E., Salcedo, G.: Peruvian negotiating culture: a exploratory study. J. Econ. Finance
Adm. Sci. 13(25), 9–33 (2008)
8. Russeft, D., Preskill, D.: Evaluation in Organizations: A Systematic Approach to Enhancing
Learning, Performance and Change, 1st edn. Perseus, Cambridge
9. Hall, T., Reed Hall, M.: Hidden Differences, 2nd edn. Anchor, New York (1990)
10. Miles, M.: Negotiating with the chinese: lessons from field. J. Appl. Behav. Sci. 39(4), 453–472
(2003)
11. Graham, J., Lam, M.: The Chinese Negotiation. Harvard Business Review Press, Boston (2003)
Phone Calls Speech-to-Text:
A Comparison Between APIs
for the Portuguese Language

Nilton M. Iinuma and Massaki de O. Igarashi

Abstract Automation and artificial intelligence technologies are bringing improve-

ments in efficiency and cost savings to companies in different areas. These techno-
logical developments and software use, platforms and infrastructure as a service have
grown significantly in recent years, especially in the field of speech-to-text conver-
sion. The rapid scalability and abstraction associated with the democratization of
technology may be the cause of this growth. Developers often make the latest tech-
nologies and solutions available through routines and programming standards, the
Application Programming Interface—API. Therefore, this paper aims to compare
the effectiveness of speech-to-speech conversion tools available by cloud solution
providers based on Portuguese telephone call recordings. For the development of this
paper, some audio recordings and text conversion were made in order to compare
the efficiency of the conversion. The test shows that Word Error Rate (WER) has a
direct impact on conversion quality.

Keywords Speech to text · API · Phone calls · Comparative

1 Introduction

The customer relationship sector of various companies has been seeking greater
efficiency and lower cost driven by technologies, for example, the use of “platform
as a service” (platform as a service—PaaS), infrastructure as a service (IaaS) and even
artificial intelligence. These technologies have changed the landscape of call centers.
They are replacing people with smart systems that aggregate these technologies.

N. M. Iinuma · M. de O. Igarashi (B)

Mackenzie Presbyterian University, Campinas, Brazil
e-mail: [email protected]
N. M. Iinuma
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 437
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_44
438 N. M. Iinuma and M. de O. Igarashi

More drastic cases have already been reported, such as the complete closure of
a call center. It was replaced by a debt renegotiation startup that uses software and
technological solutions [1].
This demand has driven the PaaS market, which has grown by almost 20% per
year [2], probably due to its scalability and abstraction capability. This growth also
seems to have been favored by the “democratization of technology” through the free
access to these advanced technologies made available by any developer.
This access is available through routines and programming standards, popularly
called API (Application Programming Interface—API). According to [3], the use
of PaaS, IaaS and SaaS (software as a service) brings some advantages such as: not
going through the company’s firewall, elasticity (rapid scalability), cost optimiza-
tion (due to resource sharing), available for all, virtualization (easy resource alloca-
tion), operation and management functionality, different levels of hiring, IP address
provided via DHCP per VM (IP Static/global are also available but at additional cost),
private VLANs, IP addresses, VPN, round-robin DNS for load balancing, content
compression, geographic load distribution (example: Big-IP global traffic manager).
The technology-minded companies such as IBM, Google, Microsoft, and Amazon
already offer services that use artificial intelligence; for example, in the speech-to-
text converters, artificial intelligence is used to combine language and grammar
structures with speech signal processing, in order to get more accurate identification
of the words.
Therefore, this research, which aimed to test and compare the efficiency of some
voice-to-text conversion APIs from leading cloud solution providers, can be used for
future implementations in further researches in this area.

2 Theoretical Reference

Artificial intelligence, more specifically neural networks, as reported by [4], can

be used to automatically build a training dataset used for speech recognition from
YouTube videos. Several filters and post-processing steps were used in this work and
the amount of noise and the Word Error Rate (WER) were evaluated (in this study
the data sets presented results ranging from 15.8 to 36.6%) after transcription.
Another subarea of artificial intelligence reported in the bibliography is natural
language processing (NLP), a core of cognitive computing that enables the processing
of natural language diction. NLP is a very extensive topic and includes many subdivi-
sions (of which: natural language understanding, natural language generation, knowl-
edge base building, dialog management systems, intelligent tutor system, speech
processing, data mining–text mining–text analytics) [5].
Among the technologies available and necessary for speech transcription, the
speech-to-text tool plays an important role in this regard. However, it is important to
note that the telephone system may have additional difficulties, such as distortion.
Voice digitization in a telephone system is performed by sampling the voice signal.
In this process, it is common to limit the frequency range of the optimization system.
Phone Calls Speech-to-Text: A Comparison … 439

It is important to remember that humans can hear frequencies in the range of 20 Hz–
20 kHz. However, studies state that the range between 200 Hz and 3.2 kHz is sufficient
for a minimum understanding of speech. [6].
According to the Shannon–Nyquist sampling theorem [7], the sampling rate must
be at least twice the highest frequency so the sampling rate in telecommunications
systems is usually 8 kHz. But this reduction, despite understanding, still causes some
distortions in the digitized signal. Therefore, the necessity to analyze the performance
of speech-to-text systems in the context of telephone calls is justified by comparing
the conversion error rates of different API.

3 Methodology

Ethical approval for the experiment in this study was deemed unnecessary by the
Ethics Committee. As also the activities in the experiment do not pose risks greater
than those ordinarily encountered in daily life and all the subjects remained anony-
mous. As well as all participants were given consent and were also informed about the
academic objective of the research as well as the guarantee of anonymity giving the
choice of participation or not. Still considering that as it is an opinion poll, according
to Resolutions 466 and 510 of the National Health Council, there is no need to
go through the University’s Ethics Committee. This is following CNS Resolution
466/12 of the National Health Council, of December 12, 2012, (available at https://
conselho.saude.gov.br/resolucoes/2012/Reso466.pdf). Idem with Article 1 of Reso-
lution 510, of 7 April 2016, of Brazil’s National Health Council (available at http://
www.conselho.saude.gov.br/resolucoes/2016/Reso510.pdf) and also with item 8.0.5
of the Ethical Principles of Psychologists and Code of Conduct of the American
Psychological Association (available at https://www.apa.org/ethics/code/principles.
pdf).
It is a descriptive study of exploratory research. Because it sought a better under-
standing of the topic of voice-to-text conversion in order to provide a better overview
and guide research delimitation and objectives. This is a descriptive study because
a systematic data collection was performed, recording, analyzing, classifying and
interpreting the observed phenomena without, however, at any time, any interference
from the researcher.
The audio recordings were performed from six different Brazilian speakers (three
females and three males) with a “divine comedy” standard text. The speakers have
a postgraduate degree and age between 40 and 50 years. The call recordings were
performed using some APIs for telephony and text messaging services that enable the
recording of the telephone conversation between a VoIP client (an acronym for Voice
over IP) and the destine telephone. After finalized calling, the audio was downloaded
through the link provided by the web service accessed via the call token + id [8].
Then the recorded audios were treated and converted using three different APIs for
voice-to-text conversion: AWS, Google, and IBM.
440 N. M. Iinuma and M. de O. Igarashi

It is important to observe that the recordings were made using the same equipment
and the same telecommunications provider so that the results of the experiments could
be comparable.

4 Results and Discussion

4.1 Record Preparation

Initially, it is necessary to configure recorder software (Fig. 1).

To configure, it is important to insert the caller and the destination numbers, the
date and consecutively activate the start of recording.

4.2 Treatment of Recordings

Consequently, the recording audio process has some steps for a correct voice-to-text
conversion.
Figure 2 shows the workflow required to handle the audio signal before its
conversion:

Fig. 1 API documentation

Phone Calls Speech-to-Text: A Comparison … 441

Fig. 2 Workflow for audio processing and conversion

All steps described in Fig. 2 are performed through the Audacity® software [9].
Each party’s voice is separated by simply copying and pasting the recorded audio
signal.
Then, the elimination of speechless periods was performed by clicking on the
“effects” menu > ”lock silence” option (Fig. 3). The program detects as “silence”
the configured thresholds. In this case, it was configured a −20 dB threshold with
0.2 s duration for silent parameters and 0.2 for discard section parameters.
The callers can have different tones and voice volumes. Thus, it is important
to introduce a gain in the signal to equalizes the audio quality and ensures good
conversion. In the “effects” menu the “amplify” option was used. In this option, the
software already provides a dB gain suggestion that best applies to audio.
Finally, noise must be removed using the same software. By clicking on the
“effects” > ”noise reduction” menu, it is possible to detect the “noise profile” and
reduce it in the entire audio signal.

4.3 Voice-to-Text Conversion API

The experiments were performed by using the following APIs:

• Google Cloud Speech to Text [10]
• IBM Watson Speech to Text [11]
• Amazon Transcribe [12].
442 N. M. Iinuma and M. de O. Igarashi

Fig. 3 Filters applied in the lock silence option

4.4 Word Error Rates

Word Error Rate (WER) was used to compare the efficiency of speech-to-text API
algorithms. WER is a value that indicates the transcription error rate. Thus, the lowest
value means a better quality conversion [13].
WER is calculated based on the formula (1) described by Koo Alison [14]:

S+I +D
WER = (1)
NWS
• S = Substitutions. These are errors which some word is exchanged to another;
• I = Insertions. These are unspoken words that are added. It may, for example, be
a single word that has been described as more than one word;
• D = Deletions. These are missing words in the transcript.
• NWS= Number of Words Spoken. The number of pronounced words.

4.5 Analysis of Voice-to-Text Conversion Results

Figure 4 presents the WER relationship separating the conversion results by

interlocutor and the types of APIs used.
Phone Calls Speech-to-Text: A Comparison … 443

Fig. 4 WER values relating the caller and the API used

Figure 4 shows that person 2 presented the lowest WER value in all APIs while
the other persons presented a variation of WER value in all experiments in a uniform
way. Therefore, person 2 has a positive influence on the efficiency of voice-to-text
conversion. It is also noteworthy that person 2 is a university teacher, so this fact may
also have possibly influenced the result due to their greater care with pronunciation.
Figure 5 presents the same WER results, but now filtered by API type, where the
WER was listed for each party.
Regarding the conversion using the AWS API, WER ranged from 11.52 to 22.32%
for people 1, 3, 4, 5 and 6 (Fig. 5); and between 5.13 and 6.45% for person 2.

Fig. 5 WER values relating to API and interlocutors

444 N. M. Iinuma and M. de O. Igarashi

Fig. 6 WER médiuns for APIs tested

Figure 6 shows the average WER values calculated based on the APIS conversion
results used.
From the average WER values obtained in Fig. 7, it can be seen that the Google
API presented the lowest average WER value (14.90%), thus, the highest efficiency
of all APIs. However, the value is very close to the value obtained by the IBM API
(15.23%) and slightly less than the value of the AWS API (16.5%); it demonstrates
that despite differences at values, APIs have close conversion errors and thus similar
efficiencies. Noteworthy in this case is the average WER value obtained by the AWS
API (16.50%), which was approximately 10.7% higher than Google’s average WER,
i.e., it was virtually 10.7% less efficient than Google’s and 7.69% less efficient than
IBM.

5 Conclusions and Future Works

The WER analysis considering different persons in the experiments shows a higher
conversion efficiency of the Google API, followed by the IBM API and lastly the
AWS API. However, it is important to highlight that these conclusions refer to 18
recordings samples in total and a small sample of people (just 6 interlocutors). There-
fore, it is recommended a larger number of experiments and people samples for future
experiments and more assertive conclusions.
It follows that the transcript was closer to the original text in the transcript using
the Google API (GCP), which had the lowest Word Error Rate—WER. However,
the WER difference in error rate was not very significant (10.7% between AWS and
GCP and 2.21% between IBM and GCP).
Phone Calls Speech-to-Text: A Comparison … 445

References

1. Scheller, F., Capelas, B.: Inteligência Artificial toma conta de call centers (2019). https://
link.estadao.com.br/noticias/inovacao,inteligencia-artificial-toma-conta-de-call-centers,700
03023316
2. Krancher, O., et al.: Key affordances of platform-as-a-service: self-organization and continuous
feedback. J. Manage. Inf. Sys. (Bern–Switzerland) (2018). https://doi.org/10.1080/07421222.
2018.1481636
3. Rhoton, J.: Cloud Computing Explained: Implementation Handbook for Enterprises. Recursive
Press (2009)
4. Lakomkin, E., et al.: KT-Speech-Crawler: Automatic Dataset Construction for Speech
Recognition from YouTube Videos. University of Hamburg, Hamburg, Germany (2019)
5. Moreno, A., Redondo, T.: Text Analytics: the Convergence of Big Data and Artificial Intelli-
gence. Universidad Autónoma de Madrid and Instituto de Ingeniería del Conocimiento, Madrid
Espanha (2016)
6. Pinheiro, P.: Tutoriais Telefonia Fixa - Ciclos Evolutivos: O Início da Digitalização (2004).
https://www.teleco.com.br/tutoriais/tutorialciclos/pagina_4.asp
7. Haykin, S.S., et al.: Sinais e sistemas. Bookman, Porto Alegre (2001)
8. Total Voice: A startup que se reinventou para crescer seis vezes em um ano (2019). https://sci
nova.com.br/total-voice-startup-que-se-reinventou-para-crescer-seis-vezes-em-um-ano/
9. AUDACITY: Audacity (2019). https://www.audacityteam.org
10. GCP: Google Cloud Speech to Text (2019). https://cloud.google.com/speech-to-text
11. IBM. Watson Speech to Text (2019). https://speech-to-text-demo.ng.bluemix.net
12. AWS: Amazon Transcribe (2019). https://docs.aws.amazon.com/pt_br/transcribe/latest/dg/
how-it-works.html
13. Popovic, M., Hremann, N.: Word Error Rates: Decomposition over POS Classes and Applica-
tions for Error Analysis. University Aachen, Germany (2002). http://citeseerx.ist.psu.edu/vie
wdoc/download?doi=10.1.1.310.4832&rep=rep1&type=pdf
14. Koo, A.: How to Calculate Word Error Rate. rev.ai (2019). https://www.rev.ai/blog/how-to-cal
culate-word-error-rate/
Production of Cantonese Lexical Tones
by Native Speakers of Brazilian
Portuguese: A Comparative Analysis

João Vítor Possamai de Menezes and Adriano Vilela Barbosa

Abstract This study presents a quantitative analysis of the production of lexical tone
by speakers of a non-tonal language. More specifically, we look at how well native
speakers of Brazilian Portuguese are able to produce the six tones of Cantonese. An
experiment was conducted where six subjects listened to a number of tone words
previously produced by a native speaker of Cantonese. For each word they heard, the
subjects were asked to try to reproduce it as best as they could. No further instructions
were provided. A total of 150 tokens per subject were produced (5 phonetic strings ×
6 tones × 5 repetitions). The quality of the non-native productions was assessed by
comparing them to the productions of the native speaker. This was done by extracting
the F0 contours from all recorded tokens (both native and non-native), approximating
them by polynomials (first, second, and third order) and then using the polynomial
parameters to group the tokens into six classes corresponding to the six lexical tones
of Cantonese. The analysis was performed in two ways: (1) by using an F0 onset ×
offset graph already proposed in the literature and (2) through a classification task
using a Linear Discriminant Analysis (LDA). In both cases, each subject was analyzed
separately. The F0 onset × offset graphs presented differences between native and
non-native data; the non-native productions showed great variability while the native
productions were better defined, presenting less variability. Classifiers trained on the
Cantonese native speaker’s data resulted in accuracies above 90%, whereas those
trained on the Brazilian Portuguese speakers’ data resulted in accuracies between 40
and 60%.

J. V. P. de Menezes · A. V. Barbosa (B)

Graduate Program in Electrical Engineering, Universidade Federal de Minas Gerais, Av. Antônio
Carlos 6627, Belo Horizonte, MG 31270-901, Brazil
e-mail: [email protected]
J. V. P. de Menezes
e-mail: [email protected]
A. V. Barbosa
Department of Electronic Engineering, Universidade Federal de Minas Gerais, Av. Antônio Carlos
6627, Belo Horizonte, MG 31270-901, Brazil

© The Editor(s) (if applicable) and The Author(s), under exclusive license 447
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_45
448 J. V. P. de Menezes and A. V. Barbosa

Keywords Lexical tone · Non-native production · Discriminant analysis ·

Cantonese · Brazilian Portuguese

1 Introduction

About 70% of all known languages are tonal [1]. These languages are characterized by
the use of lexical tones, which is the use of pitch to distinguish lexical or grammatical
meaning [1]. Tonal languages are spread over East Asia [2], Africa [3], the Americas
[4], and other regions [5].
The differences in meaning between fundamental frequency (F0), pitch, and tone
are important in this work. F0 is an acoustic term defined as the physiological
frequency of vibration of the vocal folds. Pitch is a psychoacoustic term related
to the perception of F0. Thus, F0 and pitch are terms associated with the production
and the perception of speech, respectively. Although high F0 values are associated
with a high pitch, and low F0 values are associated with a low pitch, the relation
between them is not linear. In turn, a tone is a linguistic unit, expressed in terms of
F0 and perceived in terms of pitch that manifests itself as lexical tone within words
and as intonation within sentences. In intonation, tones are responsible for associ-
ating different meanings to sentences, whereas lexical tones, in tone languages, are
responsible for contrasting different words [1]. As an example, the syllable [yau] in
Cantonese means “worry” when uttered with a high-level tone, and it means “paint”
when uttered with a high-rising tone [1].
Tone language learning and perception are recurrent research subjects [6, 7],
mainly due to the significant number of speakers and learners of these languages.
Many of these learners are speakers of non-tonal languages who have to familiarize
themselves with the structure of the new language and learn how to produce and
perceive lexical tones. Since the ability to produce sounds on a target language is
crucial to enable communication between native and non-native speakers, non-native
lexical tone production is being researched in several tone languages, such as Thai
[8] and Mandarin [9], and also in pitch-accent languages such as Norwegian [10]
and Swedish [11].
This work focuses on Cantonese, a tonal language with six lexical tones that are
generally described by their corresponding F0 contours. Figure 1 shows a simple
representation of these contours as straight lines [12], whereas Fig. 2 shows a more
accurate representation based on F0 tracking of recorded data [13]. Based on these
figures, the six lexical tones can be classified as level, if the F0 remains practically
constant throughout its production, or contour, if the F0 changes as the tones are
being produced. In Cantonese, tones 1, 3, 6 (T1, T3, and T6) are level tones whereas
tones 2, 4, 5 (T2, T4, and T5) are contour tones.
Previous studies about non-native production of lexical tone and pitch accent
have focused on different aspects of production, such as accentedness ratings [8], the
improvement in production after perception training [9], and the ability to produce
sounds distinguishable by native speakers [10]. General results suggest that non-
native speakers have more difficulty in producing level tones than contour tones [8],
that the perception affects the production of lexical tones [9], and that the ability to
Production of Cantonese Lexical Tones by Native Speakers … 449

produce F0 contours and levels are not mastered in parallel, with the level tones being
more challenging than the contour tones [10]. From these results, we hypothesize that
non-native speakers in our experiment will (1) encounter difficulties in producing
native-like lexical tones, because they will not benefit from any previous perceptual
training and (2) encounter more difficulty in the production of Cantonese level tones
(T1, T3, and T6) than in the production of contour tones (T2, T4, and T5).

Fig. 1 F0 contours of the six lexical tones of Cantonese. They are here drawn didactically, focusing
on the initial and end frequencies of each tone. Adapted from [12]

Fig. 2 F0 contours of the six lexical tones of Cantonese. They are here the mean contours of 198
native productions [13]
450 J. V. P. de Menezes and A. V. Barbosa

This work presents two main contributions: (1) the use of Brazilian Portuguese
speakers, with no familiarity with Cantonese, as subjects and (2) the use of discrim-
inant analysis [14] in order to assess the ability of these speakers to produce the six
lexical tones of Cantonese. We also look to reproduce the F0 onset × offset graphs
proposed by Barry and Blamey [15] with the Brazilian Portuguese speakers.
The objective of this work is to compare native and non-native production
of Cantonese lexical tones. More specifically, we compare the production of a native
speaker of Cantonese (henceforth CN speaker) to those of six native speakers of
Brazilian Portuguese (henceforth BP speakers). Some of the questions we try to
address in this work are: (1) How do BP speakers perceive and produce Cantonese
tone words? (2) Do these speakers have more difficulty in the production of level
tones or of contour tones? (3) Are they able to produce all the six tones equally well?
If not, which tones are more challenging to them?
The rest of this paper is organized as follows. Section 2 describes the data acquisi-
tion process for the CN speaker and the BP speakers. Section 3 presents the techniques
used in the analysis. Section 4 presents and discusses our results, and Sect. 5 presents
the conclusions.

2 Materials

The Cantonese data was provided by collaborators (see the acknowledgment section)
and only used as a stimulus in this study. The data consists of 30 tone words (five
phonetic strings bearing each of the six lexical tones) produced by a female, 24-
year-old CN speaker. The phonetic strings used were: “fan,” “fu,” “hau,” “soej,” and
“wai.” Each tone word was repeated five times, resulting in 150 tokens.
The Brazilian Portuguese data was recorded in a speech production experiment.
Before the experiment, 1 the subjects were given an information sheet explaining
them, among other things, the objectives of the study and also informing them about
the voluntary and anonymous nature of their participation. The subjects then signed
a consent form agreeing with the experiment terms and proceeded to the actual data
collection. All subjects were native speakers of Brazilian Portuguese (three females,
three males) ranging from 20 to 37 years. Some of them showed a background in
music (2) and knowledge of foreign languages (6) such as English, Spanish, French,
or Italian. None of them had any previous knowledge of Cantonese or any other tone
language. The recording took place in an acoustically treated sound booth, using a
MacBook Pro (Mid 2014). The 150 original Cantonese stimuli (tone words) were

1 ethical approval for the experiment in this study was deemed unnecessary by the Ethics Committee

of the Federal University of Minas Gerais, since the activities in the experiment do not pose risks
greater than those ordinarily encountered in daily life and all the subjects remained anonymous.
This is in accordance with Article 1 of Resolution 510, of April 7th 2016, of Brazil’s National
Health Council (available at http://www.conselho.saude.gov.br/resolucoes/2016/Reso510.pdf) and
also with item 8.0.5 of the Ethical Principles of Psychologists and Code of Conduct of the American
Psychological Association (available at https://www.apa.org/ethics/code/principles.pdf).
Production of Cantonese Lexical Tones by Native Speakers … 451

randomly presented to each subject, one stimulus at a time. For each word they
heard, the subjects were asked to try to reproduce it as best as they could. No further
instructions were provided; the subjects did not even know that what they were
listening to were Cantonese words. One recording was made for each subject and
for each Cantonese audio token, resulting in 150 tokens produced by each subject
(150 × 6 = 900 non-native tokens in total). The subjects were allowed to pause at
any time.

3 Methods

F0 contours were extracted from each token using Praat [16]. The rest of the analysis
was performed in R [17]. Initially, the F0 contours exported by Praat were imported
into R by means of the rPraat library [18]. The imported F0 contours were then
approximated by polynomials of different orders (first, second, and third).
The polynomial orders were chosen based on the F0 contours shown in Figs. 1
and 2. A first-order polynomial would be sufficient to characterize Cantonese lexical
tones as displayed in Fig. 1, but not the lexical tones as displayed in Fig. 2. Given
that the data used here consists of natural recordings, its F0 contours should be
more similar to those in Fig. 2, which would be better approximated by higher-order
polynomials, such as second and third order. Figures 3 and 4 show the F0 contours
for the CN speaker and for the BP speaker six upon which our analysis was based.
In this paper, we also compare results obtained with each polynomial approximation
order, possibly suggesting an optimal polynomial order for Cantonese F0 contour
approximations.

Fig. 3 F0 contours produced by the CN speaker for the word [fu] with all of the six lexical tones
452 J. V. P. de Menezes and A. V. Barbosa

Fig. 4 F0 contours produced by the BP speaker 6 for the word [fu] with all of the six lexical tones

Finally, each F0 contour was represented by 2, 3, or 4 polynomial coefficients.

Based on these coefficients, a linear classification task was carried out using Linear
Discriminant Analysis (LDA).

3.1 F0 Onset × Offset Graph

As a theoretical basis for their proposal of a F0 onset × offset graph, Barry and Blamey
[15] cite the five acoustic dimensions that define tone differences form a listener’s
perspective: (a) average pitch, (b) direction, (c) length, (d) extreme endpoint, and
(e) slope [19]. All of those dimensions, except for length, can be captured by such
a graph. The onset was defined as the first point of the F0 contour and the offset as
the endpoint, so they are able to define the frequency at which the contour starts the
frequency at which it ends.
In the F0 onset × offset graph, each repetition is a point, generating a cloud
of points. To measure how clear the speaker acoustically differentiates each lexical
tone, Barry and Blamey [15] proposed two indices (I 1 and I 2 ) that can be calculated
based on ellipses centered at each tone’s mean value and with axes as wide as the
standard deviation of each tone repetitions. The area of each ellipse illustrates how
consistent the speaker is for each tone: The smaller the ellipse, the more consistent
is the speaker.
Production of Cantonese Lexical Tones by Native Speakers … 453

The first index (I 1 ) is based on the three more distinguishable tones, T1 (high
level), T2 (high rise), and T4 (low falling), and on the concept of tonal space. Being
T1, T2, and T4 normally at the outer boundaries of the graph, the tonal space is
defined as the area of the triangle whose vertices are the mean values of T1, T2, and
T4. Index 1, described in (1), is then calculated as the reason between the area of the
tonal space triangle (At ) and the area of T1, T2, and T4 ellipses (A1,2,4 ). The higher
its value, the greater the chance of overlapping between tones.

I1 = At /A1,2,4 (1)

The second index (I 2 ) is based on all of the six tone ellipses, giving it an advantage
over Index 1. It measures how separate they are from one another and how homo-
geneous they are among themselves. This index, described in (2), is calculated as
a reason between the average distance between the center of every ellipse to each
other (Davg ) and the average of the sum of the radius of every ellipse (S avg ).

I2 = Davg /Savg (2)

3.2 Classification Task

The classification task consists of a Linear Discriminant Analysis, which is a method

used both for dimensionality reduction and classification. As a classifier, the LDA
tries to define linear separators between classes, whose distributions are estimated as
normal and is popular when the classification task consists of more than two classes
[14].
Some of our previous studies with the same native Cantonese lexical tone data
showed that native-produced lexical tones can be classified based on their F0 contour
almost error-free (95% of accuracy for third-order polynomial approximation). In this
direction, each subject (one native CN speaker and six non-native BP speakers) will
have its own samples classified separately to verify how accurate the classification
is. The benchmark is the classification accuracy achieved for the CN speaker, and
the closer to it the BP speaker’s classification accuracy is, the better their lexical tone
production is.
Firstly, a baseline accuracy value was obtained for each subject using a single
LDA. After that, more meaningful values were obtained by performing twelve five-
fold cross-validation runs and taking the mean and standard deviation of the 60 (12
runs × 5 folds) accuracy values.
454 J. V. P. de Menezes and A. V. Barbosa

4 Results

The results come from the F0 onset × offset analysis and the classification task. For
each of them, results for every approximation are presented (first, second, and third
order).

4.1 F0 Onset × Offset Graph

Figures 5, 6, and 7 show the F0 onset × offset graph for the CN speaker with
polynomial approximations of third, second, and first order, respectively. The figures
show similar separation between, and variation within, tones. According to Barry
and Blamey’s results [15], T1, T2, and T4 should lie at the outer edges of the graph,
defining the tonal space, whereas T3, T5, and T6 should lie within this space, tending
to overlap with each other. This was also observed in the graphs for the CN speaker,
in Figs. 5, 6, and 7.
Figures 8 and 9 show the F0 onset × offset graphs for two BP speakers using
the third-order polynomial approximations. Figure 8 (BP speaker 6) shows a graph
similar to the CN speaker’s, with a large and well-defined tonal space with differences
within tones smaller than differences across tones. On the other hand, the graph shown
in Fig. 9 (BP speaker 5) is an example of poorly defined tones with large overlaps
between the ellipses. Some instances of T1, T2, and T4 are well separated from the
others, but the ellipses associated with T3, T5, and T6 significantly overlap. Figures 8
and 9 highlight how different the ability of BP speakers to produce Cantonese lexical
tones may be.

Fig. 5 F0 onset × offset graph based on the F0 contour approximated by a third-order polynomial.
Subject: CN speaker
Production of Cantonese Lexical Tones by Native Speakers … 455

Fig. 6 F0 onset × offset graph based on the F0 contour approximated by a second-order polynomial.
Subject: CN speaker

Fig. 7 F0 onset × offset graph based on the F0 contour approximated by a first-order polynomial.
Subject: CN speaker
456 J. V. P. de Menezes and A. V. Barbosa

Fig. 8 F0 onset × offset graph based on the F0 contour approximated by a third-order polynomial.
Subject: BP speaker 6

Fig. 9 F0 onset × offset graph based on the F0 contour approximated by a third-order polynomial.
Subject: BP speaker 5
Production of Cantonese Lexical Tones by Native Speakers … 457

Fig. 10 Index 1 × Index 2 graph. The results from the first-order approximation are in red, from
the second-order approximation are in yellow, and from the third-order approximation are in blue

An I 1 × I 2 plot is shown in Fig. 10. The subjects in this figure may be divided
into two groups based on index I 1 : those with values smaller than 1 and those with
values larger than 1. Separability based on I 2 is not as clear. From the definition of the
indices, a well-defined tonal space means a higher value for I 1 and good separation
between tones means a higher value for I 2 . These characteristics can, however, have
effects on both indices. Apart from the CN speaker, BP speakers 4 and 6 showed
higher values for I 1 , whereas BP speakers 1, 4, and 6 showed higher values for I 2 .
This suggests that their productions of Cantonese lexical tones are more similar to a
native’s than the production of the other subjects.
One effect that stands out in Fig. 10 is the difference in the indices between
the different order polynomial approximations from the CN speaker. They form
almost a straight line, with the first-order approximation at the bottom end and the
second-order approximation at the upper end. This may suggest that a second-order
polynomial approximation is more appropriate for the F0 onset × offset graph.

4.2 Classification Task

The classification analysis attempted to measure the similarity of non-native and

native productions of Cantonese lexical tone. The benchmark is the linear classifi-
cation result of native production, which resulted in accuracy above 90% in some of
our previous studies with the same data.
458 J. V. P. de Menezes and A. V. Barbosa

Table 1 Baseline accuracies for each subject

Single LDA
Subject First order Second order Third order
Accuracy Accuracy Accuracy
BP 1 0.53 0.67 0.47
BP 2 0.47 0.43 0.50
BP 3 0.43 0.53 0.33
BP 4 0.63 0.67 0.60
BP 5 0.43 0.50 0.43
BP 6 0.50 0.50 0.47
CN 0.83 0.90 0.83

The baseline accuracy values obtained with single LDA for the three polynomial
approximations are shown in Table 1. Aside from the CN speaker, with accuracies
near 90%, BP speakers 1, 4, and 6 achieved the best overall classification accuracies,
with accuracies ranging from 47 to 67%.
In order to obtain a more robust measure of how good the lexical tone production
of each BP speaker is, we performed twelve fivefold cross-validation runs using the
same LDA classifier. The results are shown in Table 2. According to the baseline
results from a single LDA, the CN speaker had the best results, around 90%, and the
BP speakers 1, 4, and 6 performed better than BP speakers 2, 3, and 5. All of the
subjects, however, showed above chance (16.67%) accuracy rates. The CN speaker
also showed the lowest Standard Deviation (SD) value of all, suggesting productions
with smaller variability.

Table 2 Accuracies for each subject using 12 trials of fivefold cross-validation

LDA with 12 trials of fivefold cross-validation
Subject First order Second order Third order
Accuracy SD Accuracy SD Accuracy SD
BP 1 0.51 0.094 0.57 0.092 0.55 0.084
BP 2 0.45 0.059 0.43 0.075 0.43 0.075
BP 3 0.39 0.075 0.40 0.074 0.40 0.083
BP 4 0.66 0.070 0.66 0.083 0.66 0.070
BP 5 0.42 0.087 0.43 0.071 0.45 0.090
BP 6 0.53 0.082 0.53 0.096 0.54 0.083
CN 0.91 0.051 0.90 0.050 0.89 0.053
Production of Cantonese Lexical Tones by Native Speakers … 459

Table 3 Confusion matrix for all of the subjects using third-order polynomial approximation
All subjects Reference
Third order 1 2 3 4 5 6
Prediction 1 27 4 9 0 3 3
2 1 22 1 0 5 2
3 4 1 12 1 7 8
4 1 0 2 25 2 6
5 0 8 3 1 15 5
6 2 0 8 8 3 11
Bold values represent correct predictions

To also measure how well each of the tones was classified, confusion matrices for
all of the productions of each subject, as well as for all of them together, were created.
The general confusion matrix, shown in Table 3, presents the predictions made for
the 210 test samples (20% of the total samples) using a third-order polynomial
approximation. The tones more accurately predicted were T1, T2, and T4, the tones
which delimit the tonal space [12]. That can be explained by the relative proximity
between T3, T5, and T6 and the fact that T1, T2, and T4 stand commonly apart from
each other, with no big superposition with any other tone, as shown by the graphs in
Figs. 1, 2, 3, 4, and 5.
Since, from the results in Table 2, BP speaker 3 achieved the lowest and BP
speaker 4 achieved the highest accuracy among non-native speakers, we will present
their confusion matrices, as well as the confusion matrix for the CN speaker. The
trend shown in Table 3 that T1, T2, and T4 are better classified than the other tones
can also be observed for the BP speakers in Tables 4 and 5. The CN speaker, on the
other hand, achieved an almost perfect classification, as shown in Table 6.
According to Table 2, the classification task did not present significantly different
results when the polynomial order changed. This suggests that the classification of
F0 contours have similar accuracies in 2, 3, and 4 dimensions.

Table 4 Confusion matrix for BP speaker 3 using third-order polynomial approximation

BP speaker 3 Reference
Third order 1 2 3 4 5 6
Prediction 1 2 1 2 0 0 0
2 0 3 0 0 1 0
3 1 0 1 0 2 0
4 1 0 0 2 0 1
5 0 0 2 0 2 4
6 1 0 0 0 3 0
Bold values represent correct predictions
460 J. V. P. de Menezes and A. V. Barbosa

Table 5 Confusion matrix for BP speaker 4 using third-order polynomial approximation

BP speaker 4 Reference
Third order 1 2 3 4 5 6
Prediction 1 5 0 0 0 0 0
2 0 5 1 0 1 1
3 0 0 0 0 0 3
4 0 0 1 4 1 0
5 0 0 0 0 2 0
6 0 0 3 1 1 1
Bold values represent correct predictions

Table 6 Confusion matrix for the Cantonese native subject using third-order polynomial
approximation
CN speaker Reference
Third order 1 2 3 4 5 6
Prediction 1 5 0 0 0 0 0
2 0 4 0 0 0 0
3 0 0 5 0 0 2
4 0 0 0 5 0 0
5 0 1 0 0 5 0
6 0 0 0 0 0 3
Bold values represent correct predictions

5 Conclusion

The goal of this work was to compare the production of Cantonese lexical tones
by speakers unfamiliar with that language with those produced by a native speaker
of the language. The BP speakers had no knowledge of tonal language structure, a
condition meant to prevent speakers from making any conscious effort to produce
specific lexical tones, for they faced each Cantonese tone word as unknown audio.
The reproduction of the F0 onset × offset graphs proposed in [15] with the data
we collected showed that BP speaker’s productions presented high variability within
and across subjects when compared to the CN speaker’s productions. This means (1)
that some BP speakers’ F0 contours produced graphs quite different from other BP
speakers and (2) that the F0 contours produced by some BP speakers were not like
the graphs expected by Barry and Blamey [15], showing significant overlap between
all of the tones.
Results from the classification task showed that the BP speakers’ accuracy was in
the range between 39 and 66%, much below the accuracy of the native CN speaker
(between 89 and 91%). The accuracy was interpreted here as a measure of how good
Production of Cantonese Lexical Tones by Native Speakers … 461

a non-native speaker could produce lexical tones: the closer they were to the native’s
accuracy, the better.
Comparisons between the classification task results and the F0 onset × offset
graph (and its related I 1 and I 2 indices) showed that the subjects with best results
were BP speakers 1, 4, and 6, the same subjects with higher values of I 2 . Also, the
tones which were more accurately classified were T1, T2, and T4, which are the
boundaries for the tonal space defined in the F0 onset × offset graph.
Our results point, therefore, in the direction that BP speakers can produce some
lexical tones better than others (namely T1, T2, and T4). This may, however, be
biased by our method, which was a linear classifier. A perception task could be used
to determine how good the BP speakers’ productions are based on the native listener’s
perspectives.
Regarding the effect of polynomial order on the analysis, the F0 onset × offset
related indices produced similar results across orders, except for the CN speaker,
which showed different index values for different polynomial orders. An effect of
the polynomial order on the classification task could not be perceived.
An additional result is that the male BP speakers (1, 4, and 6) performed better
than the female BP speakers (2, 3, and 5) both in the F0 onset × offset indices and in
the classification task. However, further tests with a larger number of subjects should
be carried out in order to establish a significant relationship between the gender of
the subjects and of the native speaker.
Next steps in the analysis of BP speakers’ production of Cantonese lexical tones
can be (1) the recording of more native data (only a female speaker was used in this
experiment); (2) the recording of more non-native speakers; (3) the execution of a
perception stage, which would require native Cantonese listeners to judge the non-
native lexical tone productions and (4) the usage of more methods, such as non-linear
classifiers.

Acknowledgements Support for this work was provided by CAPES (Coordination for the Improve-
ment of Higher Education Personnel, Brazil) to João Vítor Possamai de Menezes and by FAPEMIG
(Fundação de Amparo à Pesquisa do Estado de Minas Gerais, Brazil) to Adriano Vilela Barbosa
(process APQ-03701-16). The Cantonese data was kindly provided by Dr. Denis Burnham (the
MARCS Institute at the Western Sydney University, Australia).

References

1. Yip, M.: Tone, 1st edn. Cambridge University Press, Cambridge (2002)
2. Yip, M.: The tonal phonology of Chinese. Ph.D. Dissertation, MIT (1980)
3. Odden, D.: Tone shift and spread in Taita I. Stud. Afr. Linguist. 30(1), 75–110 (2001)
4. Soares, M.F.: Traços acústicos das vogais em Tükuna. Cadernos de Estudos Linguísticos 7,
137–175 (1984)
5. Maddieson, I.: Tone. In: Dryer, M.S., Haspelmath, M. (eds.) The World Atlas of Language
Structures Online. Max Planck Institute for Evolutionary Anthropology, Leipzig (2013)
462 J. V. P. de Menezes and A. V. Barbosa

6. Burnham, D., Ciocca, V., Stokes, S.: Auditory-visual perception of lexical tone. In:
EUROSPEECH 2001 Scandinavia, 7th European Conference on Speech Communication and
Technology, 2nd INTERSPEECH Event, Aalborg, Denmark, 3–7 September (2001)
7. Lee, K.Y.S., Chan, K.T.Y., Lam, J.H.S., van Hasselt, C.A., Tong, M.C.F.: Lexical tone
perception in native speakers of Cantonese. Int. J. Speech-Lang. Pathol. 17(1), 53–62 (2015)
8. Wayland, R.: Non-native production of Thai: acoustic measurements and accentedness, ratings.
Appl. Linguist. 18(3), 345–373 (1997)
9. Wang, Y., Jongman, A., Sereno, J.A.: Acoustic and perceptual evaluation of Mandarin tone
productions before and after perceptual training. J. Acoust. Soc. Am. 113(2), 1033–1043 (2003)
10. Steien, G.B., van Dommelen, W.A.: The production of Norwegian tones by multilingual non-
native speakers. Int. J. Bilingualism 22(3), 316–329 (2016)
11. Kaiser, R.: Do Germans produce and perceive the Swedish word accent contrast? A cross-
language analysis. TMH-QPSR 51(1), 93–96 (2011)
12. Cheng, S.T.T., Lam, G.Y.H., To, C.K.S.: Pitch perception in tone language-speaking adults
with and without autism spectrum disorders. I-Perception 8(3), 1–15 (2017)
13. Wong, P., Leung, C.T.: Suprasegmental features are not acquired early: perception and produc-
tion of monosyllabic cantonese lexical tones in 4- to 6-year-old preschool children. J. Speech
Lang. Hearing Res. 61(5), 1070–1085 (2018)
14. James, G., Witten, D., Hastie, T., Tbishirani, R.: An introduction to statistical learning: with
applications in R, 1st edn. Springer, New York (2013)
15. Barry, J.G., Blamey, P.J.: The acoustical analysis of tone differentiation as a means for assessing
tone production in speakers od Cantonese. J. Acoust. Soc. Am. 116(3), 1739–1748 (2004)
16. Boersma, P., Weenink, D.: Praat: doing phonetics by computer [Computer Program]. Version
6.1.03. Retrieved 1 Sept 2019 from http://www.praat.org/ (2019)
17. R Core Team: A Language and Environment for Statistical Computing. R Foundation for
Statistical Computing, Vienna, Austria (2019)
18. Bořil, T., Skarnitzl, R.: Tools rPraat and mPraat. In. Sojka, P., Horák, A., Kopeček, I., Pala, K.
(eds.) Text, Speech and Dialogue. Springer (2016)
19. Gandour, J.T.: The perception of tone. In: Fromkin, V.A. (ed.) Tone: A Linguistics Survey, 1st
edn. Academic Press, New York (1978)
Smart Campus IoT Guidance System
for Visitors Based on Bayesian Filters

Alvaro Aspilcueta Narvaez , Dennis Núñez Fernández ,

Segundo Gamarra Quispe , and Domingo Lazo Ochoa

Abstract In this work, we proposed an indoor location system that makes use of a
Raspberry Pi embedded computer and WiFi signals to guide a person inside a region
of the faculty of Electrical and Electronic at Universidad Nacional de Ingeniería,
Peru. The main advantage with similar indoor location systems like beacons or RFID
technology is that the presented system does not require additional hardware since it
makes use of the pre-installed WiFi routers. The experimental tests show promising
results, achieving a location accuracy of 92.31%.

Keywords Indoor location · WiFi · Bayes filter · Embedded computer

1 Introduction

The existing navigation systems like Global Positioning System (GPS) offer a precise
location in outside environments but an inaccurate location inside common buildings
[1]. This problem of localization has attracted the interest of researchers and devel-
opers due to the high demand for such systems for indoor navigation, immersive
experiences, asset tracking, augmented reality, and more.
During the last decade, several approaches for indoor location systems [2] such
as radio frequency identification (RFID), wireless local area networks (WLAN), and
Bluetooth among others have been proposed [3]. In [4, 5], an indoor positioning

A. Aspilcueta Narvaez · D. Núñez Fernández (B) · S. Gamarra Quispe · D. Lazo Ochoa

Universidad Nacional de Ingeniería, Lima 15333, Peru
e-mail: [email protected]
A. Aspilcueta Narvaez
e-mail: [email protected]
S. Gamarra Quispe
e-mail: [email protected]
D. Lazo Ochoa
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 463
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_46
464 A. Aspilcueta Narvaez et al.

system for smart buildings was used to obtain relatively good results; however, the
system makes use of RFID technology, which produces an additional cost for hard-
ware implementation. Regarding indoor location using WiFi signals, [4] makes use
of Monte Carlo (MC) filter for a precise WiFi-based indoor localization, but the
system is intended for tracking objects with a high precision employing relative
high computing processing, which is not suitable for real-time applications since
it demands high computing resources and a small division of the regions. In recent
work, [6, 7] proposes a novel, incremental approach that reduces the energy consump-
tion of WiFi localization by scanning just a few selected channels. The results are
remarkable and provide a way for the implementation of indoor location on embedded
systems. In addition, Indoor Google [8] provides guidance; however, its accuracy is
not enough (5–15 m).
In this paper, we propose an indoor location system for a region of the faculty
of Electrical and Electronic of Universidad Nacional de Ingeniería, Peru. Unlike
the indoor location systems presented before, which rely on external devices like
RFID hardware or Bluetooth beacons, the proposed system does not demand external
equipment or additional costs. Therefore, employing an embedded computer and pre-
installed access points, our indoor location system is able to provide a good position
accuracy and the guidance to the destination region.

2 Proposed Method

2.1 Overview

Our system performs indoor localization and guidance on an embedded computer

using WiFi signal levels. In this way, the system works on small embedded devices
with low computational resources. With the aim to achieve response time and power
computational constraints, a simple and effective Bayes recursive estimator was used
in this work.
The methodology used for this work has two phases: offline and online phase. At
the offline phase, the WiFi levels were captured several times in every region, thus
obtaining a received signal strength indicator (RSSI) table per region. Later, the RSSI
data were divided into training and testing datasets in order to perform the evaluation.
Then, the normalized histograms are cleaned using the Gaussian distribution. Finally,
the localization system is evaluated offline using average accuracy and confusion
matrix. At the online phase, the Bayesian filter was implemented in a Raspberry Pi
embedded computer. This filter uses the captured WiFi signal levels and the cleaned
RSSI database to predict the most likely region. Figure 1 shows the diagram of the
proposed methodology.

2.2 Data Recollection

First, the principal regions of the faculty of Electrical and Electronic of Universidad
Nacional de Ingeniería are identified and divided into several areas. Figure 2 shows
Smart Campus IoT Guidance System … 465

Fig. 1 Scheme of the proposed interior locating system

Fig. 2 Sampling points and destination in the third-floor scheme

such a distribution of regions. After division, the main WiFi information like service
set identifier (SSID), media access control (MAC) and RSSI is recollected via a
custom application in for the Raspberry Pi. Thereby, 150 samples in each region
at different times during three non-consecutive days/nights were captured. Finally,
filtering processing for non-stable access points (APs) was carried out, so the final
dataset was constructed using only the best APs in order to have the best results.
466 A. Aspilcueta Narvaez et al.

2.3 Bayesian Filter

Several relevant papers about RSSI-based indoor location systems assume Gaussian
probability density function (PDF) [9, 10]. It is justified by the relation of the noise of
the radio receiver together to the influence of average white Gaussian radio channel
noise, which is modeled by a Gaussian PDF. Therefore, original histograms could be
cleaned by approaching each histogram to a Gaussian PDF. In Eq. (1), the formulation
of a Gaussian PDF is presented. It is described by two parameters: μ (mean) and σ
(standard deviation).

1 (x−μ)2
f (x|μ, σ 2 ) = √ e− 2σ 2 (1)
2π σ 2

In order to achieve response time and computational requirements, a straightfor-

ward and effective estimator is used, and it is based on the Bayes recursive estimator.
This is able to infer the posterior using sensed and prior knowledge (see Eq. (2)).
Being A the event we want the probability, and B the new evidence that is related to
A. Therefore, the posterior P(A|B) is calculated by the likelihood P(B|A) (probability
of observing the new evidence) and the prior P(A) (probability of our hypothesis
without any additional prior information). P(B) is the marginal likelihood, which is
the total evidence probability.

P(B|A)P( A)
P(A|B) = (2)
P(B)

The algorithm presented below corresponds to the implementation of the Bayes-

based estimator. This takes the AP tables and current WiFi measurement as inputs and
returns the estimated region and its probability. Thereby, the algorithm recursively
calculates the probability of the posterior region (line 11), and then, the probability
(line 12) and predicted region (line 13) are calculated. Algorithm 1 was implemented
in Python for evaluation of the performance of the Bayes estimator and for inference
on the Raspberry Pi platform.
Smart Campus IoT Guidance System … 467

2.4 Guidance System

The flowchart of the operation of our guidance system is shown in Fig. 3. The
start screen allows the manual entry of the destination point through a drop-down
menu, based on the plane of sampling points shown in Fig. 2. Once selected, the
user pushes the central button in order to start the guidance system. The destination
region is displayed on the screen.
Then, the system calculates the current node where the user is located, which is the
returned result of the indoor location algorithm based on the Bayes filter, see Fig. 3.
Based on the current node and the destination node initially entered, the graph of the
nodes is expressed in an adjacency matrix of the distances and angles between the
nodes (based on the plane of the faculty facilities). Then, the algorithm of Dijkstra
calculates the shortest route and returns the next node to go, which is based on the
minimum route.
From the provided map of the facilities, the orientation with respect to the magnetic
north is also shown, and the use of a magnetometer was added to indicate the orien-
tation in degrees from the magnetic north; so, it works like a compass, no matter the
direction of the user it will always orient to the north direction. In order to have the
same direction, it was previously calibrated. Some considerations were taken such
as a small variation because the screen generates a magnetic field, see Fig. 4.
468 A. Aspilcueta Narvaez et al.

Fig. 3 Flowchart of the smart campus guidance system program

Smart Campus IoT Guidance System … 469

(a) Location of the magnetometer (b) Calibration of magnetometer

Fig. 4 Hardware and calibration related to the magnetometer

Using the angle with respect to the magnetic north, and the angle between the
current node and the next node, which was determined by the Dijkstra algorithm, the
direction of the guiding compass is independent of the direction of the user to the
next node, which follows the minimum route to reach its destination.
When the user arrives at their final destination, the system finishes working and
the next short message will appear on the screen of the Raspberry Pi board: “You
have arrived at your destination!!!”.

3 Experimental Results

In order to obtain the performance and other metrics for the indoor localization
system, we evaluated it on the testing dataset. The overall accuracy and the confusion
matrix (see Fig. 5) define the results. This matrix for the proposed system shows high
accuracy for the diagonal values and low accuracy for the others, and it means that the
system has a high accuracy per class/region. In addition, the accuracy of the indoor
location system is about 92.31%.
For the evaluation of the system, we use a Raspberry Pi model 3B+ , an LCD 3.5
screen, an MPU9250 sensor, and an external battery. It was conducted in different
regions and several positions. As expected, the system correctly predicted the loca-
tions in all regions. Figure 6 shows some results of the test. As we can see, our system
correctly predicts the location.
470 A. Aspilcueta Narvaez et al.

Fig. 5 Confusion matrix for the proposed system

4 Conclusions

This paper introduced the implementation of a Smart Campus WiFi-based indoor

location system to guide visitors in a region of the faculty of Electrical and Elec-
tronic of Universidad Nacional de Ingeniería, Peru. With the aim to accomplish a
high positioning accuracy and a low computational power consumption, the proposed
system makes use of Bayes recursive estimator. Experiments show promising results,
obtaining an accuracy of 92.31%. As further work, the processing could be performed
in remote servers in order to have a better response time. Thus, the proposed method-
ology is not limited to this work, but can also be applied to similar localization tasks
such as immersive experiences, augmented reality, and asset tracking, among others.
Smart Campus IoT Guidance System … 471

Fig. 6 Online results on the Raspberry Pi. a Result in region 3.10. b Result at region 3.1

Acknowledgements We thank Ph.D. Marco Zúñiga Zamalloa for their time and inspiring teaching.
This work was funded by the Research Institute of the Faculty of Electrical and Electronic
Engineering (IIFIEE UNI).

References

1. Motte, H., Wyffels, J., De Strycker, L., Goemaere, J.-P.: Evaluating GPS data in indoor environ-
ments. Adv. Electr. Comput. Eng. 11(3), 25–28 (2011). https://doi.org/10.4316/AECE.2011.
03004
2. Sakpere, W., Adeyeye Oshin, M., Mlitwa, N. B.: A state-of-the-art survey of indoor positioning
and navigation systems and technologies. S. Afr. Comput. J. 29(3), 145–197 (2017)
3. Brena, R. F., García-Vázquez, J. P., et al.: Evolution of indoor positioning technologies: a
survey. J. Sensors 2017, 21 (2017)
472 A. Aspilcueta Narvaez et al.

4. Dhital, A., Closas, P., Fernández-Prades, C.: Bayesian filters for indoor localization using
wireless sensor networks. In: 2010 5th ESA Workshop on Satellite Navigation Technologies and
European Workshop on GNSS Signals and Signal Processing (NAVITEC), pp. 1–7. Noordwijk
(2010)
5. Moreno-Cano, M. V., Zamora-Izquierdo, M. A., Santa, J., Skarmeta, A. F.: An indoor local-
ization system based on artificial neural networks and particle filters applied to intelligent
buildings. Neurocomputing 122, 116–125 (December 2013)
6. Brouwers, N., Zuniga, M., Langendoen, K.: Incremental Wi-Fi scanning for energy-efficient
localization. In: 2014 IEEE International Conference on Pervasive Computing and Communi-
cations (PerCom), pp. 156–162. Budapest (2014)
7. Lin, K., Chen, M., Deng, J., Hassan, M. M., Fortino, G.: Enhanced fingerprinting and trajectory
prediction for IoT localization in smart buildings. IEEE Transact. Autom. Sci. Eng. 13(3),
1294–1307
8. Indoor Google Maps. (Online Web, accessed on January 2014). Available: http://maps.google.
com/help/maps/indoormaps/
9. Chruszczyk, L.: Statistical Analysis of Indoor RSSI Read-outs for 433 MHz, 868 MHz, 2.4 GHz
and 5 GHz ISM Bands. Int. J. Electr. Telecommun. 63(1), 33–38 (2017)
10. Kaji, K., Kawaguchi, N.: Design and implementation of WiFi indoor localization based on
Gaussian mixture model and particle filter. In: 2012 International Conference on Indoor
Positioning and Indoor Navigation (IPIN), pp. 1–9. Sydney, NSW (2012)
Research of the Academic Performance
of University Students Through
Statistical Models

Olga Solano Dávila , Luis Nuñez Lira , Olga Bolaños Solano ,

and Néstor Mamani-Macedo

Abstract In this work, there were analyzed some factors, such as demographic
variables, study habits, and academic stress, which influence academic performance,
especially in students of the faculty of mathematical sciences of a public univer-
sity. A cross-sectional, descriptive, and explanatory study was used from a statistical
approach, which was made up of 311 students. The data were collected during the first
semester of 2019 using the academic stress inventory, the vicuna study habits inven-
tory, and the demographic variables, age, gender, professional school, and socioeco-
nomic level; subsequently, the weighted average of the students was obtained. The
results show that variables that influence academic performance are: age, profes-
sional school, study habits, and the dimension reactions to academic stress and the
variables that do not influence performance are: gender, socioeconomic level, level
of stress, stressors, and coping strategies for academic stress.

Keywords Academic performance · Demographic variables · Study habits ·

Academic stress

O. Solano Dávila (B) · N. Mamani-Macedo

Facultad de Ciencias Matemáticas, Universidad Nacional Mayor de San Marcos, Lima, Peru
e-mail: [email protected]
N. Mamani-Macedo
e-mail: [email protected]
L. Nuñez Lira
Facultad de Educación, Universidad Nacional Mayor de San Marcos, Lima, Peru
e-mail: [email protected]
O. Bolaños Solano
Facultad de Psicología, Universidad Nacional Mayor de San Marcos, Lima, Peru
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 473
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_47
474 O. Solano Dávila et al.

1 Introduction

Over the years, experts warn that education in Latin America has the worst academic
performance. The statistical data of the countries state that the region is below the
general patterns of academic achievement. Countries like Peru, Colombia, Brazil,
and Argentina are among the lowest 10th percentile whose students have a lower
performance in the disciplines of math, science, and reading [1].
A study habit is a researched behavioral standard that is automatically observed
in the face of frequently determined circumstances on a regular basis, where the
subject does not have to meditate on how to proceed. Vicuña proposes as fundamental
aspects of learning habits, method, ordering, time, and distractibility [2]. According to
academic stress, it is considered a methodical procedure, of an adaptive nature, mainly
psychological, which is observed (a) at the time the student is in academic settings
and has to make a list of requirements to be evaluated, and these are considered
stressors (entrance); (b) when these produce methodical disorder (overwhelming
state) it reveals a list of signals (signs of instability); and (c) when this instability
pressures the student to do coping activities (exit) to renew systemic stabilization [3].
The aim of this paper is to show a statistical model to explain academic performance
based on academic stress, study habits, and sociodemographic variables.

2 State of the Art

Several researchers have found that there is an association between study habits and
university achievement [4, 5]. This means that the more convenient the habits of
learning, the better the academic achievement. However, students do not have study
habits, a plan to study, and certain care has their scope of study and the motivation
with their studies not to disapprove of the enrolled courses. These results should be
applied a program of improvement of the habit of learning, taking into account the
organization of tests, work, planning of your time, attitude, physical condition, and
the issues associated with the place of study.
From other points of view, there are findings that indicate that there is no asso-
ciation between the habit of studying with academic achievement and moral values
among students. This means that students have good learning habits and more moral
values; and that the influence of academic achievement is due to factors such as
the community, family, economic environment, the cognitive-affective and sensitive
structures of each student [6]. However, in the process of training students, not only
the cost of studying and their techniques support an appropriate development of study,
but other causes of the particular individual such as dignity, self-representation, self-
judgment, own capacity, incentives, social capacities, balance and management of the
sensitive universe, as well as parental, governmental or physiological causes (student
health) can have a positive or negative impact on the task of learning [7]. Related
references on academic stress indicate that this affects the academic achievement of
Research of the Academic Performance … 475

students [8] and is strongly associated with the perceived physical and mental health
dimensions [9].
Other stress-related researches have not been able to verify associations with other
variables, in addition to academic performance [10]. The main stressors in students
are the excess of homework and university work, as well as the short time to perform
them [10]; other investigations have found that stressors are considered most relevant
examinations, contradictory approach work by teachers, inconsistent approach work
by teachers, and lack of practical to theoretical classes [11]. For [12], the most
affecting stressors in students are teacher methodological deficiencies, examinations,
and interventions in public. In [13], the author concluded that students are stressed by
the fact of competing with classmates, academic overload, teacher character, exams,
the tasks requested by teachers, limited time to do homework and not understand
well the topics analyzed in class. On the other hand, most students suffer stress
during the academic year, and academic, psychosocial, and environmental stressors
are associated with perceived stress [14].

3 Methodology

The quantitative, descriptive, and explanatory approach was used because this design
intends to explain academic performance based on study habits, academic stress,
and sociodemographic variables in students of the Faculty of Mathematical Sciences
(FMS) of a higher university center [15].
The authorization of the director of the Administrative Direction of the FMS of
a public university was obtained to carry out the research. The students signed an
informed consent before applying the instruments used in the research, the Vice-
Dean’s Office was also asked to list of enrolled students from the third cycle to the
tenth cycle, and the head of enrollment provided this information in a Microsoft
Excel file.

3.1 Population Sampling

The universe in this study is constituted by students from the third to the tenth cycle
enrolled in the FMS of a public university on semester 2019-I, and the unit of anal-
ysis is each student with the characteristics mentioned above. Random probabilistic
sampling was used, separated by groups according to the characteristic of the profes-
sional school, proportional to the percentage of students in four schools [16]. Each
professional school was used as a stratum, in total four, which are: mathematics,
operation research, statistics, and computing science. The selection of students for
the study was random in each school. According to this sampling scheme, the sample
size was calculated, with a significance level of 5%, a margin of error of 4.47%, the
proportion of students with moderate academic stress level in each of the professional
476 O. Solano Dávila et al.

school and historical information is provided by the registration office of the FMS
[16], and the sample size was 311 students.

3.2 Data Collection Techniques and Instruments, Validity,

and Reliability

The general characteristics were considered as age, gender, professional school,

socioeconomic level, then the inventory of study habits was considered, whose
purpose is to recognize in the student the appropriate and inappropriate habits
expressed at the time of studying and, finally, the academic stress inventory that aims
to identify the stress attributes that undergraduate and graduate students experienced
during their university education.
The instrument that was used was divided into three parts: (1) general character-
istics of the student; (2) inventory of study habits [2]; and (3) inventory of academic
stress [17]. The length of the application of the instruments was from 15 to 20 min,
which was carried out during the month of June of this year. To perform the internal
consistency reliability analysis for this inventory of academic stress, Cronbach’s
Alpha was used. It was found that the value of this coefficient for the full scale was
0.89. For the validity of the instrument construct, the confirmatory factor analysis
(CFA) was used [18], and the goodness-of-fit indexes were: goodness-of-fit index
(GFI) = 0.964, adjusted goodness-of-fit index (AGFI) = 0.958, normed fit index
(NFI) = 0.946, and relative of fit index (RFI) = 0.942 superiors to 0.90, so it is
confirmed that the proposed model is adequate, with the three dimensions of the
inventory of academic stress: symptoms, stressors, and coping strategies. Regarding
the reliability of the study habits scale, it was considered KR-20 reliability, reliability
of the entire scale was 0.74, which means that reliability is acceptable [19, 20].

4 Results

For the logistic regression analysis [21], we use the academic performance variables,
0: approved and 1: disapproved, as a dependent variable and the explanatory variables
are study habits, academic stress, stress level, and sociodemographic variables: age,
gender, professional school, and socioeconomic level. The explanatory variable β is
the logistic regression model parameter, Wald is the test statistic, which evaluates
the significance of each parameter, D f degrees of freedom, the p − value is the
probability associated with the test statistic which also evaluates the significance of
each one of the parameters, exp(β) is the exponentiation of the β coefficient, and
(IC) at 95% exp(β) is the interval confidence for exp(β) at 95% confidence.
When analyzing the study habits variable, which was considered as categorical,
the category (positive level) was taken as a reference and we can observe that the
Research of the Academic Performance … 477

positive value of β for the first category (trend level(−)) indicates that in students
who had this type of study habits increases the possibility of disapproving courses
in relation to those who had positive study habits. Students who have very negative
or negative study habits are 4.477 times more likely to disapprove courses (Exp
(β = 4.477)).
For the variable ‘professional school,’ four professional schools were considered:
computer science, statistics, operations research, and mathematics. Mathematics
school was used as a reference in order to calculate the odds ratio with respect
to the other professional schools. In an attempt to reflect certain patterns, it is veri-
fied that students of the professional school of statistics are 8.40 times more likely
to disapprove courses compared to students from other professional schools (Exp
(β = 8.4)).
As a result, Table 1 shows the variables that influence academic performance and
these are: age, professional school, study habits, and the dimension of reactions to
academic stress, at a 5% level of significance. On the other hand, Table 2 shows the
variables that do not influence performance that are gender, socioeconomic level,
stress level (mild, moderate, deep), stressors, and coping strategies for academic
stress, at a 5% level of significance ( p −value greater than 0.05 level of significance).

Table 1 An estimated logistic regression model with the significant variables of the academic
performance of FMS students of a public university—Semester 2019-I
Variable B β (odds)a Wald Df p − value exp(β) IC 95% exp(β)
Lower Higher
Age −0.055 8.200 1 0.004 0.946 0.911 0.983
Professional 39.715 3 0.000
schoolb
Professional 0.899 6.553 1 0.010 2.458 1.235 4.892
school (1)
Professional 2.128 27.835 1 0.000 8.400 3.810 18.520
school (2)
Professional 1.627 25.452 1 0.000 5.087 2.704 9.569
school (3)
Study habitsc 10.825 3 0.013
Study habits 1.499 5.363 1 0.021 4.477 1.259 15.917
(1)
Study habits 0.379 0.658 1 0.417 1.461 0.584 3.651
(2)
Study habits −0.189 0.128 1 0.721 0.828 0.294 2.330
(3)
Reactions to 0.026 4.542 1 0.033 1.026 1.002 1.050
stress
a Odds value regards mathematics school
b Professional school (1): computing science, professional school (2): statistics, and professional
school (3): operations research
c Study habits (1): Very negative or negative. Study habits (2): trend (−). Study habits (3): trend (+)
478 O. Solano Dávila et al.

Table 2 An estimated logistic regression model with the non-significant variables of the academic
performance of FMS students of a public university—Semester 2019-I
Variable Wald Df p − value
Sex 0.184 1 0.668
Socioeconomic levela 3.147 3 0.370
Socioeconomic level (1) 1.258 1 0.262
Socioeconomic level (2) 0.718 1 0.397
socioeconomic level (3) 1.903 1 0.168
Stressors 1.916 1 0.166
Coping strategies 0.001 1 0.976
Stress levelb 0.819 2 0.664
Stress level (1) 0.212 1 0.645
Stress level (2) 0.777 1 0.378
a Socioeconomic level (1): high, socioeconomic level (2): medium and socioeconomic level (3): low
b Stress level (1): mild and stress level (2): moderate

The logistic regression model to predict the probability if a student is approved

in its weighted average the Semester 2019-I is the following:

1
π̂ = , 0 ≤ π̂ ≤ 1 (1)
1 + e−z

where

z = −0.055 Age + 0.899 School(1) + 2.128 School(2)

+ 1.627 School(3) + 1.499 Study Habits(1) + 0.379 Study Habits(2)
− 0.189 Study Habits(3) + 0.026 Reactions

The probability of a student approved is when π̂ is greater than 0.5 and less than
one and the probability of a student disapproved is when π̂ is greater than zero and
less to 0.5.

5 Discussion

The results show that age, professional school, study habits, and reactions to academic
stress influence academic performance. This is consistent with other researchers
whose findings indicate associations, but lack study habits, a study plan, implying
the necessity to apply improvement plans in attitudes-time planning-physical
development-motivation [22–24].
Research of the Academic Performance … 479

It is important to note that students having study habits generate greater academic
performance because they generally manage their time well because they have atti-
tudes and aptitudes for cognitive development and the desire to excel. On the other
hand, stress generates the stimulation of stressors, generating physiological and
psychological responses, and diseases can be generated. Other studies also agree
that study habits and university achievement are associated, where students obtained
good academic performance [23, 24].
Training processes in education not only consist of having study habits but also,
it is important to have knowledge of the development of techniques and technolo-
gies that require the development of skills in students. The statement by [8] is also
true that academic stress causes poor performance. And this stress is generated not
only by the excess of tasks but the evaluations, incoherent works, or lack of theoret-
ical–practical classes. On the other hand, in the study of [25], it refers to multiple
causes of academic performance, and it indicates that it is caused by the interaction
of multiple social, personal, institutional-academic factors that differ from a group
of subjects. In addition, the author also believes that there are significant differ-
ences between university satisfaction, the index of study habits, levels of support in
university students, satisfaction with teachers, and the socioeconomic level. Like-
wise, according to [26], the self-esteem variable influences academic performance.
In our study, stress reactions or symptoms influence academic performance, and this
is due to the fact that 77.2% of students have moderate to deep stress, also, in physical
reactions, students state that they are drowsy or greater need to sleep, with respect
to psychological reactions they express that they have restlessness (inability to relax
and be calm), they also affirm that they have concentration problems, in light of these
results, academic stress is present in the student.

6 Conclusions and Recommendations

In this work, we present a statistical model that explains the academic performance
according to the variables’ study habits, academic stress, and demographic variables,
in this sense, we conclude that the variables that influence academic performance
are: age, professional school, study habits, and reactions to academic stress, and
the variables that do not influence performance are: gender, socioeconomic level,
stress level, stressors, and coping strategies for academic stress ( p − value greater
than 0.05 level of significance). It is recommended that the FMS authorities design
appropriate policies to help students who experience academic stress so that they can
receive adequate treatment at the university clinic and improve their mental health.
Also, the authorities may develop programs for students who have low academic
performance so that they are taught the management and practice of study habits.
480 O. Solano Dávila et al.

References

1. OCDE, Pisa 2015: Resultados Clave (2016). https://www.oecd.org/pisa/pisa-2015-results-in-

focus-ESP.pdf. Last accessed 14 Oct 2019
2. Vicuña, L.: Inventario de hábitos de estudio. CEDES, Lima (1998)
3. Barraza, A.: Un modelo conceptual para el estudio del estrés académico. Revista Electrónica de
Psicología Iztacala 9(3), 110–129 (2006). http://www.psicologiacientifica.com. Last accessed
11 Oct 2019
4. Atsiaya, E., Maiyo, J.K.: Study of the relationship between study habits and academic achieve-
ment of students: a case of Spicer Higher Secondary School, India. Int. J. Educ. Adm. Policy
Stud. 7(7), 134–141 (2015). https://doi.org/10.5897/IJEAPS2015.0404
5. Garbanzo, G.: Factores asociados al rendimiento académico en estudiantes universitarios, una
reflexión desde la calidad de la educación superior pública. Educación 31(1), 43–63 (2007).
https://www.redalyc.org/pdf/440/44031103.pdf. último acceso 15 Oct 2019
6. Mondragón, C. M., Cardoso, D., Bobadilla, S.: Hábitos de estudio y rendimiento académico.
Caso estudiantes de la licenciatura en Administración de la Unidad Académica Profesional
Tejupilco, 2017. Revista Iberoamericana para la Investigación y el Desarrollo Educativo. 8(15),
1–25 (2018). http://dx.doi.org/10.23913/ride.v8i15.315
7. Enriquez, M. F., Fajardo, M., Garzón, F.: Una revisión general a los hábitos y técnicas de estudio
en el ámbito universitario. Psicogente 18(33), 166–187 (2015). http://doi.org/10.17081/psico.
18.33.64
8. Alfonso, B., Calcines, M., Monteagudo, R., Nieves, Z.: Estrés académico. Edumecentro 7(2):
163–178 (2015). http://scielo.sld.cu/pdf/edu/v7n2/edu13215.pdf. Last accessed 15 Oct 2019
9. Chau, C., Vilela, P.: Variables asociadas a la salud física y mental percibida en estudiantes
universitarios de Lima. Liberabit 23(1), 89–102 (2017). https://doi.org/10.24265/liberabit.
2017.v23n1.06
10. García, A., Del Toro, A. Y., Cisneros, E., Querts, O., Cascaret, X.: Algunas variables
psicosociales asociadas al bajo rendimiento académico en estudiantes de primer año de
medicina. Medisan 21(4), 433–439 (2017). http://scielo.sld.cu/scielo.php?script=sci_arttext&
pid=S1029-30192017000400007. Last accessed 15 Oct 2019
11. Cano, S. T., Ramos, J. A., Medina, M. G.: Análisis del estrés académico en estudiantes de
ingeniería como estrategia para el aprendizaje significativo. ANFEI Digital 2(5), 1–8 (2016).
http://www.anfei.org.mx/revista/index.php/revista/article/view/280/921. Last accessed 15 Oct
2019
12. Zárate, N. E.: Estrés académico en estudiantes universitarios: medidas preventivas. Revista de
la Alta tecnología y la Sociedad 9(4), 92–8 (2017). http://www.academiajournals.com/revista-
alta-tec-y-sociedad. Last accessed 15 Oct 2019
13. Rodríguez, B., González, M. D. P., Blanco, L. E.: Estresores académicos percibidos por estu-
diantes pertenecientes a la escuela de enfermería de Ávila, centro adscrita la Universidad de
Salamanca. Revista Enfermería C y L. 6(2), 98–105 (2014). http://www.revistaenfermeriacyl.
com/index.php/revistaenfermeriacyl/article/view/131/104. Last accessed 15 Oct 2019
14. Zárate-Depraect, N. E., Soto-Decuir, M. G., Martínez-Aguirre, E. G., Castro-Castro, M. L.,
García-Jau, R. A., López-Leyva, N. M.: Hábitos de estudio y estrés en estudiantes del área de
la salud. FEM: Revista de la Fundación Educación Médica 21(3), 153–57 (2018). http://sci
elo.isciii.es/scielo.php?script=sci_arttext&pid=S2014-98322018000300007. Last accessed 15
Oct 2019
15. Khaleef, F. F., Saud, H., Nasser, M. F., Saliman, R. T., Elhaj, A.: Stress and it is impact among
medical students in Hail University, KSA. Int. J. Adv. Res. 6(4), 1381–1387 (2018). http://dx.
doi.org/10.21474/IJAR01/6999
16. Hernández, T., Mendoza, C.P.: Metodología de la Investigación, 6a edn. McGraw Hill, México
(2018)
17. Scheaffer, R., Mendenhall, W., Ott, L.: Elementos de muestreo, 6ª edn. Paraninfo Cengage
Learning, Madrid (2007)
Research of the Academic Performance … 481

18. Barraza, A.: Propiedades psicométricas del Inventario SISCO del estrés académico. Psicologia-
Cientifica.com. 9(10) (2007). http://www.psicologiacientifica.com/sisco-propiedades-psicom
etricas/. Last accessed 15 Oct 2019
19. Aldas, J., Uriel, E.: Análisis multivariante aplicado con R, 2ª edn. Editorial Paraninfo, Madrid
(2017)
20. Kuder, G.F., Richardson, M.W.: The theory of the estimation of test reliability. Psychometrika
2(1), 151–160 (1937)
21. Hosmer, D. W., Lemeshow, S., Sturdivant, R. X.: Applied Logistic Regression. New Jersey,
Wiley (2013)
22. Chilca, L.: Autoestima, hábitos de estudio y rendimiento académico en estudiantes univer-
sitarios. Propósitos y Representaciones 1(5), 71–127. (2017). http://dx.doi.org/10.20511/pyr
2017.v5n1.145
23. Espinoza, J. L.: Relación entre los hábitos de estudio y el rendimiento Académico de los
alumnos de escuelas profesionales acreditadas. Revista Científica Tzhoecoen 9(4), 29–40.
(2017). https://doi.org/10.26495/rtzh179.423933
24. Gilavand, A.: Investigating the study habits among students of iranian medical sciences univer-
sities: a review. Indo Am. J. Pharm. Sci. 4(12), 4734–4738 (2017). https://doi.org/10.5281/zen
odo.1133164
25. Sreelekha, V., Indla, Y. R., Reddy, R., Rameswarudu, M., Swathi, A., Yamini, D., et al.: Study
habits and academic performance of first year MBBS students. Int. J. Med. Sci. Public Health
5(9), 1831–1835 (2016). https://go.galegroup.com/ps/anonymous?id=GALE%7CA466520
071&sid=googleScholar&v=2.1&it=r&linkaccess=abs&issn=23204664&p=HRCA&sw=w.
Last accessed 15 Oct 2019
26. Nicho J.: Relación entre la autoestima y el rendimiento académico de los estudiantes de la e.b.c.
Tecnológica de la Facultad de Educación de la Universidad Nacional José Faustino Sánchez
Carrión. Universidad Nacional, Perú 2(5), 606–615 (2013)
Model of Sustainable Digital
Transformation Focused
on Organizational and Technological
Culture for Academic Management
in Public Higher Education

Jessie Bravo , Janet Aquino , Roger Alarcón , and Nilton Germán

Abstract Currently, public universities in Peru are undergoing changes due to new
market requirements, aspects that regulatory entities took into account to rethink their
operational requirements, therefore, it is necessary that their processes are adapted
and comply with these new characteristics regulated according to law, especially
academics who relate directly to teachers and students, these processes must be
transformed taking into account emerging technologies. To solve this problem, a
sustainable digital transformation model was developed, consisting of four dimen-
sions: infrastructure, business architecture, digitalization, and people. In each dimen-
sion, indicators were used to assess at what stage of maturity it is found, which by
integrating them allowed us to know the level of general maturity of the digital
transformation of the institution in which our case study was applied to the effects
of knowing the digital divide. This model was validated in a national university of
Peru, specifically in the process of managing the digital teaching portfolio.

Keywords Digital transformation · Academic management · Maturity level

J. Bravo (B) · J. Aquino · R. Alarcón · N. Germán

School of Computer and Informatics Engineering, Pedro Ruiz Gallo National University,
Lambayeque, Peru
e-mail: [email protected]
J. Aquino
e-mail: [email protected]
R. Alarcón
e-mail: [email protected]
N. Germán
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 483
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_48
484 J. Bravo et al.

1 Introduction

At present, public universities in Peru under the licensing context have led them
to adapt to the new requirements of both infrastructure and technology, so they
are looking for ways to improve their academic services, reduce costs and achieve
the basic quality conditions. These conditions, in addition to being an indispens-
able requirement for the functioning of universities, will allow students to receive an
educational service according to the needs and expectations of the student population
and society in general. On the other hand, with the information and communications
technologies present in the market, which provide tools to be competitive, it is neces-
sary to create new scenarios adapted to the measures and needs of public universities.
The purpose of this study is to provide a model of sustainable digital transforma-
tion in order to solve the needs of public universities in Peru with the main axis,
the person, through the measurement of their digital and organizational maturity,
allowing digitalization processes streamline management and improve the quality of
the educational service.

2 State of the Art

It can be stated that according to [1] digital transformation in the public sector means
new ways of working with stakeholders, building new service delivery frameworks
and creating forms of relationships. Digital transformation is mainly associated with
the need to use new technologies to remain competitive in the Internet era, where
services and products are delivered both online and offline. Others define digital
transformation as a way to rebuild business models following the needs of customers
through the use of new technologies [2].
In the research conducted by Bernhard [3] on the current state of digitalization of
adult, continuing education in Switzerland and Germany identifies that information
technology infrastructure, staff development, and management/leadership are the
most relevant challenges to implement digital media in continuing education orga-
nizations for adults. According to [4], digital transformation processes must have
an innovative model for how to work with students with the aim of improving the
scope of teaching. Similarly, [5] consider that digital transformation introduces a
new business model through the implementation of new business logic to create and
capture value, it affects the entire company and its ways of doing business. Jackson
[6] examines problems that have always needy ambidexterity, also of innovation
and considering the past trajectories that had caused problems in education. For
Aguiar et al. [7] the multidimensional environment of the modern university must be
considered as something fundamental to address the innovation needs demanded by
society, the bearer of the largest commission. Without a doubt, the progress made by
the Internet makes it the main protagonist of the great technological revolution of the
twenty-first century. Its inescapable presence has constituted it as a great scenario in
which new ways of learning, thinking, communicating, doing and acting take place.
Model of Sustainable Digital Transformation … 485

According to [8] digital transformation in higher education was referenced in

seven dimensions, governance dimension was included in the process on its own,
differing in that our proposal is based only on four dimensions and a transversal
activity that allows the monitoring and tracking of this entire transformation process.
In [9] a conducted research on cultural diversity and new communication
ecosystem of the emerging digital culture are analyzed in the discourses that guide
the Information and Communication Technologies (ICT) policies in education in the
cities of Brazil, Argentina and Chile.

3 Proposal

Based on a review of the literature, a sustainable digital transformation model for

academic management in public universities is proposed, where, according to the
exploratory analysis performed, it is evident that the personal factor is relevant to the
success of the digital transformation required in the university. Our model integrates
aspects of organizational culture and digital culture. The proposed model focuses
on four dimensions: person, business architecture, digitalization, supported in the
infrastructure dimension as shown in Fig. 1.
• Person Dimension (P): It is one of the most important dimensions public higher
education, it is observed that the main limiting factor of an adequate digital trans-
formation is the person, so it is necessary to measure the digital culture (DC) and
organizational culture (OC) of the people involved, thus referring to teachers, as
the main agent of change. Regarding organizational maturity, key aspects such
as institutional identification, soft skills and organizational climate have been
considered, since this will allow recognizing limitations and defining organiza-
tional change strategies. The maturity model of digital culture takes into account
three dimensions: the use of technology, digital skills and communication, and
the Internet. Within this digital ecosystem, different profiles of users have been
found, which present totally different habits.
• Infrastructure Dimension (I): Every public university must have the physical
infrastructure, which includes classrooms, laboratories, workshops, administra-
tive offices, leisure spaces, and sports fields. The technological infrastructure
such as the Internet connection, cloud computing, servers, multimedia projectors,
smart boards and a security system which includes aspects such as surveillance,
safe areas, gauging, drills, use of fire extinguishers and security protocols are
considered.
• Business Architecture Dimension (BA): This dimension has been divided into
two levels: user and manager. The user-level includes the processes that interact
directly with students, teachers, and graduates. At the manager level, it has been
subdivided into directive processes and academic management processes.
• Digitalization Dimension (D): Based on the business model dimension, it has
been divided into three levels: service platform digitalization, the directive
digitalization level and the academic digitalization.
486 J. Bravo et al.

Fig. 1 Model of sustainable digital transformation

For the calculation of the score in each dimension:

n
Scorecriteria
ScoreDimension = i=1 (1)
n
For the calculation of the model score:
nd
j=1 ScoreDimension
ScoreModel = (2)
nd
Tables 1 and 2 describe the proposed maturity levels for the person dimension.
For the remaining dimensions (infrastructure, business architecture, and digital-
ization) the maturity levels are defined according to Table 3.
Model of Sustainable Digital Transformation … 487

Table 1 Maturity levels regard organizational culture

The maturity level of organizational culture General score Average score
Indifferent 0–15 0–5
Interested 16–29 6–9
Involved 30–44 10–15
Committed 45–60 16–20

Table 2 Maturity level regards digital culture

The maturity level of digital culture General score Average score
Primitive 0–17 0–5
Basic 18–34 6–11
Intermediate 35–51 12–17
Advanced 52–68 18–23

Table 3 Maturity levels by infrastructure, business architecture, and digitalization

Dimension Infrastructure Business architecture Digitalization
General score by maturity levels
Incipient 0–8 0–8 0–8
Basic 9–16 9–16 9–16
Managed 17–25 17–25 17–25
Optimized 26–34 26–34 26–34

Finally, the level of maturity (LM) of the institution’s digital transformation model
was evaluated by separately first the dimensions of the person and second the three
remaining dimensions, infrastructure, business architecture, and digitalization. The
formulas to apply for our digital transformation model are the following:

LMTD = Index I + IndexBA + Index D (3)

LM P = IndexOC + IndexDC (4)

4 Validation

To validate the sustainable digital transformation model presented, we applied a

case study to show that the proposal successfully resolves the application of ICT
in the academic management of university higher education. Figure 2 shows the
methodologies chosen for each dimension in the process of digitalization of the
teaching portfolio of a national university.
488 J. Bravo et al.

Fig. 2 Result of the application of the model of sustainable digital transformation

• Person dimension: For this dimension, two surveys were applied, one to measure
the organizational culture and the other the digital culture. Regarding the organi-
zational culture, 216 surveys were applied, this was validated regarding its content
and reliability using the Cronbach Alpha obtaining a good result (0.812). It was
observed that for 35 people its level is interested, which represents 43.52%, in
terms of institutional identification, instead, 190 people are committed, it repre-
sents 87.96% in the management of soft skills and 130 people are at the level
involved (60.19%) in relation to the organizational climate. For the digital culture,
228 surveys were applied, this was validated regarding its content and reliability
using the Cronbach Alpha obtaining an acceptable result (0.799), from these
results it can be observed that referring to the digital culture in the university 153
people (67.11%) have an intermediate maturity level and 46 people (20.18%) have
an advanced maturity level that will allow us to successfully face the proposed
technological changes.
Model of Sustainable Digital Transformation … 489

• Business architecture dimension: This dimension includes two levels, for the
student user, their relationship is indirect through the tutoring service provided
by teachers and socialization of the syllabus. The teaching user is the main entity
that will allow him to manage his academic activity related to the activities:
teaching, research, tutoring and social responsibility. Regarding the second level:
manager, it will be the responsibility of the director of the department and part of
the director of the school. The department director is responsible for generating
and controlling the digital teaching portfolio every academic year. The school
directors are responsible for the academic follow-up, the teaching portfolio being
an important tool for such activity.
• Digitalization dimension: For this case study, academic digitalization has been
implemented through mandatory use of institutional mail, digital portfolio digi-
talization using the institutional virtual platform such as the classroom tool. The
activities carried out by the school director and department director that are
currently digitized, related to the teaching digital portfolio process, represent
66.7%.
To determine the level of maturity based on the three dimensions of infrastructure,
business architecture and digitalization, based on the indicators expressed in Table 3,
the following was obtained:

LMTD = 18.22 + 25.0 + 22.23 = 65.46 (5)

It can be seen that the university maturity level of digital transformation in the
teaching digital portfolio process is at the managed-level, but there are still improve-
ments to be made that will be achieved thanks to the monitoring and tracking based
on the strategies mentioned in the following section to achieve the optimized level.
Regarding the person dimension, the organizational and digital culture was eval-
uated, obtaining the following results, organizationally the teachers are involved and
in terms of the digital aspect, their level is intermediate.

LMOC = 74.07 (6)

LMDC = 67.11 (7)

4.1 Tracking and Monitoring

The purpose we propose is that our model allows for sustainability, this means that
the University has the capacity to ensure its continuity and long-term positioning,
as well as contributing to the continuous improvement of the basic conditions of
academic quality, to we have considered the following strategies:
490 J. Bravo et al.

• The people dimension: improvement of the organizational climate, perma-

nent training of teachers in terms of management and digital skills, awareness
campaigns.
• The infrastructure dimension: implementation of the school management offices
and academic departments with technological equipment, improvement of
Internet bandwidth, adequately define the policies for access to information and
technological equipment.
• The business architecture dimension: elaborate the procedure for executing the
tutorials performed on the students, continue with the follow-up of the teaching
digital portfolio.
• The digitalization dimension: follow up the class session plans, verify the tutorials
executed, and determine indicators to evaluate the teacher’s teaching activity.

5 Conclusions

The result of the proposed model demonstrates that the initial level of digital maturity
in teachers was basic and it was possible to reach the intermediate level. The initial
level of organizational maturity is pondered as interested, reaching the level involved.
The initial level of infrastructure, digitalization and architecture of the business were
pondered as basic, reaching the level managed. Also, the organizational culture is
essential to achieve the strategic objectives, since the mastery of the technological
tools is not enough if the teacher does not commit to his institution. As future work,
a much more exhaustive analysis of the teacher’s characterization should be made to
identify which criteria are relevant and determine a profile of the university teacher
for its digital transformation.

References

1. Mergel, I., Edelmann, N., Nathalie, H.: Defining digital transformation: results from expert
interviews. Gov. Inf. Q. 36, 1–16 (2019)
2. Berman, S.: Digital transformation: opportunities to create new business models. Strat. Leader.
40(2), 16–24 (2012)
3. Bernhard-Skala, C.: Organisational perspectives on the digital transformation of adult and contin-
uing education: a literature review from a German-speaking perspective. J. Adult Contin. Educ.
25, 20 (2019)
4. Ariño Martín, L. A.: Digital transformation: key points and considerations for reflection of
the digital transformation in the university. RUIDERAe: Revista de Unidades de Información,
pp. 1–20 (2018)
5. Verhoef, P. C., Broekhuizen, T., Bart, Y., Bhattacharya, A., Qi Dong, J., Fabian, N., Haenlein,
M.: Digital transformation: a multidisciplinary reflection and research agenda. J. Bus. Res. Paris
1–13 (2019)
6. Jackson, N.C.: Managing for competency with innovation change in higher education: examining
the pitfalls and pivots of digital transformation. Bus. Horiz. 62, 761–772 (2019)
Model of Sustainable Digital Transformation … 491

7. Aguiar, B., Velásquez, R., Aguiar, J.: Innovación docente y empleo de las TIC en la Educación
Superior. Espacios 40(2), 8–19 (2018)
8. Almarz Menéndez, F., Maz Machado, A., López Esteban, C.: Analysis of the digital transforma-
tion of Higher Education Institutions. A theoretical framework, Revista de educación mediática
y TIC, España 181–202 (2016)
9. Vivanco, G., Gorostiaga, J.: Cultura digital y diversidad: perspectivas de discursos de políticas
TIC-Educación. Cadernos de Pesquisa 47(165), 1016–1043 (2017). https://doi.org/10.1590/198
053144261
University of Brasilia’s Potential
for Bioeconomy Services and Innovation

Flávio Duque Estrada Soares Pereira and Wagner Augusto Fischer

Abstract Bioeconomy is an emerging concept that has been more and more inte-
grated into discussions concerning sustainable development. Such a theme became
one of Brazil’s main strategies for science, technology and innovation, spurring
several initiatives seeking to better understand the country’s bioeconomic capacity.
The present study was one of several conducted in Distrito Federal attempting to
assess the state and its surrounding region’s economic strengths and weaknesses, part
of Projeto Brasília 2060. The aim of this study was to perform a descriptive analysis
of the University of Brasília’s potential in the field of Bioeconomy. We conducted
semi-structured interviews with respondents from 74 laboratories from four different
institutes/faculties of the Darcy Ribeiro Campus. Our results showed that 48 labo-
ratories had the latent capacity to start/expand activities related to technological
innovation or service provision to firms, government or society. Most laboratories
seemed more inclined to provide services than technological innovation. Using the
content analysis methodology, we were able to detect six categories of obstacles that
hinder UnB’s potential to Bioeconomy.

Keywords Innovation policy · Technology transfer · Content analysis · Research

management

1 Introduction

Bioeconomy is an emerging concept that has been progressively more integrated

into the global agenda concerning sustainable development. Such term came into
the spotlight due to the climate change scenario and the perspective of several non-
renewable resources being depleted in the near future. The European Union, for

F. D. E. S. Pereira (B) · W. A. Fischer

Instituto Brasileiro de Informação em Ciência e Tecnologia—IBICT, Brasília, DF 70070-912,
Brazil
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 493
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_49
494 F. D. E. S. Pereira and W. A. Fischer

example, is structuring policies towards a more resource-efficient and sustainable

economy, promoting innovation, food security and environmental protection [1].
Similarly, Bioeconomy became part of Brazil’s main strategies for science, tech-
nology and innovation [2, 3]. A few reasons for this theme’s inclusion in the national
strategy are the country’s commitments in the Paris Agreement to reduce emissions
of greenhouse gases, as well as Brazil’s potential in this field thanks to its biodiver-
sity. Ever since it gained a strategic status, several Brazilian initiatives have aimed
to better understanding and developing of the country’s bioeconomic capacity.
One of the projects that dealt with this subject was Projeto Brasília 2060, carried
out by the Brazilian Institute for Information in Science and Technology (IBICT). Its
goal was to develop studies to better comprehend the potentials for sustainable devel-
opment in Brasilia’s Metropolitan Area. Bioeconomy was, among other thematic
axes, considered one of the most promising fields for this economic region. The
present study was part of such project and its objective was to analyse the University
of Brasilia’s (UnB) capacity in Bioeconomy. UnB was chosen for this study because
it is Distrito Federal’s oldest and largest University.
Although there is more than one definition to Bioeconomy in official strategic
documents from the Brazilian Government [2, 3], we opted to use the one from the
Organization for Economic Cooperation and Development (OECD): economic activ-
ities that come from innovation in fields related to biological sciences, culminating
on the development of sustainable products, processes and services [4].

2 Methods

2.1 Defining the Sample

Ethical approval for the experiment in this study was deemed unnecessary, as the
research was designed to investigate laboratories, not human individuals. Our ques-
tions were aimed at evaluating the laboratory itself as part of a public university. Also,
the activities in the survey do not pose risks greater than those ordinarily encountered
in daily life, and all the interviewed subjects remained anonymous. All participants
that opted to participate were also informed about the academic objective of the
research, as well as ensured preservation of their anonymity. This is in accordance
with Article 1 of Resolution 510, of 7 April 2016, of Brazil’s National Health Council
(available at http://www.conselho.saude.gov.br/resolucoes/2016/Reso510.pdf) and
also with item 8.0.5 of the Ethical Principles of Psychologists and Code of Conduct
of the American Psychological Association (available at https://www.apa.org/ethics/
code/principles.pdf). Furthermore, Brazilian Law no. 12.527, of 8 November 2011
states that public universities (and therefore its workers—see Art. 7; III) have the
obligation to provide any information requested related to its resources (Art. 7; VI),
organization, activities, projects, goals and indicators (Art. 7; VII).
University of Brasilia’s Potential … 495

UnB is a federal university comprised of five different campi, including the

Fazenda Água Limpa (FAL), an open campus/laboratory (a farm) destined to conduct
experiments in Biology, Veterinary, Agronomy and other research fields. According
to their respective websites, there are 14 different laboratories on the Ceilândia
Campus (CC), most of them carrying out health sciences research. Ten research
laboratories are located on the Planaltina Campus (PC), most of which are dedicated
to Bioeconomy and Biotechnology studies. All of the five laboratories on the Gama
Campus (GC) perform engineering research. The Darcy Ribeiro Campus (DRC),
on the other hand, was the first one created for UnB, accommodating over 400
laboratories from several academic fields.
Because Darcy Ribeiro Campus (DRC) is the largest and oldest of the campi,
we chose to restrict our research to this site, since it is expected that its researchers
would be more familiar with the university’s mechanisms for technology transfer and
innovation. The next step of our study was to identify which institutes or faculties
in DRC congregate most of the laboratories that perform (or have the potential to
perform) research and services in the field of Bioeconomy. Therefore, we analysed
all information from laboratories and research groups available on UnB’s website.
Our analysis indicated that most laboratories with research and activities related to
Bioeconomy on DRC were located on the Institute of Biology (IB), the Faculty of
Health Sciences (FHS), the Institute of Chemistry (IC), the Faculty of Agronomy and
Veterinary (FAV) and the Faculty of Technology (FT), specifically on the Department
of Environmental Engineering.

2.2 Survey Design

Once our sample universe was defined, we developed and applied our survey
consisting of seven questions, two of which aimed to identify the respondent (Q1,
Q2), four were to delineate the activities, infrastructure, collaborations and funding of
the laboratory (Q3, Q4, Q6, Q7), and one question was to obtain information about
the laboratory’s current situation when it comes to innovation and technological
transfers (Q5). The survey is detailed below:
• Q1. Name of the interviewed researcher
• Q2. Researcher’s contact
• Q3. Research themes under way at the laboratory
• Q4. Main equipment in use at the laboratory
• Q5. Laboratory’s capacity (latent or consolidated) to provide services to the
society or either the private or public sector, and its capacity (latent or consoli-
dated) for technological production/innovation. Here, we specifically asked the
researchers to:
– cite the main services and technologies that could be provided/developed by
the laboratory.
496 F. D. E. S. Pereira and W. A. Fischer

– talk about the barriers that hindered the laboratory’s potential to develop new
services and technologies.
– report what would be necessary to expand the provision of
services/technological innovations that already existed in the laboratory.
• Q6. Research collaboration (I) within the same institute/faculty, (II) with
different institutes/faculties within the same university, (III) with other Brazilian
universities and (IV) with other countries.
• Q7. Is the laboratory being funded?

2.3 Survey Application

All interviews were conducted in loco, along three months (from July to September
2018), with whichever researchers (researcher-professors or either Ph.D. or post-
doctorate students when the professors were not available) were present in the
laboratories during the visitation period. The interviews were conducted in a semi-
structured manner, and the respondents’ answers were written on a printed form
containing the seven questions, being posteriorly transcribed to a Microsoft Excel
table. All respondents agreed to participate in the study and were informed of its
objectives. In order to acquire a representative description of what factors possibly
hindered DRC’s laboratories capacity in the field of Bioeconomy, we used content
analysis to identify recurring topics in the respondent’s answers [5, 6]. Francis and
collaborators [7] recommend establishing a sample size for the first round of analysis
(referred to as initial analysis sample), then performing a certain number of inter-
views to confirm if the data is saturated (i.e. no new topics appearing). The authors
recommend at least three interviews following the initial analysis sample to confirm
saturation.
Therefore, our visits to the laboratories/correspondent researchers started on the
Faculty of Agronomy and Veterinary (FAV), followed by the Institute of Biology (IB),
Institute of Chemistry (IC) and Faculty of Technology (FT). The answers obtained
were analysed (as detailed in our Sect. 2.4: content analysis), and the initial categories
of topics were proposed.
Afterwards, we conducted five interviews in the Faculty of Health Sciences (FHS)
to test the hypothesis of data saturation. Indeed, no new topics emerged during this
process. The FHS was chosen for this purpose because its laboratories acting in the
Bioeconomy field only represent a small percentage of its total laboratories. Even
though there were new mentions of topics already presented, data from FHS was
excluded entirely, since only five interviews would not constitute a representative
sample from the faculty. We consider this was not harmful to our analysis, since
there were evidences that responses obtained from the other institutes were enough
to reach saturation (i.e. an optimal sample size [6]).
In total, we visited 74 laboratories on the DRC: FAV (16 laboratories), IB (38), IC
(12) and FT (8), accumulating over 100 respondents. Laboratories on Darcy Ribeiro
Campus frequently have more than one researcher-professor responsible for carrying
University of Brasilia’s Potential … 497

out different lines of research. In order to avoid overstatement of the answers provided
in our survey, we grouped the answers of all researchers from the same laboratory.

2.4 Content Analysis

Once the answers from the 74 laboratories were tabulated, we proceeded to the pre-
analysis of their content. The organization of the material is followed by its “floating
reading” [8], with the objective of impregnating the researcher with its content [9].
After reading the entire material, we proposed six categories, or topics, related to
the capacities for technological innovation and service provision, as well as two
binary variables for the presence/absence of innovation or service provision and the
possibility for expanding current activities. Each category was associated with certain
terms and themes which appeared in the sample of interviews [10].
The next step encompassed counting which categories were represented in each
of the answers, being assigned the value 1 (one) if terms or themes related to that
topic were present and 0 (zero) if none were present. We aggregated all data from
each institute before presenting it in order to preserve the identity of the respon-
dents. Furthermore, a topic was considered relevant to our study only if it reached
a minimum of five occurrences in the sample. Other topics that appeared more than
once but failed to meet this criterion will also be addressed in our next section.

3 Results and Discussion

Our results and discussion will be divided into three sections. The “Findings from the
survey” section will focus on the answers obtained from questions Q5–Q7, presenting
an overview of the results. The following “Barriers for innovation and services related
to Bioeconomy” will detail the findings presented in the preceding section. Finally,
“Limitations of the study” will discuss the methodological limitations in this research
as well as recommendations for future studies.

3.1 Findings from the Survey

The main themes/categories detected during the floating reading of question Q5’s
answers were “need for physical space”, “need for equipment maintenance”, “lack
of equipment”, “need for laboratory technicians”, “need for laboratory consumables
and reagents” and “bureaucratic problems”. A brief summary of the categories with
examples of answers that fit in each of them is presented in Fig. 1, including state-
ments that indicate if the laboratory had an established capacity in technological
production or service provision, as well as the latent capacity to start or expand such
activities.
498 F. D. E. S. Pereira and W. A. Fischer

Fig. 1 Categories/dimensions with examples of answers that fit in each of them

The absolute frequency of answers per laboratory that fits into each category can be
seen in Table 1. On the same table is the number of laboratories that declared having
provided services to the society or either the private/public sector or had developed
technology (established capacity) or had the capacity to start/expand such activities
(latent capacity).
The topics are shown in Fig. 1 and Table 1 were not the only ones that appeared
during the interviews. Others, such as difficulty to conform to standards required

Table 1 Absolute frequency of answers per institute that fit into each topic/category, including
declared established and latent capacity
Topic Institute (or faculty)
FT IC FAV IB Total
Established capacity 4 9 6 9 28
Latent capacity 6 9 7 26 48
Need for physical space 2 3 0 0 5
Equipment maintenance 3 0 0 2 5
Lack of equipment 2 7 3 5 17
Need for laboratory technicians 3 3 4 10 20
Need for laboratory consumables or reagents 1 1 1 4 7
Bureaucratic problems 0 3 1 5 9
Interviewed laboratories 8 12 16 38 74
University of Brasilia’s Potential … 499

for laboratory certification (such as INMETRO or ANVISA) and need for funding
showed up in the answers with a total absolute frequency of 2 and 4, respectively. The
topics related to certification could either be a specific scenario to the laboratories
that reported them or even another type of problem associated with the “bureaucratic
problems” dimension. The need for funding, on the other hand, was most likely
linked to a specific need of the laboratory related to other topics (such as equipment
maintenance or acquisition) that was not properly disclosed by the interviewees.
It is also important to highlight that none of the presented categories/themes were
mutually exclusive. For example, researchers could declare to have an established
capacity to provide services or produce technologies, and also believe their laboratory
could expand such activities.
When it comes to resources, 44 of the laboratories declared being funded at the
moment, and the most frequent source cited was the Research Support Foundation
of Distrito Federal (FAPDF). Most of the other funding sources cited were also
from the public sector. The question Q6 of the form (about research collaboration)
showed that DRC laboratories in the field of Bioeconomy have a high amount of
cooperation within institutes/faculties, with other Universities and research centres
of Distrito Federal (especially EMBRAPA) and Universities from other Brazilian
states or countries. Interestingly, the cooperation among institutes from the same
campus was relatively small.
Our results show that 28 of the 74 interviewed laboratories (37.8%) had an estab-
lished capacity to perform technological production/innovation or services (Table 1).
In this group, 13 (46.4%) also declared they could expand such activities under certain
conditions. Most of the 48 laboratories with latent capacity were more inclined
to perform services to the society or the private/public sector than technological
production. Such services include advising companies and government in the field
of Bioeconomy, providing analyses of materials and components and even creating
certifications for products. However, several barriers compromise the consolidation
of these capabilities.

3.2 Barriers for Innovation and Services Related

to Bioeconomy

Since this paper characterizes a case study, the results obtained from the sample
represent only the situation of DRC’s laboratories acting in the field of Bioeconomy.
Therefore, even though we shall compare our findings with others from the literature,
it is important to note that only one other article [11] had a similar scope, also being
conducted at UnB, specifically on its Support Centre for Technological Development
(SCTD).
According to the interviewees, the “need for consumables or reagents” impacted
their laboratory’s capacity to supply continuous services to the society, especially
due to their elevated costs. Similarly, as reported by the respondents, the necessity of
500 F. D. E. S. Pereira and W. A. Fischer

“equipment maintenance” made it difficult for their laboratories to perform certain

types of analyses or services continuously to the private sector. However, that does
not mean the laboratory’s equipment was not working or out of use. Instead, the lack
of proper maintenance made it difficult for the laboratories to ensure the accuracy of
their analyses to companies. Both of these topics are related to one of the reasons why
laboratories seek to provide services or technology transfers, which is the attainment
of necessary resources to carry on their research activities [12, 13].
Laboratories’ “need for physical space” appeared only in answers from the Insti-
tute of Chemistry and the Faculty of Technology. The appearance of this topic on the
IC was related to the need for scaling up the activities in order to achieve what the
respondents considered an adequate technological production or service provision.
This topic’s presence on the FT, however, was mostly related to their current phys-
ical installation. The Department of Environmental Engineering, where our research
took place, had increased its staff and activities over the years inside the FT. Such
facilities had consequentially become too small to accommodate the Department,
and the University was taking action to reallocate them into a new building.
The “lack of equipment” was one of the most frequent topics cited in our sample
(17 times), and it brings some relevant considerations to our study. It was observed
that this topic’s mention on the IC was usually related to the need for scaling up
processes of technology production or service provision (much like the category
“need for physical space”). Respondents from the IC also stated that sometimes
companies reach out to the laboratories requesting types of analyses that require
specific equipment that is not available. Interestingly, the IC has at its disposal an
Analytical Centre (Central Analítica do Instituto de Química—CAIQ), a laboratory
located on the IC’s facilities with a pool of equipment and personnel dedicated to
operating them. It seems, however, that CAIQ is not enough for IC’s researchers to
develop their full potential for innovation and services and is more often used to
assist them in their academic work.
To other institutes and faculties, on the other hand, the “lack of equipment”
seems to impact both the provision of services/innovation and their academic work.
However, the fourth question in our survey (main equipment in use at the laboratory)
showed that several of the equipment listed by a laboratory as a necessity was usually
available in other laboratories on the campus. This shows that researchers on the same
campus often do not know other laboratories’ work and installations, and a result
that was corroborated by our sixth question (research collaboration), which showed
that cooperation between institutes and faculties within the same university are less
frequent than within institutes or between different universities. For example, no
laboratory outside the Institute of Chemistry cited any use of IC’s Analytical Centre
(CAIQ).
The topic “bureaucratic problems”, which appeared nine times across our sample,
encompassed both the slow internal processes of the University and the lack of prepa-
ration that professors-researchers have to deal with them. This last aspect also appears
in other studies about technology transfer [12–14]. When it comes to this dimension,
our interviewees demonstrated some frustration with UnB’s bureaucracy (although
there were also positive opinions towards its institutions dedicated to technology
University of Brasilia’s Potential … 501

transfer and research support). For example, some respondents have reported that it
takes at least six months for the university to approve the necessary arrangements
for the laboratory to provide services to a firm, a reality that was also described
in another study conducted on UnB [11]. The respondents disclosed their previous
experiences with know-how and technology transfer, stating that some companies
were not willing to wait for the extensive time it takes for cooperation between
the private sector and university to be approved. Costa and Cunha [12] explain that
small-medium enterprises (SMEs) often find this bureaucracy inadequate and lack
the proper stimuli to invest in innovation. Bigger enterprises, on the other hand, may
choose to execute their research and development internally, whilst international
companies may simply opt to import technologies.
The category “need for laboratory technicians” was another barrier to the labora-
tories’ latent capacity, appearing with the highest absolute frequency in the sample.
According to the interviewees, personnel that works on the laboratories is most often
needed to execute research activities, being impossible to allocate the researchers
to perform external services. As stated by previous studies [12, 13], the Brazilian
Governmental policy for evaluation of higher education institutes may contribute
to this issue, since the number of researchers’ publications are more relevant to
their professional evaluation than work developed in cooperation with companies (a
scenario also seen in Argentina [14]).
Barra et al. [15] have studied correlations between academic excellence, local
knowledge spillovers and innovation in Europe. Their study shows a possible
trade-off between academic impact/reputation and appropriation of scientific knowl-
edge/innovation by local firms, since they had found negative correlations between
a number of publications per university staff and local innovation. Barletta et al. also
found similar results in Argentina [14], by observing a negative correlation of scien-
tific production (i.e. papers) and technological transfers. The authors also found
a negative correlation between public funding and technology transfer, whilst the
same type of funding was positively correlated with scientific productivity. Barletta
et al. argued that the design of public scientific funding in Argentina mostly targets
researchers with higher levels of publications, failing to provide incentives for scien-
tists to cooperate with firms [14], similar to the manner Brazilian scientific policy is
structured.
Interestingly, several respondents stated that performing analyses of food,
produce, environmental components or microorganisms on a producer or company’s
request would provide data that could be used in their researches. Likewise, a study
conducted in the United States [16] found positive correlations between scientific
publications and innovation developed in partnerships of small firms and universities.
This indicates that interactions between the private sector and universities produce
both scientific papers and patent deposits, and one form of knowledge output does not
necessarily crowd out the other. Such a result, however, may be due to greater matu-
rity of the United States’ Universities when it comes to innovation and interaction
with firms.
Furthermore, the interactions between firms and university researchers may also
bring positive knowledge spillovers between both sectors. For example, cooperation
502 F. D. E. S. Pereira and W. A. Fischer

with laboratories may contribute to a company’s human resource improvement and

ability to carry out research and development. Those characteristics have a role in a
firm’s capacity to absorb new technologies, a relevant aspect of technology transfer
[11, 12, 14, 17, 18].
From the laboratories’ point of view, such interactions are also advantageous. For
instance, a researcher may develop new technology without understanding how to
increase its production in scale or how to insert it in the market. Cooperating with
companies can help improve the scientists’ understanding of their own researches
applicability [14]. The improvement of a laboratory’s ability to create functional
prototypes or marketable products consequentially contributes to further attract new
firms for possible partnerships [17].
Another intriguing aspect observed in this study is that most of the researchers
seemed more inclined to interact with firms, government and society through service
provision, instead of other mechanisms such as technology licensing. Yet, most of
the literature on technology transfer from universities seems to deal with patent
deposits and their licensing to companies. Our study might be indicating that other
forms of cooperation between universities and firms/society are undervalued in this
field. Indeed, broad research conducted in Brazil [18] suggested that the relevance
of patents to technological transfer varied according to the technology type. For
example, transferring processes and techniques to the industry usually did not involve
patents, being instead correlated to consulting activities. Even considering their
entire sample, “only 14.1% of respondents pointed to the use of patents to transfer
technology” [18].

3.3 Limitations of the Study

Although the methodology adopted in this study was useful to reveal the current
situation of innovation of DRC’s laboratories in the field of Bioeconomy, more steps
would be needed for an in-depth analysis. Our research managed to capture six
categories of barriers that possibly hinder Bioeconomy innovation and the provision
of services. However, there may be other issues that could impact such laboratories’
capacities. Other works [12, 13], for example, suggest that divergences between
academic and firm cultures are hindrances to university-company cooperation, a
topic that did not appear in our sample.
Also, our study does not address how each topic is related to one another. For
instance, the “need for equipment maintenance” and “need for laboratory consum-
ables or reagents” could be related to a hidden underlying factor. Factor analysis is
a statistical tool widely used by Marketing, Psychology and Health Sciences [19–
22] and could be applied to estimate possible relationships among the categories.
It would be necessary, however, to create a more detailed survey encompassing the
categories presented in this study and others taken from the academic literature in
order to use factor analysis. Scussel and Demo [6] used a similar methodology to
propose a scale that measures the relationship between costumers and luxury brands.
University of Brasilia’s Potential … 503

Finally, our study indicates a large latent Bioeconomy capacity of Darcy Ribeiro
Campus’s laboratories, especially when it comes to providing services to the society
and the private/public sector. However, we have only obtained information from
the suppliers of such services. Only by assessing the actual demand for what the
university offers can we truly estimate its Bioeconomy potential.

4 Conclusion

The scope of this study was to perform a descriptive analysis of the current situation of
Darcy Ribeiro Campus’s laboratories when it comes to innovation and service provi-
sion in the field of Bioeconomy. Our survey revealed that several laboratories had the
latent potential for these activities, and respondents were more inclined to cooperate
with companies and society by providing services than through technology transfer
to firms. Six categories of obstacles to the realization of this potential emerged from
the answers. We highlight that the constant requirement for researchers to publish
academic results may be undermining their capacity/availability to innovate in the
paths adopted to transfer the scientific knowledge from university to society. Further-
more, narrowing the relationship between academics and companies may contribute
to both the sector’s intrinsic abilities to generate innovation. It is also important to
note that, although our study indicates a large latent capacity of DRC’s laboratories
in the field of Bioeconomy, we have only surveyed the laboratories themselves (i.e.
the suppliers of products and services). New studies considering society’s demands
for such products and services will help to improve our understanding of UnB’s true
potential in the field of Bioeconomy.

Acknowledgements We would like to thank the researchers from UnB’s Darcy Ribeiro Campus for
their availability and disposition to contribute to this work. This study was funded by Projeto Brasilia
2060, executed by the Brazilian Institute for Information in Science and Technology (IBICT).

References

1. European Commission: A sustainable Bioeconomy for Europe: strengthening the connection

between economy, society and the environment. Brussels (2018)
2. Brasil: Estratégia nacional de ciência, tecnologia e inovação 2016–2022. MCTIC, Brasília
(2016)
3. Brasil: Plano de ação em ciência, tecnologia e inovação em bioeconomia. MCTIC, Brasília
(2018)
4. Organisation for Economic Co-operation and Development: The Bioeconomy to 2030:
Designing a Policy Agenda. OECD Publishing (2009)
5. Bardin, L.: Content Analysis. Livraria Martins Fontes, São Paulo (1977)
6. Scussel, F., Demo, G.: Os aspectos relacionais do consumo de luxo no Brasil: o desenvolvimento
da escala de percepção de relacionamento de consumidores de luxo e a análise da influência
504 F. D. E. S. Pereira and W. A. Fischer

da personalidade de marca sobre a percepção de relacionamento com marcas de moda de luxo.

Braz. Bus. Rev. 16(2), 174–190 (2018)
7. Francis, J.J., et al.: What is an adequate sample size? Operationalising data saturation for
theory-based interview studies. Psychol. Health 25, 1229–1245 (2010)
8. Seramim, R. J., Walter, S. A.: What does Bardin say that authors do not show? Study of Brazilian
scientific productions from 1997 to 2015. Adm. Ens. PESQ 18, 241–269 (May, 2017)
9. Campos, C. J. G., Turato, E. R.: Content analysis in studies using the clinical–qualitative
method: application and perspectives. Rev. Lat. -am. ENF 17, 259–264 (March, 2009)
10. Santos, F.M.: Análise de conteúdo: a visão de Laurence Bardin. Rev. Eletrôn. Edu. 6, 383–387
(2012)
11. Ferreira, C. L. D., Ghesti, G. F., Braga, P. R. S.: Desafios para o processo de transferência de
tecnologia na Universidade de Brasília. Cad. Prospec. Salvador 10, 341–355 (July, 2017)
12. Costa, V. M. G., Cunha, J. C.: A universidade e a capacitação tecnológica das empresas Rev.
Adm. Contemp. 5, 61–81 (January, 2001)
13. Closs, L.Q., Ferreira, G.C.: University-industry technology transfer in the Brazilian context: a
review of scientific studies published from 2005 to 2009. Gest. Prod. 19, 419–432 (2012)
14. Barletta, F., Yoguel, G., Pereira, M., Rodríguez, S.: Exploring scientific productivity and transfer
activities: evidence from Argentinean ICT research groups. Res. Policy 46, 1361–1369 (2017)
15. Barra, C., Mayetta, O.W., Zotti, R.: Academic excellence, local knowledge spillovers and
innovation in Europe. Reg. Stud. 53(7), 1058–1069 (2019)
16. Audretsch, D. B., Link, A. N., Hasselt, M.: Knowledge begets knowledge: university knowledge
spillovers and the output of scientific papers from U.S. Small Business Innovation Research
(SBIR) projects. Scientometrics 121, 1367–1383 (October, 2019)
17. Pereira, M. F., Melo, P. A., Dalmau, M. B., Harger, C. A.: Transferência de conhecimentos
científicos e tecnológicos da universidade para o segmento empresarial. RAI 6, 128–144
(September, 2009)
18. Póvoa, L.M.C., Rapini, M.S.: Technology transfer from universities and public research insti-
tutes to firms in Brazil: what is transferred and how the transfer is carried out. Sci. Publ. Policy
37(2), 147–159 (2010)
19. Auerswald, M., Moshagen, M.: How to determine the number of factors to retain in exploratory
factor analysis: a comparison of extraction methods under realistic conditions. Psychol.
Methods 24, 468–491 (2019)
20. Clevers, E., et al.: Factor analysis defines distinct upper and lower gastrointestinal symptom
groups compatible with rome iv criteria in a population-based study. Clin. Gastroenterol. H.
16, 1252–1259 (2018)
21. Calobrace, M.B., et al.: Risk factor analysis for capsular contracture: a 10-year sientra study
using round, smooth, and textured implants for breast augmentation. Plast. Reconstr. Surg. 141,
20S–28S (2018)
22. Hooker, J.L., et al.: Psychometric analysis of the repetitive behavior scale-revised using
confirmatory factor analysis in children with autism. Autism Res. 12, 1399–1410 (2019)
A Proposal for Modeling
of the Management of Talent
Recruitment and Training in Peruvian
Sports Centers

Arturo Laredo , Luzmila Pro Concepcion ,

and Nestor Mamani-Macedo

Abstract This study was determined by the need of entities of the Peruvian state to
know whether relevant institutions are properly implementing a sports management
process at the national level because attracted and recruited sports talents mostly
arrive by their own free will, and many such talents drop out during their training
path. This research was conducted using qualitative and quantitative methods, with
a descriptive approach where instruments such as interviews and observation were
applied to coaches, physical education teachers, athletes, and professional personnel
involved in sports management. This study aimed at designing a sports management
model based on information technology to help integrate the management indicators,
which allowed the identification of more talented athletes, to reinforce the skills of
such athletes and guide them to international competitions and tournaments. In this
manner, the early dropout of talented athletes can be avoided.

Keywords IT model · Sports management · Sport management indicators

1 Introduction

Sports have always been considered important, similar to other disciplines, because
it facilitates the acquisition of different individual skills, such as motor skills, in
addition to the learning skills; further, engaging in sports improves the health of the
population. Thus, government entities must focus on sports disciplines, and sports

A. Laredo · L. P. Concepcion
Faculty of Systems and Computer Engineering, Universidad Nacional Mayor de San Marcos,
Lima 15000, Peru
e-mail: [email protected]
L. P. Concepcion
e-mail: [email protected]
N. Mamani-Macedo (B)
Faculty of Mathematical Sciences, Universidad Nacional Mayor de San Marcos, Lima 15000, Peru
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 505
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_50
506 A. Laredo et al.

institutions must implement a responsible sports management (SM) system, using

indicators that facilitate talent identification among the youth participating in high-
level competitions. To properly implement an SM system, information technologies
(IT) must be used.

2 Theoretical Foundation

There are few studies on the use of IT in relation to sports management and similar
topics, such as the impact of IT on the effectiveness and efficiency of SM models.
Rodríguez [1] provides a relationship between information and communication tech-
nologies (ICTs) and sports. There is an insufficient amount of documentation on
SM models; consequently, each country designs the best possible structure and tech-
nology adaptation for sports management in its own manner [2, 3]. Given the scarcity
of literature on SM using IT models, studies have been found on technological inno-
vation originally intended for different functions and applied in other scientific fields,
such as medicine [4, 5]. Nevertheless, there is no doubt that the ICTs tools allow
obtaining relevant data about sports [6], although some argue that their use is not yet
significant [7].

2.1 Theoretical Models Used to Optimize SM in Attracting

Talent

Factors such as performance levels, skills, human resource capabilities, and their
level of commitment and identity toward the institution are important for sports
managers [8] because the sports performance largely relies on it. Thus, an SM
model aware of these factors based on quality and continuous improvement allows
identifying policies improving sports activities, which information might be neces-
sary for state agencies regards three aspects: the political, processes, and the total
quality [3]. Information management model for sports organization: a model aimed
at the implementation and development of five components (human capital, organi-
zational memoirs, the operation and integration of institution members, information
centers, and ICTs) that constitute a continuous improvement process in information
management [9–13].

2.2 IT Applications for SM

The following technologies can be used to generate an SM database:

• Business-to-business (B2B): technology that allows connecting the application
services of two companies at a minimum cost, improving revenue stream of both.
[10]
A Proposal for Modeling of the Management … 507

• Motion capture: is a live movement recording, recreating such movement on a

digital model using computer images [12].
• Big data: describes the large volume of data—both structured and unstructured—
that floods a business on a day-to-day basis [12].
• Cloud computing: allows the user to access a catalog of standardized services to
meet their business needs in a flexible and adaptive manner [13].

3 Problematic Situation

Rodríguez [1] argues that the negative results related to the use of ICTs in SM are
due, among others, to the reluctance of coaches and athletes to take advantage of
these technologies to improve their performance. Slusarczyk [2] considers that the
poor use of ICTs, as in sports, produces major organizational gaps because the people
in charge of business management are generally not familiar with the technology.
Currently, the Peruvian Sports Institute (IPD) only implements information systems
for administration and not for SM because of a lack of knowledge of the standard
systems.
General Problem: There are few models of SM using IT that accurately measure
SM processes and data in their entirety and in an efficient and timely manner to attract,
develop, and produce more sports talents. All processes are aimed at recruitment,
selection, training, promotion, and high competition. This is a challenge faced every
day by entities that are dedicated to training, development, and competition in sports
disciplines.
Specific Problems:
• Inadequate use of ICTs by those involved in sports recruitment and training, and
the lack of a model of relevant indicators to measure talents that are not discovered
at an early age in an efficient manner.
• A lack of policies and strategies by the public entities that deal with sports
development.
• A lack of an SM model with adequate indicators to measure talents that are not
discovered at an early age in an effective manner.

4 Methodology

This study comprises applied research that uses a pre-experimental design for which
we created the necessary conditions or adapted existing ones to clarify the character-
istics and relations of the optimal IT resources required to integrate and support SM
processes. The research objective is to design an SM model based on IT, indicators,
and implementation using qualitative and quantitative methods, as shown next in the
phases of the methodology used.
508 A. Laredo et al.

Fig. 1 Research methodology for defining a sports management model using information
technology applications for sports management

Six IT program procedures used in Phase 1 of the SM Model during the “Informa-
tion Collection and Analysis” activities are chosen as shown in Fig. 1. Based on this,
a management model using IT applications and SM evaluation elements is proposed,
as shown in Table 1.
Figure 2 shows the IT tools used for model validation:
• Motion capture technology: a mobile application that captures bone height
through scanning; speed through video recordings; the distance between two
points.
• Telepsychology: Instant Message (IM) mobile/desktop applications for interac-
tive evaluations.
• Big data and Cloud computing: for data storage and analysis.
• Biomedical technology: a medical calculator system (mobile/desktop applica-
tion) about metrics and equations of people healthcare.
A Proposal for Modeling of the Management … 509

Table 1 Evaluation of the sports management model based on IT applications

Evaluation IT applications
elements B2B Motion Big Cloud Telepsychology Biomedical
capture data computing technology
technology
Content 10 10 10 10 5 5
distribution
Strategy 10 10 10
management
without
considering
the talent
acquisition
model
Athlete 10 10
tracking
Measure 10 9
mental health
Measure 6 8 5 8 10
physical
performance
Availability of 10 10 10 10 8 8
information

5 Validation

After having completed the SM model based on IT applications, we adapted the

model to the Experimental Sports Educational Center of the IPD (CEDE) in June
2018. Subsequently, we assessed the model through the indicators shown by the
different applications and compared the current state against the post-implementation
state (after six months).

6 Discussion

Table 2 shows that the number of athletes referred with projection highly competitive
to receive support from their federations in the two periods increased by 14%. On
the other hand, the number of specialists using data capture technology increased by
9%. The indicators number of athletes evaluated and psychologically supported and
the number of athletes whose training can be sponsored was zero in June 2018 and
5% and 10% of the total in December 2018, respectively. This was due to the use
of telepsychology and data capture technologies through smartphones with installed
applications that measured with greater accuracy and transmitted the information in
seconds to the server, thereby considerably minimizing the time previously used by
510 A. Laredo et al.

Fig. 2 Information technology applications selected for the model

trainers to evaluate each applicant. With relation to synergy indicators, the number
of athletes subsidized (with a grant) by IPD increased by 7%, and the number of
athletes registered for international competitions increased by 12% from June 2018
to December 2018. In the same period, the number of athletes with economic and
social support from the IPD and the Ministry of Education went from 0 to 4%. This
was due to the use of biomedical technology and economic and social applications
that were measured with accuracy data of applicants.
In the present model proposed for SM at the CEDE-IPD (Fig. 3), we established
three main management processes: selection (attracting sports talents), training
(follow-up and monitoring), and projection (athletes are referred to federations,
clubs, or sports centers); subsequently, the athletes are to be qualified as high-
performance athletes that are directly supported by the state and/or sponsors. Simi-
larly, we designed a set of indicators per sports process: effectiveness Indicators
for the selection phase and efficiency and synergy indicators for the training phase
as described by the indicators in Table 2. The “sports strategies” will address the
model execution according to the geographical areas that each federation prioritizes
to execute.
Based on the SM indicators, we integrated these technologies into a new SM model
wherein the new evaluation criteria used from other disciplines that converge with
the health and psychology of athletes have supported better analysis and decision
making through the use of multi-disciplinary information. Further, another finding is
A Proposal for Modeling of the Management … 511

Table 2 Six-month comparison of the indicator evaluation

Indicators Jun-18 Dec-18 % Increasea
Effectiveness
Number of applicants 57 %b 78 %c
Number of candidates 0 – 78 100% –
evaluated based on the
anthropometric
characteristics, nutrition,
social, economic, cultural, and
psychological data
Number of applicants visually 57 100% 0 – –
evaluated by the coach
Efficiency
Number of athletes referred 12 21% 27 35% 14%
for projection (federations)
Number of athletes evaluated 0 – 13 17% 17%
and psychologically supported
Number of athletes whose 0 – 4 5% 5%
training can be sponsored
Number of specialists using 1 2% 8 10% 9%
data capture technology
Synergy
Number of athletes subsidized 4 7% 11 14% 7%
by the Peruvian Sports
Institute
Number of athletes registered 10 18% 23 29% 12%
for international competitions
Number of athletes benefiting 0 – 3 4% 4%
from economic and social
support from the Peruvian
Sports Institute and the
Ministry of Education
(MINEDU)
a The percentage of increase from June 2018 until December 2018
b The percentage of the total number of applicants in June 2018
c The percentage of the total number of applicants in December 2018

that the number of data provided to CEDE physical education teachers for evaluating
the recruitment and selection processes of sports talent is greater and better addressed
in relation to the processes of talent analysis.
512 A. Laredo et al.

Fig. 3 Proposed model for sports management in the experimental sports school or college

7 Conclusions and Recommendations

This work proposes a sports management model with the help of cutting-edge IT
technologies that can be used in the recruitment of talent athletes by the Peruvian
government. To operationalize the model, we define a set of indicators about effec-
tiveness, efficiency, and synergy. The model was validated in a real scenario getting
significant improvements in the results benefiting the athletes.
As future work we will develop productivity Indicators for the “Projection”
phase, these indicators will be customized according to each type of sports disci-
pline. Likewise, the model will be applied to other educational centers specialized in
high competition sports that exist in some regional governments of Peru to receive
feedback and improve the model itself.

References

1. Rodríguez, Q.M.: Information and communication technologies (ITC) in physical education.

Theor. Rev. 1(1), 75–86 (2015)
2. Slusarczyk, A. M.: Analysis of Business and ICT Strategies. Área de Innovación y Desarrollo
p. 3c Empresa 5(1), 44 (2016)
3. Moreno, P.Y.: Modelo de gestión deportiva para el municipio de Quibdó. Elsevier 5(12), 75–194
(2014)
A Proposal for Modeling of the Management … 513

4. Carballo, M., Barrero M., Villalazón, A.: FiMAN: Sistema Computarizado para Analisis de
Movimientos Digitales. Investigación y Desarrollo (2016)
5. García, X., Salguero, A., Molinero, O., De la Vega, R., Ruiz, R., Márquez, S.: El papel del
perfil resiliente y las estrategias de afrontamiento sobre el estrés-recuperación del deportista
de competición. Kronos 14(1) (2015)
6. Tiching: Las TIC en la Educación Física (2013). Available from: http://blog.tiching.com/las-
tic-en-la-educacion-fisica/
7. Alvarado, P. C.: Uso de las TICs por parte de los profesores de Educación Física. Lecturas
educación física y deportes, Revista Digital.. EFDeportes.com. Año 22-Nº230 (July, 2017)
8. Núñez, R.G., Zueck, M.C., Marín, R., Soto, M.C.: Modelo gerencial para potenciar la práctica
deportiva y recreativas. Revista Española de Educación Física y Deportes 420, 25–38 (2018)
9. Cantos, M., Reyes, J.: El nuevo modelo de Gestión Pedagógica y su impacto en las escuelas de
educación básica del cantón Cañar, Ecuador. Revista Killkana Sociales 2(4) (December, 2018)
10. CYSEND: B2B API para recarga de móviles (21 March, 2019). Available from: https://www.
cysend.com/merchant/es/b2b-api-para-recarga-movil
11. Arteneo. Captura de movimiento o motion capture: ¿cómo funciona? (25 January,
2515/2019). Available from: https://www.arteneo.com/blog/captura-de-movimientos-motion-
capture-como-funciona/
12. Ramírez, N. L.: Campeones de la Transformación Digital: 10 líderes españoles Madrid: Profit
Editorial (2018)
13. Moreno, M.: Computación en la nube. CEMA 566, 1–17 (2015)
Mentoring Process: An Assessment
of Career, Psychosocial Functions
and Mentor Role Model

Neusa Maria Bastos Fernandes dos Santos , Mariana Juer ,

Igor Polezi Munhoz , Alessandra Cristina Santos Akkari ,
Rodrigo Guimarães Motta , and Marianna Konyosi Miyashiro

Abstract This work has focused on studying the relationships between mentor
and mentee, as well as exploring the mentoring process as a current organizational
demand and the perspectives of these actors regarding mentoring functions. So, a
unique case study was developed, using as a model a large multinational company
in the housewares sector that has an official mentoring program for young talent.
Two research instruments were applied, which addressed the expectations of the
mentees and the assumptions of the mentors regarding career functions, psychoso-
cial functions, mentor role models and confidence present in the relationship, and
then a statistic treatment was performed using the SPSS software. It was observed
that although the expectations of program mentors and mentees are not entirely
compatible, no program dysfunction was observed. There are no different expecta-
tions clearly evidenced among peers, but differentiation in levels. It is suggested a
new look at how to accompany and train the participants in the mentoring process,
or even to define the pairs in order to avoid possible disagreements that may cause
frustrations.

Keyword Mentoring process · Mentor · Mentee · Career

1 Introduction

Mentoring emerged as a tool for personal and professional development and growth
in companies in the 1970s and has been evolving with changes in the world of

N. M. B. F. dos Santos · M. Juer · R. G. Motta · M. K. Miyashiro

Pontifical Catholic University of Sao Paulo (PUC-SP), Sao Paulo, Brazil
e-mail: [email protected]
I. P. Munhoz (B)
Federal Institute of Education, Science and Technology of Sao Paulo (IFSP), Sao Paulo, Brazil
e-mail: [email protected]
A. C. S. Akkari
Mackenzie Presbyterian University, Campinas, Brazil

© The Editor(s) (if applicable) and The Author(s), under exclusive license 515
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_51
516 N. M. B. F. dos Santos et al.

work [1]. Nowadays, the mentoring process has become a critical tool for supporting
and leveraging successful careers, particularly for interns and junior professionals
entering the job market [2].
In recent decades, the concept of work has changed not only from the perspective
of those who occupy the posts, but also the type of work performed and the format
of how it is performed. In this scenario of change, outsourcing is another fact to be
considered, as it has already become a visible element today, even though there is a
debate about how much this could mean a deterioration of working conditions [3].
If in the 1980s, professionals wanted to move up the hierarchy until they reached
the highest possible step in the organization where they began their careers, today’s
intellectual workers have more choices, more demands and more challenges ahead.
From this point, temporary and part-time jobs increased. The individual is not
expected to spend his entire life in a corporation, as this may sound like resignation.
Careers are less limited to traditional barriers, more flexible according to profiles, and
more directed to projects and temporary tasks, which implies greater responsibility
for professionals for their own careers [3, 4].
It is in this context of dynamism and rapid evolution of organizational paradigms
that evolve concepts of development of individuals to achieve success by organi-
zations. So, a new set of mental models is needed, which migrates the thinking of
command and control to coordinate and cultivate it. In this sense, there is expanding
coaching, as evidenced by the dozens of books and articles on the subject published,
mostly in the United States, Australia, Canada, and England. However, there are
other modalities of intervention, such as the mentoring to be studied in this article,
which is confounded, even because the authors’ distinctions on the topics were not
consensual [5].
Ensher and Murphy [1] state that although mentoring studies have advanced, there
is the criticism to be made about their development. The first concerns the use of
research results in corporations. Human resources professionals and managers do not
appear to access research material published in academic journals. Many of these
studies have been compiled into books, but, according to the authors, only a few
more recent books draw a parallel between the academic theoretical universe and
practice. The second criticism concerns the content of studies conducted over the
past ten years, which focuses on conducting questionnaires and research to describe
the nature and benefits of mentoring, even though it is so difficult to quantify the
richness of these relationships.
Therefore, the aim of this paper was to study mentoring relationships between
mentors and mentees from career, psychosocial functions, and mentor role model
perspective. This research aims to contribute to both academia and the corporate
universe by seeking a better understanding of mentee expectations and mentor
assumptions about mentoring functions, as well as the level of confidence in the
mentoring process.
Mentoring Process: An Assessment … 517

2 Theoretical Framework

Mentoring, as a field of academic study, is among the literature on career, leadership,

and development of leaders. In the late 1970s, the psychologist Levinson applied, in
the United States, a survey among successful professionals from various areas and
all stressed the importance at some point in their careers of the figure of a mentor as
a deciding factor [1].
Thus, mentoring becomes an essential theme in current labor relations and its
main objective is to provide greater job satisfaction by helping in the development
and improvement of relationships in this environment and the pursuit of the desired
results for professional growth. It is noteworthy that individuals do not depend on
their companies to participate in a mentoring relationship [6].
Many definitions of what becomes a mentor have been created without consensus
among the authors of the subject. Thus Chandler and Kram [6] adopt the definition
of Ragins which will also be adopted in this article: individuals with advanced expe-
rience and knowledge who are committed to providing support and upward mobility
to their protections’ careers.
The role of the mentor is not to indicate the right paths, but to raise questions,
open possibilities and the mentor should define the path to be taken in order to win a
challenge. It is interesting to note that some authors refer to the junior professional
who receives help from the most experienced, the mentor, as a protégé, a word of
French origin [1, 6].
The success of a formal mentoring relationship rests on a constant exchange
between mentor and mentee, in which developmental experience involves responsi-
bility and effort on both sides [4]. Appreciation and connection between the parties in
formal relations play an essential role in their willingness to devote time and energy
to the development of the process. While these factors naturally occur in informal
relationships, the same may not occur in formal relationships.
Although Burke [7] has shown that mentors can have around fifteen different
functions, the current literature has pointed to two essential functions, which have
consequences [8], being the career functions and psychosocial functions. Young and
Perrewé [9] noted that mentors value relationships more in which mentee behavior is
career-oriented, and mentee places more emphasis on behaviors related to mentors’
psychosocial issues. When each other meets common expectations, the perception
of effectiveness in the relationship’s results and the level of trust increase [10].

3 Methodology

The expectations and assumptions of the parties involved in the young talent
mentoring program of a housewares sector multinational, present in more than 170
countries, with around 60 manufacturing units and technology centers in 12 countries,
were assessed. The company has a formal mentoring program with clear objectives
518 N. M. B. F. dos Santos et al.

to develop leadership within the organization. Although the organization has formal
mentoring programs in all countries where it operates, this project focuses on the
mentor-mentee relationships involved in the youth talent program in Brazil.
This research was comprised of two adapted questionnaires by Ensher and Murphy
[1]. The first was divided into three sets of questions about mentoring functions
(career, psychosocial and mentor role model) and compares expectations of the
mentee versus assumptions of the mentor. Mentor and mentee evaluated each question
by giving a note regarding the perception of what actually happened in the relationship
and another on what they consider to be the ideal. The second questionnaire had a set
of questions about trust and also contemplated the real and the ideal scenario, which
made it possible to draw a parallel between the trust the mentee and the mentor feel, as
well as the level of trust they believe is necessary versus reported actual confidence.
The questionnaire on mentoring functions (career, emotional and mentor role model)
consisted of 5-point Likert scales; and the 4-point Likert confidence questionnaire.
The instrument was developed in a digital platform and, at the beginning of the
questionnaire, each respondent was informed about the academic objective of the
research as well as the guarantee of anonymity giving the choice of participation or
not as requested by the Research Ethics Committee of Pontifical Catholic University
of São Paulo (PUC-SP).
Results were evaluated using the Statistical Package for Social Science (SPSS)
version 13.0. The Cronbach’s alpha coefficient for each dimension (career, emotional,
mentor role model and confidence) for the real and ideal scenario was calculated,
according to the assessment of mentors and mentees. Cronbach’s alpha coefficient
measures the internal consistency of the scale, that is, how much the questions on
one scale correlate with the other questions on the same scale. In this exploratory
study, Cronbach’s alpha coefficients above 0.5 were considered satisfactory [11].
To compare the assessments of mentors and mentees for the real and the ideal
scenario, assessing the dispersion of responses and the existence or otherwise of
atypical responses, the Boxplot of mean scores for each dimension was developed.
The same methodology was used to compare the perceptions of female versus male
minds.
Finally, to assess the relationships between mentor and mentee evaluations for the
ideal and the real scenario, scatter plots were developed for each of the combinations
relevant to mentoring functions and confidence.

4 Results and Discussion

Of the 34 pairs who received the questionnaires, 18 questionnaires from mentors and
23 from mentees were answered. Regarding the mentor/mentee pair, 14 complete
answers were obtained. Cronbach’s alpha was higher than 0.547 (0.715 average) for
each dimension and for the real and ideal scenario according to the evaluation of the
mentors and mentees, allowing to consider the average scores of the questions that
constitute each dimension.
Mentoring Process: An Assessment … 519

Fig. 1 Career scale Boxplot for mentors and mentees, considering the real and ideal scenario

Figure 1 presents the career scale Boxplot for mentors and mentees, considering
the real and ideal scenario. For the career ladder, in general, mentors and mentees
value the real scenario as less than ideal. Moreover, it turns out that the ideal of
mentees is slightly higher than the ideal of mentors. However, the reality of the
mentees is slightly lower than that of the mentors; that is, both recognize that, with
regard to mentor career support, mentors perform less than optimally. However,
mentors believe that they are doing better support than they really are, according to
their mentees.
In fact, Fig. 2 assesses the relationship between the real and ideal scenario for
(a) mentors and for (b) mentees, considering the career scale. It is observed that
all mentors consider the actual scenario to be less than ideal. All study participants
understand that, for the career, the mentor should ideally provide at least medium
support. Evaluating the real scenario, it is noted that only in three cases is the career
support between good and great, but in the three cases, it is the same mentor (5, 15
and 29). Most mentees find that the support the mentor provides for their careers
is less than optimal support. However, there is a mentee who considers the actual
support to be the ideal and two mentees who consider the real to be greater than the
ideal. All mentees believe that ideal career support should be at least medium; and
most believe that it must be between good and great.
In fact, research published by Nature Reviews Cancer with scientific mentors
has pointed to mentoring as a tool that aims to help young scientists identify their
strengths and weaknesses so that they can make appropriate career choices and
explore their potential, as well contribute to the ethical aspect and integrity in
the scientific field. Through scientific mentorship, it was possible to achieve more
collaborative environments, with resource sharing and research honesty [12].
520 N. M. B. F. dos Santos et al.

Fig. 2 Relationship between the real and ideal scenario for a mentors and for b mentees, considering
the career scale
Mentoring Process: An Assessment … 521

Fig. 3 Boxplot of average career scale scores for mentees, according to gender, considering the
real and ideal scenario

Figure 3 shows the Boxplot of average career scale scores for mentees, according
to gender, considering the actual and ideal scenario. For the career scale, there is a
difference between females and males in the ideal scenario only in relation to vari-
ability, because, in this scenario, men present greater variability. In the real scenario,
women exhibit greater variability and a slightly lower perception of support than
men.
Other work in the literature [13], when analyzing the success rate of the mentoring
process by gender, found that the success ratings did not differ by mentor/mentee
gender, so that the perception of success increased with increasing frequency of
mentor-mentee interactions.
It is noted that for emotional scale (Fig. 4) there is great variability between female
and male. Also, in the ideal scenario, the expectations of male are slightly lower than
those of female. In the real scenario, female are even more dissatisfied than male.
Interestingly, other work in the literature [14], dealing with socio-emotional issues
involved in the mentoring process with a sample of refugees and migrants, pointed out
that the mentoring process acts as a tool for social inclusion, developing friendship
between mentor and mentee.
522 N. M. B. F. dos Santos et al.

Fig. 4 Emotional scale Boxplot for mentees, according to gender, considering the real and ideal
scenario

Figure 5 shows the mentor role model scale Boxplot for mentors and mentees,
considering the actual and ideal scenario. For the role of the mentor model role,
there is great variability of the real scenario, especially for the mentees. There is an
interesting divergence of opinion regarding the role played by mentors in evaluating
their own mentors and mentees. Both recognize that the real role is less than ideal,
but the mentee’s assessment of mentors is superior to that of mentors in their self-
assessment.
In this sense, it is interesting to note that the work of Esbenshade et al. [13],
with 138 mentor-mentee peers, pointed out that 78.2% of mentors found that the
mentoring process helped them succeed as professionals and 92% reported that it
helped in overall career development.
Figure 6 shows the Boxplot of the mean scores of the confidence for mentors and
mentees, considering the real and ideal scenario. It was observed that for confidence
scale there is great variability of the actual scenario for the minds. Overall, mentors
and mentees report to the real scenario a reasonable level of trust in the partner, close
to their expectations, although still below them. Their assessments of the real scenario
are similar, but the ideal scenario of the minds is superior to that of the mentors, that
is, there is a divergence of opinion about the necessary trust in relationships.
Mentoring Process: An Assessment … 523

Fig. 5 Mentor role model scale Boxplot for mentors and mentees, considering the real and ideal
scenario

Fig. 6 Confidence scale Boxplot for mentors and mentees, considering the real and ideal scenario
524 N. M. B. F. dos Santos et al.

5 Conclusion

It was observed that mentees have high expectations for both career and mentor role
models in the mentoring process. In this study, for the confidence scale, there is great
variability of the real scenario for the mentees. In general, mentors and mentees
report to the real scenario a reasonable level of trust in their partner, close to their
expectations. A divergence of views was observed regarding the necessary trust in
relationships. There is a healthy level of trust between peers, but can still improve.
Regarding roles, although there is no consensus between mentee expectations and
mentor assumptions, the results are positive.
Thus, this study pointed to a positive outcome of the program, showing the pres-
ence of mentoring functions and the existence of the trust, but the lack of consensus
regarding the role of the process is apparent.
It is suggested a new look at how to accompany and train the participants in the
mentoring process, or even to define the pairs in order to avoid possible disagree-
ments that may cause frustrations. Finally, studies that bring organizations closer to
academia are necessary to enable the evolution of mentoring to develop and improve
relationships within organizations.

References

1. Ensher, E.A., Murphy, S.E.: Power Mentoring: How Successful Mentors and Protégés Get the
Most Out of their Relationships. Jossey-Bass, San Francisco (2005)
2. Johnson, M.O., Gandhi, M.: A mentor training program improves mentoring competency for
researchers working with early-career investigators from underrepresented backgrounds. Adv.
Health Sci. Educ. Theory Pract. 20(3), 683–689 (2015)
3. Maloni, M., Gligor, D., Cheramie, R., Boyd, E.: Supervisor and mentoring effects on work-
family conflict in logistics. Int. J. Phys. Distrib. Logist. Manage. 49(6), 644–661 (2019)
4. Lankau, M.J., Riordan, C.M., Thomas, C.T.: The effects of similarity and linking in formal
relationships between mentors and protégés. J. Vocat. Behav. Athens, GA 67, 252–265 (2005)
5. Krausz, R.R.: Coaching executivo: a conquista da liderança. Nobel, São Paulo (2007)
6. Chandler, D. E., Kram, K. E.: Mentoring and developmental networks in the new career context.
In: Gunz, H., Peiperl, M. Handbook of Career Studies Sage Publications, Thousand Oaks (2007)
7. Burke, R.J.: Mentors in Organizations. Group Organ. Stud. 9, 353–372 (1984)
8. Santos, N.M.B.F.: Programas de mentoring: aprendendo com a realidade canadense. Rio
Grande: Interfaces Brasil/Canadá 7, 251–265 (2007)
9. Young, A.M., Perrewé, P.L.: What did you expect? An examination of career-related support
and social support among mentors and protégés. J. Manag. 26(4), 611–632 (2000)
10. Mancuso, C.A., Berman, J.R., Robbins, L., Paget, S.A.: What mentors tell us about acknowl-
edging effort and sustaining academic research mentoring: a qualitative study. J. Contin. Educ.
Health Prof. 39(1), 29–35 (2019)
11. Robinson, J.P., Shaver, P.R., Wrightsman, L.S.: Measures of Personality and Social Psycho-
logical Attitudes. Academic Press, San Diego, San Diego (1991)
12. Clynes, M., Corbett, A., Overbaugh, J.: Why we need good mentoring. Nat. Rev. Cancer 19(9),
489–493 (2019)
Mentoring Process: An Assessment … 525

13. Esbenshade, A. J., Kahalley, L. S., Baertschiger, R. Dasgupta, R., et al.: Mentors’ perspectives
on the successes and challenges of mentoring in the COG Young Investigator mentorship
program: a report from the Children’s Oncology Group. Pediatr. Blood Cancer 66(10) (2019)
14. Preston, J.M., Prieto-Flores, Ò., Rhodes, J.E.: Mentoring in context: a comparative study of
youth mentoring programs in the United States and continental Europe. Youth Soc. 51(7),
900–914 (2018)
Compensation Program: A Quantitative
Study of the Top Management Team’s
Level of Knowledge and Satisfaction

Neusa Maria Bastos Fernandes dos Santos , Rafael Maia Chagas ,

Alessandra Cristina Santos Akkari , Igor Polezi Munhoz ,
Fernando Fukunaga , and Ligia Kaori Matsumoto Hirano

Abstract This article aimed to evaluate the level of knowledge and the degree
of satisfaction of the top management team (TMT) of a global holding company
regarding compensation policies and programs. Thus, a questionnaire was devel-
oped and applied to 55 professionals who hold managerial positions in the target
company. The instrument was divided into three parts, encompassing demographic
data, knowledge level and degree of satisfaction. The data treatment included the
application of statistical tests (chi-square, Fisher significance, Spearman correla-
tion, and multiple linear regression), obtaining a correlation between the level of
knowledge and the degree of satisfaction with compensation policies and programs.
However, no influence of demographic profile on the degree of satisfaction of TMT
members was observed. The results showed that there is evidence that the degree of
satisfaction is directly related to the level of knowledge about the company’s compen-
sation program, despite being a weak correlation (rho = 0.4020). Moreover, the level
of knowledge and degree of satisfaction is independent of the area of activity (p =
0.469 and p = 0.615, respectively) and the different hierarchical levels (p = 0.824
and p = 0.957, respectively) of TMT members. Also, variable compensation proved
to be the most important measure within the definitions of a compensation program
(R = 0.7760), including the level of knowledge as a measure. This paper points
to the need for future studies that address ways to improve corporate transparency
regarding compensation programs.

Keyword Compensation program · Satisfaction · Top management team

N. M. B. F. dos Santos · R. M. Chagas · F. Fukunaga · L. K. M. Hirano

Pontifical Catholic University of Sao Paulo (PUC-SP), Sao Paulo, Brazil
e-mail: [email protected]
A. C. S. Akkari
Mackenzie Presbyterian University, Campinas, Brazil
I. P. Munhoz (B)
Federal Institute of Education, Science and Technology of Sao Paulo (IFSP), Sao Paulo, Brazil
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 527
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_52
528 N. M. B. F. dos Santos et al.

1 Introduction

The issue of remuneration within companies has gained importance and great atten-
tion from academic studies in recent years, mainly considering the analysis of their
amounts (excessive or not), the architecture of total remuneration programs and rela-
tionships of conflicts between agents. Usually, studies on compensation focused on
the company’s strategic level (top management team—TMT) relate the company’s
financial performance and the individual performance of professionals occupying
these levels [1].
In addition to performance reviews, other factors may influence the success of
full compensation programs, such as the professional profile and level of knowledge
of the eligible about the component items of the programs. Within the components
of a total compensation program, three main factors are identified: fixed compensa-
tion, variable compensation (short and long term) and benefits, and the impact on
satisfaction may or may not be equal [2].
Other factors may influence the degree of satisfaction of professionals of the corpo-
rate management team, which would not necessarily be linked to financial remuner-
ation, such as career development, level of knowledge about the total remuneration
program itself, hierarchical level, area of activity, age group, time in company, gender,
education, and place of operation [3].
For an analysis that covers factors other than the most commonly studied, a
survey was conducted in a multinational company in the logistics segment. Addi-
tionally, a literature review on the subject found that academic research on compen-
sation programs in companies is infrequent and, when conducted, have quantitative
objectives related to the earning capacity of professionals and impacts on companies.
Therefore, the objective of this study was to evaluate the level of knowledge and
satisfaction of TMT members regarding the company’s compensation policies and
programs, as well as to identify the sociodemographic profile of the professionals
and to analyze correlations.

2 Theoretical Framework

The concept of job satisfaction has been the subject of studies since the mid-1930s,
due to the models proposed by Taylor and Fayol and the appreciation of the human
factor. For these two scholars, motivation and job satisfaction would result from
premiums and salary improvements; for Elton Mayo, workers had their own goals
and objectives. Many studies have attempted to conceptualize job satisfaction and
motivation, however, a widespread concept would be Locke’s (1976) conceptualizing
job satisfaction as “an emotional state, positive or pleasure, resulting from work or
your work experiences” [4].
Recently, other concepts have brought to light a multifactorial sense of worker
motivation as the instrument presented by [5], which is based on five dimensions
Compensation Program: A Quantitative Study … 529

(satisfaction with colleagues, satisfaction with management, satisfaction with salary,

and satisfaction with promotions, satisfaction with the nature of the work).
The compensation program of executives is a topic that has gained great impor-
tance and discussion in recent decades. Jensen and Murphy [6] argue that changes in
the economy and global operating environment of companies as the main responsible
for this relevance and transformations.
In the 1970s, the executive compensation structure was basically a monthly salary
and bonuses tied to corporate accounting results, treated individually and without
details of the professional’s valuation metrics and parameters.
Over the next ten years, no major changes in executive compensation programs
were observed, with the 1980s being the period of major acquisitions and mergers,
one of the main responsibilities of executives to operationalize these movements [7].
It was in the 1990s that executive compensation programs underwent profound
changes, linked to their structure. Long-term incentives as a form of variable remu-
neration payment have gained widespread popularity in publicly traded companies
[8]. Another item that makes up the compensation program, and which companies
maintained their structures, was the short-term reward, such as bonuses and PLR
payment linked to financial results.
In a contemporary view, the 2000s were marked by events involving large compa-
nies in accounting scandals. Backed by fraudulent accounting and forged income
statements, these companies were in surplus shares and, after uncovered irregularities
plunged into bankruptcy [7]. Pointed out as one of the causes of the situation, exec-
utive compensation systems have been criticized for benefiting short-term results,
lack of concern about business continuity and errors in the communication process
(both internal and external) [8].
Additionally, according to Murphy [9], there is great heterogeneity in remunera-
tion practices. However, executive programs often contain four components that are
often not fully known to the employee, namely base salary, annual bonus (linked to
accounting performance), and long-term actions and incentives (including restrictive
action plans and bonuses based on multi-year accounting results).

3 Methodology

To assess the level of knowledge about company compensation policies and programs
and degree of satisfaction of TMT members, as well as to identify the sociode-
mographic profile of the professionals, a survey was conducted, which included
the application of a questionnaire to a logistic company in Brazil featuring a case
study. The study company belongs to is the main provider of moving, warehousing
and transportation services in Brazil. This organization is part of a global holding
company whose main business is courier services, commercial banking, express
delivery, freight management, and logistics solutions. With German-origin capital,
the logistics solutions division is present in 60 countries with 120,000 employees.
530 N. M. B. F. dos Santos et al.

In Brazil, it has business units in all regions (11 states) and serves the automo-
tive, chemical, energy, consumer goods, pharmaceutical and health, and technology
segments.
The questionnaire was sent electronically accompanied by an invitation letter and
informed consent, presenting the study and its objectives as well as the guarantee
of anonymity, giving the choice of participation or not as requested by the Research
Ethics Committee of Pontifical Catholic University of São Paulo (PUC-SP).
The instrument applied had thirty questions, divided into three parts, and was sent
at random, without identification and without differentiation between the regions of
the country, to 70 TMT professionals from the study model company
The first part of the questionnaire included 12 questions about sociodemographic
data.
The second part addressed ten questions to measure the level of knowledge about
the company’s compensation policies and programs. There were three answer options
for each question: “no”, “partially” and “yes”. In order to measure this level, a
point-based counting method was developed, based on the Likert method [10], and
the alternative “no” represented 1 point, the alternative “partially” represented 2
points and the alternative “yes” represented 3 points. The sum of the points obtained
represented the level of knowledge through direct proportion.
The third part addressed eight statements to measure the degree of satisfaction
of TMT members in relation to the company’s compensation policy and program.
A 5-point Likert scale and the same scoring method were applied [10]. The sum
of the points obtained represented the degree of satisfaction with the policy and
total compensation program practiced by the company through direct proportion. In
obtaining the results, the following hypotheses were tested:
H1: The different hierarchical levels and the different areas of activity are equally
aware of remuneration policies and programs.
H2: Different hierarchical levels and different areas of activity are equally satisfied
with compensation policies and programs.
H3: There is a direct correlation between the level of knowledge and the degree
of satisfaction.
To confirm or reject the above hypotheses, Fisher’s statistical significance test was
applied between the score obtained in the Likert scale-based score, by hierarchical
level and area.
Descriptive statistics were contemplated with the distribution of variables in abso-
lute and relative frequency tables according to the stratification of hierarchical levels
and areas of activity. For the purpose of classifying the level of knowledge and degree
of satisfaction with the company’s compensation policy, the Likert scale indices for
these variables were distributed in percentiles (25, 50 and 75).
The assessment of the association between the degree of knowledge and level
of satisfaction and the areas of practice and functional hierarchy was obtained by
applying the chi-square test, after classification according to the cutoff value (75th
percentile), and by the non- parametric Kruskal–Wallis analysis, comparing the
absolute indices of each of the dependent variables.
Compensation Program: A Quantitative Study … 531

The correlation between the degree of satisfaction and level of knowledge was
obtained by applying the Spearman correlation coefficient for ordinal samples. The
relationship between demographic variables and the degree of satisfaction and level
of knowledge was assessed by contingency tables using the chi-square test to verify
differences in frequency distribution between groups.
In order to verify the influence of remuneration components on the participants’
satisfaction level, a multiple linear regression model with step-by-step variables input
was used, based on the absolute values of the answers to the questions included in
the third part of the questionnaire.
For all statistical evaluations, SPSS software (version 18.0, IBM Inc., USA) and
a significance level of 5% were used.

4 Results and Discussion

Among the 70 questionnaires sent, 55 were completed, which characterized the

research sample. Regarding the characterization of the sample, 56.4% of the partic-
ipants are aged between 30 and 40 years and 7.3% between 50 and 60 years. In
addition, most of them are graduated in business administration (43.6%), have a
specialization degree (72.7%) and work in the southeastern region of the country
(70.9%). Also, the respondents were mostly male (87.3%).
Regarding the level of knowledge, the minimum score obtained was 12 and the
maximum was 30. Also, the minimum score obtained was 18 and the maximum 39
for the degree of satisfaction for the TMT members.
To classify the knowledge levels and satisfaction levels, in relation to the compen-
sation program, the 25pc, 50pc, and 75pc percentile cuts were used. Thus, for the
level of knowledge it was obtained 17 points (25pc—low knowledge), 21 points
(50pc—moderate knowledge) and 23 points (75pc—high knowledge). For the degree
of satisfaction, it was obtained 24 points (25pc—low satisfaction), 31 points (50
pc—moderate satisfaction) and 35 points (75pc—high satisfaction).
Table 1 presents the general results of the application of the questionnaires with
the absolute numbers of the universe, unanswered questionnaires, research sample,
minimum and maximum score in relation to the level of knowledge and degree of
satisfaction, besides the points that represent the distribution in percentiles (25, 50
and 75).
From the test of statistical significance obtained results of p > 0.05 (p = 0.824 and
p = 0.957) and showed that there was no statistical significance when comparing the
level of knowledge and degree of satisfaction (respectively) between the hierarchical
levels. Also, it was demonstrated that there was no statistical significance (p = 0.469
and p = 0.615) when comparing the level of knowledge and degree of satisfaction
(respectively) between the internal areas of the study target company, confirming
Hypothesis 1 and 2.
532 N. M. B. F. dos Santos et al.

Table 1 General data and score that delimits the percentiles in relation to the level of knowledge and
degree of satisfaction regarding the compensation policies and program, considering the responses
of TMT members
Knowledge level Degree of satisfaction
Universe 70 70
Did not answer 15 15
Research Sample 55 55
Minimum score 12 18
Maximum score 30 39
Percentiles
25pc 17 24 Low knowledge and
satisfaction
50pc 21 31 Moderate knowledge and
satisfaction
75pc 23 35 High knowledge and
satisfaction

In fact, Lawler [11] addresses the employee’s “high involvement approach” to the
business and encourages the creation of an environment in which individuals care
about company performance, have the information they need to influence it and, as a
result, are rewarded. To contribute to these actions, employees must have knowledge
about the policies and compensation program applied by the company. However, the
results found showed that even at different hierarchical levels or areas of expertise
the level of knowledge and degree of satisfaction with the compensation policies and
the program is the same, that is, there is no statistical significance that differ the level
of knowledge and degree of satisfaction by hierarchical level and by area of activity.
Even when we grouped the areas into two larger groups (operating areas and support
areas) there was no statistical significance (p = 0.281 and p = 0.280, respectively).
The correlation assessment between the level of knowledge and degree of satis-
faction of compensation policies and the program was found to be significantly weak
positive (0.402), thus confirming Hypothesis 3. Once verified that it is a significant
correlation, we proceed to analyze the magnitude of this association, estimated to be
40.24%, that is, about 40% of the satisfaction index variation occurs concomitantly
with variations of the index of knowledge, a poor correlation.
Corroborating this result, a weak correlation between TMT member’s knowledge
of compensation policies and program and their degree of satisfaction was observed
characterized by the Spearman correlation test (rho = 0.4020).
Thus, statistically, a correlation was found between knowledge of compensa-
tion policies and programs and the degree of satisfaction, corroborating studies in
the literature that point to the participation of knowledge in strategic compensation
approaches. However, this correlation was characterized as weak by the Spearman
correlation test (rho = 0.4020), i.e., there are factors of greater relevance to the
Compensation Program: A Quantitative Study … 533

strategic approach to remuneration. The study of Steiner et al. [12] found that job
satisfaction is often influenced by factors that go beyond remuneration and effort.
In fact, Wood and Picarelli [13] stated that the strategic approach to remuneration
reflects the reality of companies operating in the globalized and competitive world,
and considered the entire organizational context, from company strategy, structure
and management style. Thus, knowledge about compensation policies and programs
becomes a critical factor for the use of the strategic approach.
Assessing the level of knowledge of an organization’s TMT members about
compensation policies and programs can identify the transparency and/or interest of
this group regarding this topic. The communication process of policies and programs
is essential to ensure their successful implementation, as it ensures a full under-
standing of the compensation model and the gains it establishes. The communica-
tion strategy can be composed of two strands: first, show employees the value of the
total compensation package (salary, benefits, bonuses, retirement plan, etc.); second,
explain the business context and be transparent by comparing the total compensation
package with the companies used in salary surveys and studies [7].
Hipólito [2] states that among the relevant points of the compensation strategy,
the profile of the TMT members is one of them. This profile may facilitate or hinder
the performance of compensation (satisfaction) policies and programs. However, it is
noteworthy that, in addition to the profile of the TMT members, another relevant point
for the success of companies’ compensation policies and programs is the alignment
between the human resources area and the other areas. In fact, the role of the human
resources area within the organization has a great influence on this alignment [14].
Considering the results obtained on the measures that satisfy TMT members in
relation to the compensation program, it was observed that short-term variable remu-
neration was cited as the most relevant measure of total compensation programs (R
= 0.7760), followed by fixed remuneration (R = 0.8760), job structure and profes-
sional development (R = 0.9160), benefits (R = 0.9370) and level of knowledge of
the policies and program applied (R = 0.9460).
It was also found that regarding age, the level of knowledge is lower in the range
of 30–40 years, which is repeated in the degree of satisfaction.
Considering the time in the current position, the level of knowledge is lower in the
range of 5–10 years and the degree of satisfaction is lower in the range of 1–3 years.
Regarding the number of direct reports, the level of knowledge is lower for those
with 10–15 employees and the level of satisfaction is lower for those with 15–20.
From a survey with data of 21,670 participants, found that according to the
increase of the promotion of these individuals in the organization, their satisfac-
tion decreased. However, with age, these individuals became more satisfied as
they changed companies. In this dynamic, the work pointed to the importance of
remuneration [15].
534 N. M. B. F. dos Santos et al.

5 Conclusion

Based on the objectives of assessing the level of knowledge and the degree of satis-
faction of the TMT members of companies in the logistics sector located in Brazil,
by hierarchy and area of activity, in relation to the compensation program, there are
the findings listed below.
First, it was observed that the level of knowledge regarding the compensation
policies and the program was not influenced by the hierarchical level or area of
activity and was below the 75th percentile, which meant a low/moderate level of
knowledge. Still, the degree of satisfaction regarding the compensation policies and
the program was not influenced by the hierarchical level or the area of activity and
was below the 75th percentile, which meant dissatisfaction in a significant portion
of the population studied.
The level of knowledge about the variable remuneration influences TMT
member’s satisfaction, although it is weak. Items related to the sociodemographic
profile of the TMT members do not show an influence on the level of knowledge and
the degree of satisfaction about the compensation policies and program.
Therefore, it was identified the need for future studies that refer to ways to improve
the transparency of companies in relation to compensation programs, determining
whether this situation is inherent in the business communication process or refers
to the cultural attitude of employees not to seek the necessary information. Another
point to be researched is the alignment between the compensation program and the
expectations of the TMT members, analyzed under the spectrum that these factors
may be related.

References

1. Clot, S., Grolleau, G., Méral, P.: Payments vs. compensation for ecosystem services: do words
have a voice in the design of environmental conservation programs. Ecol. Econ. 135, 299–303
(2017)
2. Hipólito, J.M.: Administração salarial: a remuneração por competências como diferencial
competitivo. Atlas, São Paulo (2001)
3. Frey, B., Oberholzer-Gee, F.: The cost of price incentives: an empirical analysis of motivation
crowding-out. Am. Econ. Rev. 87(4), 746–755 (1997)
4. Thaler, R.H.: Mental accounting matters. J. Behav. Decis. Mak. 12, 183–206 (1999)
5. Siqueira, M.M.M.: Medidas do comportamento organizacional: ferramentas de diagnóstico e
de gestão. Porto Alegre: Artmed 16, 265–273 (2008)
6. Jensen, M.C., Murphy, K.J.: Performance pay and top-management incentives. J. Polit. Econ.
98(2), 225–264 (1990)
7. Jensen, M. C., Murphy, K. J., Wruck, E. G.: Remuneration: where we’ve been we got to here,
what are de problems and how to fix them. Negotiation, Organizations and Market Research
Paper Series, Harvard Business School (2004). Jensen, M. C., Murphy, K. J., Wruck, E. G.:
Remuneration: where we’ve been we got to here, what are de problems and how to fix them.
Negotiation, Organizations and Market Research Paper Series, Harvard Business School (2004)
8. Bebchuk, L.A., Fried, J.M., Walker, D.I.: Executive compensation in America: optimal
contracting or extraction of rents?. National Bureau of Economic Research, Cambridge (2001)
Compensation Program: A Quantitative Study … 535

9. Murphy, K. J.: Executive compensation. In: Ashenfelter, O., Card, D. (eds.) Handbook of labor
economics, 3rd edn. North Holland, Amsterdam (1998)
10. Stefano, N., Righi, A. W., Lisboa, M. G., Godoy, L. P.: Utilização das dimensão de qualidade
e escala de likert para medir a satisfação dos clientes de uma empresa prestadora de serviços.
Anais do XXVII Encontro Nacional de Engenharia de Produção (2007)
11. Lawler, E.E.: Paying the person: a better approach to management. Human Resource Manage.
Rev. 1(2), 145–154 (1991)
12. Steiner, E., Russel, S., Vizek, A., Warlick, A.: Crew in the West Coast groundfish catch share
program: changes in compensation and job satisfaction. J. Coast. Manage. 46(6), 656–676
(2018)
13. Wood, T., Picarelli, V.: Remuneração Estratégica, 1st edn. Atlas, São Paulo (1996)
14. Urbancová, H., Šnýdrová, M.: Remuneration and employee benefits in organizations in the
Czech Republic. Acta Universitatis Agriculturae et Silviculturae Mendelianae Brunensis 65(1),
357–368 (2017)
15. Riza, S.D., Ganzach, Y., Liu, Y.: Time and job satisfaction: a longitudinal study of the
differential roles of age and tenure. J. Manag. 44(7), 2558–2579 (2018)
Physical Ergonomics Applied
to the Administrative Sector
and the Factory Floor: The Case
of a Food Industry

Kélvin Santos Aguiar , Maria Thereza de Moraes Gomes Rosa ,

Luiz Vicente Figueira de Mello Filho , João Carlos Gabriel ,
and Alessandra Cristina Santos Akkari

Abstract There is an increase in the number of individuals with disorders related

to the spine and such dysfunction is classified as the second most common factor
of leave in companies. Thus, this work is based on physical ergonomics and its
relationship with musculoskeletal diseases alleged by workers of a multinational
food industry, which was used as a study model. This research was conducted in
the administrative area and on the factory floor of the model company, applying
the validated Portuguese version of the Nordic Musculoskeletal Questionnaire, as
well as observational methods, which aimed to better understand the processes of
the industry regarding ergonomic scratches. Among the results, a difference was
observed between the distributions of the body regions with the highest incidence
of pain, so that, in the body sector, the discomforts were located mostly in the neck
(75%), while in the factory pain was found mainly in the upper limbs. shoulders
(40%), hip (14%), ankle/feet (14%) and upper back (14%). The observations made
in loco corroborate the data obtained by the research instrument since there was
excess load lifting (factory), excessive static work, repetition of tasks, improper
posture, among others. As an optimization proposal, it is suggested the creation of a
work gymnastics program and the use of quick massage, so that workers can perform
better in their daily routines based on well-being, health, and safety.

Keywords Physical ergonomics · Musculoskeletal diseases · Food industry

K. S. Aguiar · M. T. de Moraes Gomes Rosa · L. V. F. de Mello Filho · J. C. Gabriel ·

A. C. S. Akkari (B)
Mackenzie Presbyterian University, Campinas, Brazil
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 537
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_53
538 K. S. Aguiar et al.

1 Introduction

In order to avoid displacements and employee health hazards, organizations should

consider a holistic view of the process, with companies always abiding by the Regu-
latory Standards to ensure the well-being of people within organizations, following
Ergonomic precepts. In this sense, it is of great value to understand and evaluate,
in different work environments, how physical workload influences workers’ lives
regarding welfare, safety, production effectiveness, and other aspects.
Given a historical context, it is possible to observe an increase in the number of
people with spinal dysfunction caused by the working conditions. In fact, muscu-
loskeletal disorders have the highest statistical indices, being the second most
common reason for work leaves [1].
Brazil has a number of 10% of the total chronic health problems occupied by
low back pain, which is the disease that most invalidates the population. In the
international scenario, especially in the United States, it is the eleventh reason for
hospitalization, which affects people between 20 and 45 years old, thus disabling
them to perform activities, and the most affected victims are usually men in the age
range 35 years old. May be cited as the positions of greatest risk of developing such
dysfunction, general service workers, truck drivers, domestic workers, mechanics,
garbage, nursing assistants, and construction workers [2].
Ensuring the well-being, health, and safety of workers, as recommended by
Ergonomics, collaborates to optimize levels of productivity, quality, and efficiency
in performing tasks, enabling the employee to achieve personal satisfaction, as well
as reach with the success of organizational objectives [3, 4].
In light of the above, the present work aims to evaluate aspects of physical
ergonomics in the organizational environment, using as a study model a food industry.

2 Background

Ergonomics is the area of knowledge that relates to the interconnections between

humans and other elements of a system, aiming to achieve an optimization of human
well-being and overall system performance. Analyzing the domains of specialization
of Ergonomics, there is the interaction between different fields of action and disci-
plines, encompassing the social, organizational, environmental, biological, among
others.
Focusing on the physical ergonomics domain, it can be observed that this is
somewhat related to the areas of human anatomy, physiology, anthropometry, and
biomechanics. Some of the most interesting subjects to address include material
handling, posture at work, skeletal muscle disorders, and repetitive movements [5].
Every activity performed by workers during their work hours implies the perfor-
mance of effort, which can be mental or physical. The importance of knowing the
limits of each worker is latent so that there is no physical and/or mental overload,
Physical Ergonomics Applied to the Administrative … 539

avoiding weaknesses and damage in different ways to the worker and the organization
[6].
Physical workload refers to the physical requirements that an employee is exposed
to during his or her workday. To improve understanding of the physical workload, it
is necessary to evaluate the physical efforts to which the worker is subjected, so that
by analyzing the energy consumption and heart rate of a collaborator one can infer
the degree of difficulty to perform a particular task [7, 8].
The sector that covers the food industry, which includes food products and also
drinks generally, corresponds to approximately 25% of the accidents that occurred
in the industry between 2010 and 2012 in Brazil. This number can still be analyzed
in relation to the whole, indicating representativeness of about 6.5% of the total
accidents that occurred in recent years. Thus, it is noted the importance and needs
for ergonomics to be considered in industries and companies in general, especially
in the food industry, since the numbers indicate high rates of occupational accidents.
In this type of company, occupational hazards are the biggest candidates to cause
environmental discomfort due to poor working environment or operational processes
[9].
Environmental risks may include, for example, factors related to lighting, noise,
physical exertion, task repetition, radiation, monotony, improper posture require-
ments, humidity, and substances that may harm the employees’ body, among
others.
Besides chemical and biological risk, another type of environmental risk
(Brazilian Regulatory Standard 09) extremely important to be analyzed in the food
industry, and which was the main object of study of this work, are the ergonomic
risks, including postures, lifting/loading of loads, task repetition, static work, etc.
This situation can increase the degree of absenteeism in this type of organization.
Therefore, this paper aims to evaluate aspects of physical ergonomics in the
organizational environment, using as a model the Brazilian plant of the food industry.

3 Methodology

This is descriptive exploratory research and, as for the procedures, a case study was
developed, which was applied in the Brazilian plant of the food industry. The large
food industry was chosen mainly because it is a reference in its market, so that,
besides the plant located in the state of Sao Paulo (Brazil), the company has other
production plants in the U.S., China, and France. The industry operates in the market
for flavor enhancers for seasonings, sauces, etc.
The present study considered the plant operators in their different divisions of
the factory floor, as well as analyzing the professionals in the administrative sector,
in order to allow a comparative study. Many aspects that influence the physical
workload of employees were analyzed, such as lifting loads, static positions during
the workday, breaks, postures, height/conditions of the work, among others. There-
fore, the observational method was applied in the part-time journey of different
540 K. S. Aguiar et al.

areas. The non-participatory, systematic, and semi-structured method was chosen

so that a previous script was developed from some observation categories. On the
first day, 80 min were used for analysis in the factory sector, considering the pulp,
dryer, and mixture lines. On the second day, 40 min of analysis were devoted to the
administrative sector.
Also, the workers were submitted to the validated Portuguese version of the Nordic
Musculoskeletal Questionnaire (QNSO—Questionario Nordico de Sintomas Osteo-
musculares), after signing the Informed Consent Form. The QNSO instrument has
been validated in Portuguese [10] and represents the Brazilian version of the Nordic
Musculoskeletal Questionnaire (NMQ) [11] which has been widely applied to assess
the nature and severity of self-rated musculoskeletal symptoms.
The questionnaire was developed and sent electronically to the company’s
employees, in agreement with area managers. In the email, it was explained about the
exclusively academic objective of the study, the contact of the responsible researcher
was informed and it was also explained about the guarantee of the anonymity of
the participant, giving the employee the option to respond or not to the instrument
according to the general guidelines of the Research Ethics Committee of Mackenzie
Presbyterian University.
In the data treatment, the characterization of the sample was developed and the
descriptive statistics were applied (Fig. 1).

Fig. 1 Representative diagram of the methodological approach developed for this study
Physical Ergonomics Applied to the Administrative … 541

4 Results and Discussion

The sample obtained in the present study referred to 20 people, divided equally
between the factory floor and administrative sector, with greater representation of
males, as shown in Fig. 2.
With a 30% representation in the sample (Fig. 3), the most predominant age group
is between 30 and 35 years old, and within this percentage, 66.7% of the employees
were in the administrative sector and the rest in the factory. When the sample was
analyzed, it was noticed that 60% of the employees who are younger, in their 20s and
25s, are in the administrative sector, being distributed in the purchasing, marketing,
and planning areas. All respondents are over 45 years old is in the factory, centralized
in the quality control sector.

Fig. 2 Respondents by gender, considering the administrative sector and factory floor

Fig. 3 Respondents by age, considering the administrative and factory sector

542 K. S. Aguiar et al.

4.1 Data Analysis: Factory Floor and Administrative Sector

The analysis of the data collected from the application of the QNSO indicated that
70% of the respondents claimed that they felt pain, tingling or numbness in at least
one of the body parts, in the last 12 months, and the shoulders received the highest rate
of complaints with a representativity percentage of 43% (Fig. 4a). In fact, the work of
Santos et al. [12] also pointed to the shoulders as the region with the highest incidence
of pain for the shop floor worker. This result is in line with the observations made in
the food industry since it was possible to notice inadequate postures for lifting loads.
It was observed that the layout of some sectors of the factory can be a major
cause of injury to workers since the present format of machine distributions requires
manual efforts, which are not fundamental to the activities. For example, the package
is sealed manually and then the operator forwards the finished product for inspection.
Inspection equipment could be moved closer to the sealing area as according to the
current layout the operator must travel a distance of approximately 12 m to reach the
specific location. Each package (with the finished product) is about 25 kg. Therefore,
the more optimized the layout, the lower the physical effort of cargo transportation.
The observational analysis made it possible to identify a wrong degree of inclina-
tion of the worker’s spine (30º) when compared to that recommended in the literature,
which directs the lifting of the load so that the spine is at an angle of 90° to the ground
to decrease the incidence of force on the vertebrae. Some examples of future prob-
lems that may arise from such practices would be back pain, which according to the
National Institute of Social Security (INSS), is the factor that most generates work
leave. When poor posture is practiced frequently, besides generating discomfort, it
can promote problems in the muscles and joints [8].
Another point of note is that the worker needs to lower himself to feed the equip-
ment with raw material in the system, requiring physical effort to pick up the raw
material, lower it, and have to place it inside the intended compartment. In order
to avoid this physical exertion, it is recommended to use a workstation that has a
height adaptable to the employee, thus improving their well-being and decreasing
the frequency/incidence of the shoulder, back, knee problems.

Fig. 4 Respondents who reported pain in the last 12 months, according to the body region and
considering a factory floor and b administrative sector
Physical Ergonomics Applied to the Administrative … 543

Regarding their work or personal routines, 70% of respondents claimed that at no

time did the pains described prevent them from practicing their daily activities (both
personal and professional) in the last twelve months. The remaining participants
(30%) answered that they were prevented, at some point, from practicing activities
that are part of their routines, that is, they were impacted in some way.
When analyzed by gender, it is possible to observe that the female sample
presented the highest percentage of affirmative answers in relation to the impacts
and impediments of performing their routines, with a percentage of approximately
67%. In contrast, the male sample unanimously reported that they were not prevented
from practicing their personal or routine professional activities due to the declared
pain. It is also noteworthy that the age groups that were prevented from performing
any of the activities present in their routines were between 20 and 25 years and
between 40 and 45 years. The area with the most limitations due to pain was quality
control.
It was also noted that 70% of people responded that they had not been to medical
appointments in the last 12 months. On the other hand, it is observed that of the
percentage that went to some consultation during the period, 67% were women.
Thus, it was found that women are more susceptible to pain in the workplace,
increasing the likelihood of impacting their routines. In fact, a study with the Brazilian
insurance company Sul América found that women are the most affected with
back pain, which may have direct impacts on their productivity in the organization.
The study was conducted between May 2013 and March 2016, with the participa-
tion of more than 13,000 policyholders, and the female population had the highest
percentage of back pain (68.5%) and maybe related to hormonal variation [13].
It was observed that 70% of people said they did not feel any kind of problem/pain
in the week before the questionnaire was applied. Looking at the 30% who said they
felt some kind of problem, most were women (67%).
Also noteworthy is the excess of static work in the sector, i.e., through the on-site
analysis it was noticed that during the activities execution the operators are always
standing, generating fatigue, besides possible problems with repetitive movements.
After applying the QNSO instrument to the administrative sector, it was possible
to notice that 80% of the respondents felt pain, tingling, or numbness in at least one
part of the body in the last 12 months. Of all body parts analyzed, the one with the
highest pain complaint rate in the last 12 months was the neck, with 75% (Fig. 4b).
When analyzing work or personal routines, 80% of respondents said they were
never influenced by the pain described above. The remaining participants (20%)
stated that they were prevented at some point from practicing activities that are part
of their routines, that is, they were impacted in some way.
This observation is in line with the analysis made using the observational method,
since it was possible to verify an inadequate positioning of the spine during the
workday, especially of employees who use laptops. The workers begin their activities
in an acceptable position, however, as time goes by, there is a relaxation so that they
begin to tilt their body forward, tending the neck and affecting the spine. Analyzing
in a general scenario, the factor of bad posture in the administrative sector (mainly),
544 K. S. Aguiar et al.

due to the use of laptops and microcomputers, has been increasing more and more,
causing more workers to leave the company when analyzed.

4.2 Comparison Between the Analyzed Sectors

The administrative sector presented a percentage of 10% more than the factory when
studying the occurrence of pain in the body parts in the last 12 months. Analyzing the
locations of pain reported by employees, in the administrative sector, the pain was
felt more centrally in one or two regions (especially in the neck), while in the factory,
the pain was more distributed, according to respondents’ reports. This fact can be
understood since, in the administrative sector, the postures that are required of the
various functions are very similar, bringing similar consequences to their employees.
On the other hand, in the factory, there are quite different functions, requiring
different rhythms, repetitions, and postures, which can, therefore, generate more
dispersed pain, depending on the types of activities that will be performed by each
employee.
Noting the number of respondents who claimed to have been prevented from
performing any of their routine activities, it is noteworthy that in the administra-
tive sector only 20% of respondents gave affirmative answers. When analyzed in
the factory floor sector, there was an increase of 10%, thus reaching a percentage
of 30%. In the administrative area, most of the pain is caused by poor posture
during the workday, especially by people who use laptops. In the factory, besides the
postures, there is another factor that further aggravates the situation of workers, the
lifting/loading of cargo.

5 Conclusion

During the case study developed in the Brazilian plant of the large food industry, it was
noticed a diverse sample of gender and age, presenting a higher concentration in the
range between 20 and 40 years. The sample (20) was equally divided, with 50% of the
sample composed of administrative staff and the remainder by factory staff. A factory
floor and administrative sector are reported to suffer from bad postures, having long-
term consequences as well as some short and medium-term discomfort. However, the
factory sector has an aggravating factor, since, besides the bad postures and the static
and repetitive works, there is the load lifting/loading factor, impacting the distances
in the food industry. As an optimization proposal, it is suggested the development
of an occupational gymnastics program, so that the workers can perform better in
their daily routines, predicting some future injuries to occur. Another improvement
proposal, preventing muscle injuries, is the creation of the culture of quick massage
during work hours.
Physical Ergonomics Applied to the Administrative … 545

References

1. Hyunjin, B., Songa, S., Leea, D., Pyoa, S., Shinb, D., Leeb, G.: Musculoskeletal diseases of
heavy industrial workers. Phys. Ther. Rehabil. Sci. 6(2), 71–76 (2017)
2. Lax, M.: Occupational health in central New York state: a publicly funded occupational disease
outpatient clinic 25 years later. Rev. Bras. Occupat. Health 38(127), 149–161 (2013)
3. Kuijer, P.P., van der Beek, A.J., van Dieën, J.H., Visser, B., Frings-Dresen, M.H.: Effect of
job rotation on need for recovery, musculoskeletal complaints, and sick leave due to muscu-
loskeletal complaints: a prospective study among refuse collectors. Am. J. Ind. Med. 47(5),
394–402 (2005)
4. Gornostaeva, Z., Kushnaryova, I., Alekhina, E.: Increasing the Competitiveness of Small Busi-
ness Companies on the Basis of the Modern Information Platform. In: Popkova, E. G., Ostrov-
skaya, V. N (eds.) Perspectives on the Use of New Information and Communication Technology
(ICT) in the Modern Economy vol. 726, pp. 207–216 (2019)
5. Monteiro, M.S., Alexandre, N.M.C., Rodrigues, C.M.: Musculoskeletal, work-skeletal, work
and lifestyle diseases among style workers of a public health institution. J. Nurs. Sch. USP
40(1), 20–25 (2006)
6. Pinto, N.F., Murofuse, N.T., Carvalho, M.: Process and workloads and workers’ health in
sericulture: a review. Rev. Bras. Occupat. Health 40(132), 237–247 (2015)
7. Taelman, J., Vandeput, S., Vlemincx, E., Spaepen, A., Van Huffel, S.: Instantaneous changes
in heart rate regulation due to mental load in simulated office work. Eur. J. Appl. Physiol. 111,
1497–1505 (2011)
8. Widanarko, B., Legg, S., Stevenson, M., Devereux, J., Eng, A., Mannetje, A., Cheng, S., Pearce,
N.: Prevalence and work-related risk factors for reduced activities and absenteeism due to low
back symptoms. Appl. Ergon. 43(4), 727–737 (2012)
9. Vasconcelos, F.M., Maia, L.R., Almeida Neto, J.A., Rodrigues, L.B.: Risks in the workplace
in the bakery sector: a case study in two cookie industries. Gest. Prod. 22(3), 565–589 (2015)
10. Pinheiro, F. A., Troccoli, B. T., OAK, C. N.: Validation of the nordic musculoskeletal
questionnaire as a measure of morbidity. J. Public Health 36, 307–312 (2002)
11. Kuorinka, I., Jonsson, B., Kilbom, A., Vinterberg, H., Biering-Sorensen, F., Andersson, G.,
et al.: Standardised Nordic questionnaire for the analysis of musculoskeletal symptoms. Appl.
Ergon. 18, 233–237 (1987)
12. Santos, V. M., Santos, J. W., Alsina, O. L. S., Monteiro, L. F.: Application of the musculoskeletal
Nordic questionnaire to estimate the prevalence of work-related musculoskeletal disorders in
workers under temporal pressure. In: XXXVIII National Meeting of Production Engineering,
Curitiba (2015)
13. Vitta, A., Canonici, A.A., Conti, M.H.S., Simeao, S.F.A.P.: Prevalence and factors associated
with musculoskeletal pain in professionals with sedentary activities. Physiotherapist Mov.
25(2), 273–280 (2012)
Evaluation of Illuminance in Workplace:
A Case Study on the Production
of Multicolored Self-adhesive

Wellingtânia Domingos Dias , Taciana Ramos Luz ,

and Meinhard Sesselmann

Abstract Adequate lighting is important for the productivity, efficiency, and quality
of a process line. Its evaluation is indispensable to keep mistakes and consequently
accidents from happening as well as preventing the occurrence of physical and
psychological complications for the employee. It is known that work in the industrial
sector alone can be exhausting, just like any other production process which may
demand a work rate with high productivity and concentration requirements. In the
case of the production of multicolored self-adhesive labels, the lighting of the work-
place comes across as essential for the correct configuration of colors to be used in
the productive process likewise the visualization and differentiation of the various
colors during the final inspection product. The goal is to assure the best product
quality and a most satisfied client, being that this is the most compelling factor when
the wish is to keep oneself competitive on the market. This paper has as its goal the
development of a case study based on the analysis of the illuminance on worksta-
tions at a factory where multicolored self-adhesive labels are produced. The study
started with the identification of variables relating to how the level of illuminance
on the referred production sector meets the expected standards reported by the NBR
8995-1/2013 ABNT regulation. In this context and in the light of the evaluation of
the variable correctional actions were then proposed in furtherance to better the final
product quality control inspection, alongside providing greater visual comfort to the
employees, minimizing health jeopardy.

Keywords Illuminance · Industrial environment · Workstations

W. D. Dias (B) · T. R. Luz · M. Sesselmann

Post Graduate Department of Mechanical Engineering, Bioengineering Laboratory, Federal
University of Minas Gerais, 6627 Presidente Antônio Carlos Ave, Belo Horizonte, MG
31270-901, Brazil
e-mail: [email protected]
T. R. Luz
e-mail: [email protected]
M. Sesselmann
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 547
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_54
548 W. D. Dias et al.

1 Introduction

The lighting of workspace is a preponderant factor for the composition and spatial
perception and may provoke various forms of sensations upon individuals, e.g.,
fatigue, drowsiness, visual discomfort, irritability. On the whole, the level of lighting
directly affects the work done in particular spaces, which justifies the importance of
planning lighting projects through.
According to Truccolo [1], poorly dimensioned lighting projects in which the final
illuminance does not meet the truly desired environment requirements may reduce
the employee performance drastically and, in some cases, risk their health as well
as the final product quality. Iida [2] mentions that light is essential at the workplace
and reinforces productivity and final product quality as a result of the employee’s
satisfaction and comfort.
On the other hand, a well-planned industrial lighting project improves the
working environment, increases the safety of employees, provides better produc-
tivity, reduces operational costs, and, consequently, contributes to the preservation
of the environment since the saving in energy consumption is taking into account.
In order to carry out an industrial lighting project compatible with the employees’
needs, it is necessary to conduct ergonomic analysis aimed at correcting and
preventing problems that might affect the quality of life in the workplace. Although
the company may target their goals toward the optimization of productivity, this
process ought not to put the health of the employees under jeopardy. For this reason,
the implementation of given ergonomic actions may be seen as necessary; according
to Falzon [3], the ergonomics must balance the companies goals like efficiency,
productivity, reliability, quality, and durability; with the employee’s goals as safety,
health, comfort, satisfaction, and interest in work.
Likewise, it is important to emphasize that the industrial environment should
be appropriate to the standards in force, including those concerning the lighting
[4]. The Brazilian Association for Technical Standards [5], in specific the NBR
8995-1 rule, establishes the lighting requirements for indoor working environments
beneficial to the workers to perform their visual tasks in an efficient manner, safely,
and comfortably. Still, with regard to the rules, the lighting in a specific direction
may reveal yet unseen details of a visual task, leading to better visibility and a more
prompt problem-solving. This scenario is particularly important for tasks involving
fine texturing, e.g., engraving and embossing of paper.
The way to achieve an efficient but low-price industrial lighting ranges from
defining the level of illuminance required at the work environment for choosing
the more suitable types of lamps available and their degree of cost-saving. These
steps result in calculations which will, thus, lead to the required amount of lighting
instruments needed to supply the desired level of illuminance.
For the activities in the multicolored tags production, the regulation recommends
an average of 500 lx illuminance for printing and cutting automatic processes; 1500 lx
for multicolored printing color inspection; 1000 lx for typesetting; and 500 lx for
rewinding and shipping.
Evaluation of Illuminance in Workplace … 549

Dos Santos et al. [6] state that the brightest the lighting, the better the visual acuity.
Nonetheless, excessive lighting may degrade vision and even cause injury. Overly,
lit objects may be even more difficult to observe than the ones with non-sufficient
lighting shone upon. Nevertheless, if there is a rapid increase in lighting whenever
the eyes are still adapting to low luminosity, there may be reduced visual acuity.
To each adaptation circumstance, there is an optimal level of lighting. As relative
contrast is reduced, luminous intensity must be increased to maintain the acuity.
The greater the contrast, the clearer the stimulus is perceived. Theoretically, the
maximum contrast is represented by the instance in which the lighting in maximum
on the stimulus and non-existent throughout the adjacent region. It is an almost never
found instance in practice [6].

1.1 Objective

To evaluate the level of illuminance, while acknowledging NBR 8995-1 Brazilian

standard within an industrial-worksite, at which the activities of printing, “typeset-
ting” and “inspection” of multicolored self-adhesive labels are carried out.

2 Methodology

The present work was developed from the evaluation of illuminance in the worksite
of the production of multicolored self-adhesive labels. Under this perspective, the
following parameters were considered:
• Level of illuminance in the industrial environment adopted in accordance with the
Brazilian standard NBR 8915-1, which takes into account the activities carried
out in this worksite—evaluated under the ergonomics perspective.
• Practical confirmation by collecting data with a lux meter and comparing the levels
of illuminance collected to the ones listed on the standard in order to resolve any
doubts as to the correctness of the illuminance used in the working environment.
The present work was carried out, from August to November of the year 2017,
in a company located in the metro area of Belo Horizonte-Brazil that produces
multicolored self-adhesive labels.
The company demand emerged from the need to perform improvements in the
work environment and to promote the employees’ welfare, as well as to meet quality
certifications.
Regarding the evaluation of the levels of illuminance, a study case was carried
out within the production sector of the company, comprises 14 workstations.
In order to evaluate exposure to illuminance, “NBR 8995-1/2013: Lighting
of Workstations” recommendations were followed. A digital thermo-hygro-
anemometer-lux meter by the brand Lutron, model LM8000, serial AB33174,
550 W. D. Dias et al.

equipped with a light detector for more accurate data collection, and with spectral
responses within 0 and 20.000 lx, was used to verify the lighting parameters.
In effectuation of the measurements of the amount of light of the surroundings,
some care was taken regarding known factors which could lead the data collection
into mistaken responses.
As stated in NBR ISO 8995-1, another aspect worthy of the companies’ attention,
besides the gross amount of light in a given horizontal or vertical plane, it is the
uniformity of light that covers not only the activities carried out in the production
line but also the others have done around. According to Vianna and Gonçalves [7],
the industrial work requires a light uniformity to avoid visual problems due to the
contrast between light and dark lighting, where employees must make constant visual
adjustments.
Other aspects to be considered are mutability, color reproduction index, and the
color temperature. Mutability, flexibility, or versatility of light refers to the system’s
ability to adjust the lighting in accordance with the different needs specific to the
environment. Color reproduction is the variation that is observed from the color
vision of the objects under the illuminant study, in relation to a standard illuminant.
The index of color reproduction is essential in industrial sheds, depending on the
activity field, as in the printing industries, among others. Color temperature—warm,
cool, and neutral, measured in Kelvin (K) scale, is also a normative requirement.
For Vianna and Gonçalves [7], warm color sources are not advisable for manufac-
turing or office environments, once are observed a decrease in the pace of work by
the employees. For these production environments, cool or neutral light sources are
recommended. The company sector where the measurements were taken has only
artificial lighting; therefore, it was of no need the data collection in several daylight
times and/or on different seasons.
Fourteen workstations were evaluated, and the results of illuminance evalua-
tion were compared with the minimums reported by NBR 8995-1/2013 standard.
The data were collected; the illuminance level average values of each workstation
were obtained. This way, forty-two measurements were performed in each work-
station during the time interval of 10 a.m. through 11 a.m. of one single day in
September/2017. All evaluations of the workstations’ lighting were performed during
a normal working day.
For the data analysis, Microsoft Excel software was used. After collecting the
numbers, calculation of averages was made with the aid of spreadsheets in order
to compare the illuminance levels among the workstations’ values and weigh out
whether they were in accordance with the minimum requirements of NBR 8995-
1/2013 standards.
Table 1 shows the required levels of illuminance recommended by standard NBR
ISO 8995-1 for workstations according to the nature of the activity carried out in the
company under study.
Illuminance levels were registered after lux meter calibration and stabilization,
the latter being passive of taking either short or long periods of time depending on
variations on luminosity around the points of data collection. At each of these points,
three admeasurements were collected.
Evaluation of Illuminance in Workplace … 551

Table 1 Recommended illuminance for workstations by standard NBR ISO 8998-l

Environment—manufacturing Min illuminance (lx)
Printing 500
Typesetting 1000
Color inspection in multicolored printing 1500
Rewinding 500
Shipping 300

In order to define the location of the measurement points, the task area—by
definition, the partial area of the workstation within which the visual task is carried
out— was imaginarily divided into contiguous and identical squares, each of which
with the dimension of about 20 cm of side length—indicated number for small areas
[8]. The gauging is done in the center of each of these squares during the printing of the
adhesive labels; and the measurements for the illuminance levels on the surroundings
of the task areas were done through a strip range of at least 0.5 m of width within
the employee’s field of vision.

3 Results and Discussion

In the analysis of the company operations phase, visit was arranged in order to
collect information on the production process and to embrace aspects with reference
to the number of employees, workload, type of machinery run, and the sequence
of production flow. Next, three visits to each company sector were scheduled and
performed, comprising offices and production sectors altogether.
Observations and notes were held on the activities of workers who run the printing
machinery, before with the preparation and during the actual process, as well as the
“typesetting” and “color inspection” activities.
As by the analysis of the work environment conditions, despite the artificial
lighting permanently on, it was detected that the environment luminosity was not
the most adequate and did not offer good comfort conditions and thus employee
welfare required by this production process. Besides, the darker shade of walls and
shelves contributed to the darkening of the surroundings.
As for the spatial distribution, walls and shelves were located close to the
machinery, which can be responsible for the casting of shadows that may cause
harm to the employees’ eyes, mainly during the typesetting and color inspection
processes.
552 W. D. Dias et al.

Fig. 1 Representative image of the spatial distribution of the machinery in the production sector
areas

As pre-diagnosis, one might be accurate on saying that the disposition of

machinery (see Fig. 1) did not was framed well within the set spacing limits between
machinery and spare supplies shelves, according to the Regulatory Standard 12:
Machines and Equipment [9]. Printing machines were positioned too close to each
other and too close to walls, hindering maintenance. Due to this fact, it was found
out that, during the processes of typesetting and color inspection of tags, the lighting
presented was not adequate, compromising visualization of tasks by workers and
generating greater visual effort. Therefore, the focus of the analysis revolves around
the adequacy of lighting in the task area for the typesetting and tag color inspection
processes.
The lighting equipment was arranged in a localized configuration. This can be
described as a setting in which the light incites predominantly on a preferred direction;
the lamps are centered above and around the task area, providing high illumination in
a workstation or on a machine. However, the inspection workstation, which demands
lighting strictly necessary for a thorough visual examination, during and after the
printing of the multicolored labels, was not found to be lit by localized configuration.
Nevertheless, the activity station requires a 1500 lx illuminance is required.
Evaluation of Illuminance in Workplace … 553

Table 2 Results for the factor illuminance on workstations at the company under study
Production sector Number of workstations Average illuminance
Printing 7 343.55
Typesetting 6 363.20
Inspection 4 309.25
Rewinding 3 397.56
Shipping 3 215.67
Total average 325.85

3.1 Illuminance Measurements

The collected data were submitted to analysis and are separated on the following by
activity sector—printing, typesetting, inspection, rewinding, shipping. The admea-
surements found were compared to the advised minimum values indicated by the
ABNT NBR ISO/CIE 8995-1:2013 standard, in order to verify conformity.
The evaluation of illuminance on the workstations of the company under study
resulted in an average general level of 325.85 lx. Table 2 presents the average results
for the illuminance factor.
Based on Table 2, one can note that there was a significant variation of artificial
illuminance on areas where workstations were set. Therefore, it was possible to
observe that, for evaluation of illuminance, the physical environment exerts great
influence.
After the establishment of lighting patterns according to the norm to the industrial
environment values, it was possible the comparison with the average values measured
in each workstation (see Table 3).
With respect to points where an efficient illuminance is indispensable, the work
of the inspection presented the lowest average illuminance value. Data shown in
Fig. 2 makes clear that the workstations are not in conformity with the minimum
illuminance levels standards, with the exception of the shipping workstations, which
display levels a bit higher than the minimum uniformity range. As demonstrated, the
works need correction in order to reach adequacy to the norm.
The levels of illuminance at workstations were lower than the minimum estab-
lished by the NBR 8995-1/2013 standard, which indicates the company must adopt

Table 3 Comparison between results for the factor illuminance at the company under study and
the values indicated by the ABNT NBR ISO/CIE 8995-1 2013 standard
Production sector E—average illuminance measurements Illuminance (lx) Uniformity (0.7)
(lx)
Printing 343.55 500 350
Typesetting 363.20 1000 700
Inspection 309.25 1500 1050
Rewinding 397.56 500 350
Shipping 215.67 300 210
554 W. D. Dias et al.

Fig. 2 Graph—conditions of illuminance around workstations at the company under study. Average
illuminance refers to the average values obtained from data collected at the workstations (x);
illuminance (lx) (Y ) [1] and uniformity (0.7) [2] refer to the standard values extracted from the
norm

correctional actions to minimize or mitigate this fact. The present results resemble
those from De Oliveira et al. [10], who analyzed the lighting in a wood workshop; the
workstations, with the exception of only two, displayed levels of illuminance insuf-
ficient for the activities carried out. It is known that below the minimum limit recom-
mended; the meager lighting may cause visual fatigue, excessive effort and can induce
mistakes. Therefore, it was suggested that the wood workshop held adjustments in
the lighting system of the working environment.
In a study of Von Bommel et al. [11] performed in a metallurgical company,
results revealed that increasing illuminance from 300 lx to 500 lx could generate an
increase in productivity of between 3 and 11%; that is, the increase in illuminance in
the workplace brought improvements in task performance, reduction of waste, and
diminishing of number of accidents.
Following the same lead, Juslen et al. [12] examined the influence of lighting
in the employees’ work pace and in the quality of work in the sector of assembly
of electrical devices. Results showed that the work pace in the environment with
1.200 lx is 9% faster than that performed under the 800 lx detected during winter.
Based on a study done by IESNA [13], it could be observed that the accident
recidivism factor decreased by 30% when the lighting increased from 65 to 170 lx.
Such results demonstrate that there is a positive relationship between the increase in
illuminance and productivity.
On the other hand, Akbari et al. [14], from the Pearson correlation coefficient
calculated in his study, detected that the amount of lighting has no influence over
human productivity in an automotive parts’ assembly unit and even changes in
lighting have no influence whatsoever over the productivity of employees from other
units. This study is an exception; given that others such as one authored by Anizar
et al. [15] about a cigarettes’ industry were able to conclude that there actually is
a correlation between the amount of imperfect cigarette paper that goes through
inspection with the poor lighting at the lamination station, causing a decrease in
employee performance and in how much they are centered on the task they carry out.
Evaluation of Illuminance in Workplace … 555

3.2 Suggested Corrective Actions

Correctional actions were suggested to the multicolored self-adhesive labels

company in order to provide greater employee welfare, as well as optimized services
and an increase in final product quality. The study by Rangkooy et al. [16] reiterated
that illuminance-related correctional actions by many times are not efficient in a first
instance; therefore, a second and more beneficial suggestion for the lighting design
was made.
The adequacy proposals were: improving the company’s lighting design, mainte-
nance factor application done by trained professionals, performing maintenance on
the lighting system and on the surfaces at the work sectors—the adoption of light
colors for flooring, walls, and ceiling—and adjustment of physical arrangement.
As for the labels’ inspection activity, color is a paramount factor of quality, and,
therefore, when improving the lighting at the workstations the worker’s emotional
state is benefited, besides the reduction of individual work overload. Consequently,
there are great improvements in product quality for the final consumer.

4 Conclusions

With respect to the quantitative illuminance evaluation, it is concluded that all the
workstations in the industry of multicolored self-adhesive labels display average
levels of illuminance lower than the minimum required by the Brazilian standard for
the lighting of work environments.
This company demands a greater visual accuracy by workers when observing the
color quality of labels in resonance to the product’s final quality.
The field survey results show the existence of illuminance values in nonconfor-
mity with the NBR 8995/2013 standards in all workstations analyzed, which may
be impairing the final product quality. In the meantime, it is up to the company
to arrange the necessary corrections according to the standards referenced above,
seeking minimization or mitigation of this wronged factor.
In the view of the foregoing, it is concluded that the characteristics of a work
environment may alter the quality of the job carried out by the employee as well as the
final product quality, being that a workstation must be adequate to guarantee to service
quality offered by the employees together with their health. The company must
provide a pleasant environment mounted with maximum protection and security, in
order to favor the common welfare between company and collaborator.
So as to improve the company lighting system design project, a maintenance factor
must be applied, embedded by trained professional, performing maintenance on the
lighting system and on the surfaces throughout the work sectors, adopting light colors
for flooring, walls, and ceiling, in addition to adjusting the physical arrangement.
556 W. D. Dias et al.

Acknowledgements The authors thank the Federal University of Minas Gerais (UFMG) and the
Brazilian National Council for Scientific and Technological Development (CNPq) for all financial
support and the company that gave us the opportunity to study.

References

1. Truccolo, L. J.: Análise dos riscos de acidentes no setor de produção em uma empresa de
produção de móveis escolares. Monograph (Specialization)—Occupational Safety Engineering
Course, Federal Technological University of Paraná, Medianeira, 69 f. (2013)
2. Iida, I.: Ergonomia: projeto e produção, 2nd edn. Edgar Blucher Publisher, São Paulo (2005)
3. Falzon, P. (ed.): Ergonomia. Edgar Blucher Publisher. São Paulo, p. 640 (2007)
4. Padilla, J. R.: Iluminação industrial, alternativas para melhorar sua eficiência energética e facil-
itar a manutenção. Revista Lumière Electric ed. 157. Iluminação Industrial, Editora Lumière,
Eletron—PR, pp. 102–108 (2011)
5. ABNT: Associação Brasileira de Normas Técnicas. NBR ISO NBR ISO/CIE 8995-1.
Iluminância de interiores. Rio de Janeiro, ABNT (2013)
6. Dos Santos, A., Leta, F., Veloso, M.: Fatores de risco industrial causados por diferentes
percepções de cores devido à diferença de iluminantes. Revista Produção Online 5(1) (2005)
7. Vianna, N. S., Gonçalves, J. C.: luminação e Arquitetura. Geros s/c Ltda, 2th edn. SP, São
Paulo (2001)
8. Nogueira, A.: (a) Procedimento de trabalho DSA UASO-PE 14_02 L. Determinação dos níveis
de iluminância. (b) Ficha de Registo de Medição de Níveis Iluminância DSA UASO-IM 22_01
L. Determinação dos níveis de iluminância. (c) Folha de cálculo em EXCEL da Iluminância e
Uniformidade DSA UASO-IM 55_03 L. Occupational Air and Health Unit of the Department
of Environmental Health of INSA (2009)
9. ABNT: Associação Brasileira de Normas Técnicas. NR-12: Máquinas e Equipamentos. Rio de
Janeiro (2009)
10. De Oliveira, F. N., De Sousa, L. S., Gadelha, L. F.: Análise da iluminância em postos de
trabalho de uma marcenaria na cidade de Tabuleiro do Norte-CE. XXXVI National Meeting
of Production Engineering (2016)
11. Bommel, J.W., Van Der Beld, G.L., Ooyen, M.H.F.: Industrial Lighting and Productivity. Philips
Lighting, The Netherlands (2002)
12. Juslen, H., Wouters, M., Tenner, A.: The influence of controllable task-lighting on productivity:
a field study in a factory. Appl. Ergon. 38, 39–44 (2007). https://doi.org/10.1016/j.apergo.2006.
01.005
13. Rea, M. S.: The IESNA Lighting Handbook, 9th edn. Illuminating Engineering Society of
North America, New York (2000)
14. Akbari, I. J., Dehghan, H., Azmooh, H., Forouharmajd, F.: Relationship between lighting
and noise levels and productivity of the occupants in automotive assembly industry Hindawi.
Publishing Corporation. J. Environ. Public Health 527078, 5 (2013). https://doi.org/10.1155/
2013/527078
15. Anizar, K., Syahputri, R. M., Sari, I. R., Hardianti, D. A.: Changing lamp type and position
to improve lighting quality. In: IOP Conference on Series: Earth and Environmental Science,
p. 126 (2018). https://doi.org/10.1088/1755-1315/126/1/012050
16. Rangkooy, H. A., Mazaheri, A. H., Salehi, M. M. A., Khsnmirz, A. F.: Assessment and design
of illumination in a steel manufacturing company in Ahvaz, Iran. Jundishapur J. Health Sci.
9(4) (October, 2017). https://doi.org/10.5812/jjhs.64051
Smart City: A Qualitative Reflection
of How the Intelligence Concept
with Effective Ethics Procedures Applied
to the Urban Territory Can Effectively
Contribute to Mitigate the Corruption
Process and Illicit Economy Markets

Kelem Christine Pereira Jordão , David Bianchini , Yuzo Iano ,

Ana Carolina Borges Monteiro , and Reinaldo Padilha França

Abstract The public corruption and illicit market correspond to the greatest global
challenges to be overcome by governments worldwide (Transparency Interna-
tional 2017). According to the Corruption Perceptions Index 2017, for 180 coun-
tries/territories studied, most of them are making little or no progress in ending
corruption in the public sector. The public money laundered affects direct invest-
ments in infrastructure to improve the city’s performance. The corruption encourages
directly the illicit activities formation and the Organization for Economic Cooper-
ation and Development (OECD) estimates that hundreds of billions of dollars of
revenue from illicit activities are handled globally every year and was identified in
the context of National Risk Assessments as part of the major risks that could affect
countries in development. These practices have generated negative effects in society
and the cities are the context in which all these issues occur. Then, this paper has
aimed to conduct a qualitative reflection of how smart cities, which represent effi-
ciency and effectiveness places, can contribute through your intelligent environment
to mitigate the corruption process and illicit market operations.

K. C. P. Jordão (B) · D. Bianchini · Y. Iano · A. C. Borges Monteiro · R. Padilha França

School of Electrical and Computer Engineering (FEEC), University of Campinas—UNICAMP,
Av. Albert Einstein, 400, Barão Geraldo, Campinas, SP, Brazil
e-mail: [email protected]
D. Bianchini
e-mail: [email protected]
Y. Iano
e-mail: [email protected]
A. C. Borges Monteiro
e-mail: [email protected]
R. Padilha França
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 557
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_55
558 K. C. P. Jordão et al.

Keywords Smart City model · Corruption · Illicit market · Indicators · Ethical ·

Social · Technical · Digital aspects

1 Introduction

Periodically, within a Republican State, mayors are democratically elected in order to

start a new cycle for public management. They are chosen by a popular vote where it
can be through voluntary or compulsory population participation. The elected admin-
istrator has a period of four years as mayor to administrate de city, where it can be
renewed for a new period of time if approved again by popular vote. In theory, a public
administrator is chosen through the proposal’s comparative analysis that better meets
the city’s challenges and priorities. The proposals can be exemplified for improving
the inhabitant’s life quality, maximize the use of the city budget, maximize the service
quality provided to the city minimizing expenses, leveraging investments and innova-
tion, providing economic growth and social development for the municipality, as well
as the political, professional and social candidate’s background. Political alliances
between parties are established during an electoral process, in order to strengthen
candidates in the competition to assume the city hall; however, the elected public
administrator must have a political maturity to continue successful programs of the
previous administration, independent of the predecessor political party. Once elected,
the public administrator has with responsibility, among the growth and development
strategies of the municipality, the provision, and maintenance of public services that
guarantee the dynamics of a city’s urban systems, such as water, electricity, collection
and adequate treatment urban transport, sanitation, gas, telecommunications [1].
Basically, there are three municipality processual steps to contract the provision
and maintenance of public services: contracting, supervision, and payment. The
contracting is carried out mostly through bidding processes, where companies that
provide services or products, submit proposals that meet the city requirements. Once
the service or product has been contracted by the city hall, the public technical staff
is responsible to supervise the services or good provided, having as parameter the
requirements set in the agreement, attesting the delivery veracity. Finally, once the
service or goods have been delivered in compliance with the agreement requirements,
the payment occurs as previously established between the city hall and the supplier
[1–3].

2 The Public Corruption: The Typical Corruption Model

Found in Cities

Even though there isn’t a single model of public corruption, which may occur through
an inadequate supply, over-fledged or non-existent goods or services, the first charac-
teristic observed when studied the corruption in cities corresponded to the mapping
Smart City: A Qualitative Reflection … 559

(by corrupt management) strategic functions in public operations that may be manip-
ulated or fraudulent. Once identified, it is strategic to designate trusted person partic-
ipants of a corruption network, in order to operationalize the irregularities, regardless
of the corruption scheme to be established. Therefore, a second characteristic is iden-
tified: managers’ appointment (commissioned) in municipalities’ strategic positions
without technical skill or experience to the function [3, 4].
The urban systems in a city (transport, water supply, energy, sanitation, garbage,
etc.) are carried out through bidding processes established by the city hall. For
this, two usually procedures are used: (1) the “invitation letter,” for relatively low
purchases (when compared to other expenses purchase of the city hall) where is not
required a technical acquisition evaluation; (2) the “price evaluation” for purchases
goods and services with expressive values that require a technical staff analyze.
Though the “invitation letter” procedure, companies are invited by the city hall to
submit proposals, instead of the “price evaluation” procedure, where companies
specialized submit their proposals to attend the service [3–5].
The fraudulent method usually used by a corruption network (composed of public
administrators, commissioners, and private companies) is forging the participation of
competing for companies in a bidding process, by sending documentation of “nonex-
istent (ghost) companies” or inactive companies. This action has the purpose of rein-
forcing the company proposal that already has a previously established agreement
with the corrupt public administration. The second method most used and identified
in the study corresponds to fictitious bidding processes assembled and conducted
improperly by the public administration and public commissioners. A third method
identified corresponds to bids under extremely specific technical conditions, making
it difficult for other companies to submit a proposal, benefiting only the company(s)
that already has a pre-established agreement with corrupt management. All methods
aim to establish, within the formal and legal stages of procurement, the company that
will benefit a corruption scheme [3–5].
Thus, the third characteristic identified the city corruption corresponding to the
list of companies participating in a bidding process. All participants must be carefully
verified and confirmed through documentary analysis, direct contact and companies
visit the veracity of all participants. A technical commission must be formed by
specialists, with the objective of identifying any irregularities that may generate
suspicion. The company registration, the fee collection, and invoices issuance must
also be carefully checked to ensure the proper companies’ functionality, confirming
it physically and legally existence [3, 5, 6].
Once established a corruption company to supply with services or goods the urban
system by the corruption administration, the charge of them will be made through
the payment of the invoice. The invoice payment (overprice or an irregular invoice)
is divided between the members of the scheme, being paid through cash, in order do
not establish a banking link between the supplier and the public administration that
can be traced later through breach of bank secrecy.
Another feature mapped out in the city corruption scheme is regarding the forma-
tion of specialized gangs to defraud public biddings, set up vicious contests, cover
illegalities, encouraging the development of illicit markets, and contribution actively
560 K. C. P. Jordão et al.

to the municipality impoverishment [3]. The municipality impoverishment affects

directly investments to develop and improves the city’s performance, affects the new
job generation, resources to social, cultural and educational projects, and accumulates
money for the city.
A corruption administration affects indirectly the population’s perception of what
is being done with all taxes and duties payments, reinforcing a distrust and insecu-
rity environment of the public administration toward the inhabitants of the city. In a
distrusting environment, which the population feels robbed by the public administra-
tion, becomes a viable alternative for them, search for markets with cheaper prices
and paying taxes free. It is established with that, one of the other ways, for illicit
markets operation [7].

3 Illicit Markets: Working Against the City Prosperity

The illicit market is the existence, combined or isolated, of one or more categories of
goods classified as forbidden by the government and being consumed individually or
in society. The illicit markets are classified in five categories of human life, including
social (trafficking of illegal drugs, persons and arms); commercial (stifling innovation
and legitimate entrepreneurship); economic (distorting markets, diverting revenues
and legitimate employment to illicit ends); environmental (trade in endangered and
protected wildlife, timber, and toxic waste); and health (counterfeit pharmaceuticals,
drug dependency, pollution) [8]. The sectors most affected by illegal markets are
clothing, medicines, electron electronics, and auto parts.
Like any business, the illicit markets look for making profit and face the same
market challenges to have a competitive sale cost (as a structured supply chain, capil-
larity logistics/ distribution, production scale, and supplier’s diversity), but it not
incurs the research and development cost, a safety workforce, compliance produc-
tion, environmental aspects, and safety products regulations. Basically, an illicit
market demand (through the consumers perspective) is driven by aspects related
to the economic situation and consequently, budget constraints, product itself, price
and quality perceived, individual consumer characteristics (attitude toward illicit
markets), and finally by the institutional environment in which the consumer oper-
ates (risk of discovery and jurisdictions penalties, criminal laws and how efficiently
and frequently this law are applied by the authorities to combat the issue). Through
the supply perspective, illicit markets are driven by marketing opportunity, techno-
logical and logistics distribution challenges associated with an undertaking and the
risk involved. The globalization economic is another aspect that potentially, foment,
and viability this market (mainly with as counterfeit products around the world)
through of easy access to resources, technology progress, and the rapid growth of
e-commerce trade (supported by the rapid increase of different devices) [8].
Additional of that, the illicit markets touch directly the city when increase of
disorder caused by irregular occupancy of the public space by drugs consumers,
illegal street market causing (besides the illegal commerce) an unsafety and insecurity
Smart City: A Qualitative Reflection … 561

place for citizens, public and private immobile devaluation, economic inhibition,
facilitating the degradation of social environment in the city, affecting the innovation,
the employment creating, an economic growth long term and explore incorrectly the
environment and ecological sustainability [8].
Much of the concepts and technologies used by a Smart City can combat
the corruption process formulation and the illicit markets, through a cultural and
organizational cities changes [9].
The Smart City concept can organize and discipline the citizens routine through a
city workflow process (interconnectivity between news and existing process), govern-
ment transparency, a bottom-up and top-down culture between the population and
administration, a efficient cultural and social education about the illegal and risk situ-
ation involving the population, rescue of the city to citizens (valorization of what is
produced locally in the city itself), intensify the use of all kind of devices to control
the routine of the place and the population (cameras, sensors, police application,
etc.), and by the efficiency that the place will systemic and gradually present, etc.
[10].

4 Smart City Concept and Technologies

The concept of Smart City is becoming more and more relevant for academics,
policymakers, and governments, by the consistent changes generated in the urban
territory related to the functionality and improvement of numerals systems existed
in the city [11]. Some recent research has been proposing a Smart City model,
suggesting understanding and quantify the different offers and demands existed in
the Same City related with infrastructure, technologies, and social/cultural aspects,
to provide specific solutions to transform the city without losing this systemic urban
and unit vision [10].
The Smart City model promotes to work, plan, and organize the city through
sub-centers (or subgroups), without losing their systemic and unit vision, identifying
their existing and necessary infrastructure, investing to develop a digital network
and mapping the social and cultural population aspects. This model reflects the idea
that a Smart City—where people, technology, and process are connected—can only
enjoy all efficiency of technology capability available, when developing aspects of
your social and cultural intellectuality, as a tool that differences them of a simple
city [12]. In this context, telecommunications systems with a low level of abstraction
and low computational consumption are essential to connect people and systems and
favor the implementation of Internet of things (IoT) to people’s daily lives [13]. In
addition, new methodologies related to the environment and health are also part of
the concept of smart cities, because through innovative methodologies blood tests
can be performed through computers and their results sent to doctors easily and
quickly regardless of income and place where the patient is located [14–16]. Thus,
the premise of this Smart City model is your citizens, as a key to connect, explore,
562 K. C. P. Jordão et al.

suggest, and increase the dynamic of urban territory, achieving the efficiency, efficacy,
and synergy among process, systems and devices available.
The proposal of this concept has structured the analysis of the city no longer in a
macroscale, where the territory is visualized in its totality, but in a microscale, where
the local planning and the identity of the studied place are rescued, aiming at to the
individual perception of the needs and challenges of each sub-center of the city [17]
Some essential characteristics are observed and planned on this microscale:
Operational aspects: Smart Cities, in contrast to centrally planned cities, are orga-
nized on a decentralized model (sub-centers or subgroups) but connected to each
center composing a unique city [10]. In the meantime, it should seek to explore
concepts in the territory that will work to increase the density and their multi-
centricity—since the subcenters are generators of employment, infrastructure opti-
mization, housing, education, health, security, etc., within the same city—providing
the better use of existing empty spaces in regions provided with urban infrastruc-
ture and thus minimizing the need for new investments in infrastructure to meet
the demand of the city. When well-worked the subgroups density, it also promotes
a more efficient urban mobility system when it is planned to integrate the terri-
tory. This is possible by identifying and connecting the existing multi-centers within
the territory, by exploiting with the better performance the use of collective public
transport and by providing alternatives for mobility, which are accepted in the daily
life of the population by transmitting security, reliability, and practicality Examples
are the cycle paths implemented to link the dynamics of the city (residential areas
with commercial areas, universities, industrial, etc., and not just segregated bicycle
paths in neighborhoods, squares, and parks) and the use of walking as an alternative
modal, sharing of vehicles, etc. Such a change in behavior will naturally contribute
to reducing the number of vehicles on the street, thereby reducing congestion exten-
sions, and increasing the accessibility and use of alternative modes. Significant gains
in air quality in cities by reducing the emission of gaseous pollutants, consequently
improving air quality by reducing the level of greenhouse gas emissions per capita
[18].
Social aspects: Each subgroup should be closely linked to the development of
knowledge, research, innovation, culture, arts, economics, technology, of the place
studied and have in these areas their raw material to differentiate them from others
by stimulating the creation of ideas, building a culture of knowledge and intelligence
in all segments of the sub-center and thus rescuing the strengthening of the city’s
competitive advantages as a larger group [19]. The intelligent city possesses in its
DNA concepts and elements sustained by the creativity of individuals, companies,
capital, institutions of knowledge, academia, governments, educated and productive
population, etc. [20]. It is through the contributions promoted by the creativity that
a city can maintain its identity, develop its characteristics, and stimulate its potential
toward differentiation and competitiveness even in the face of globalized competition.
This is only possible by encouraging creativity, providing access to new and varied
knowledge, as well as respect for diversity and differences as they are sources for
creativity and innovation. Ideas that are connected to new technologies in a constant
Smart City: A Qualitative Reflection … 563

and growing act within the territory, improving the quality of life of its inhabitants
and working to decrease budget constraints [21].
Digital aspects: Once the sub-centers have been identified, it is also intended to
complement the organization of sites by promoting interconnection between existing
systems (commerce, security, public administration, infrastructure use, health, work,
education, environment, transportation, etc.) by making intensive use of digital tech-
nologies (applications, sensors, citizen services, etc.), thereby promoting efficiency
in services provided, process improvement, agility, quality, within the territory,
services, and citizen [11]. They promote to their inhabitants, through interconnec-
tivity, a community digital space with the construction of networks of knowledge,
and providing spaces where the exchange of information and experiences occurs.
These are the ones that enable the participatory innovation process conducted by
citizens, employees, and end-users [22]. A resume of all aspects of a Smart City
model is described in Table 1.

Table 1 Characteristics of a Smart City model

Aspects Smart City characteristics
Infrastructure • Compact (can be by neighborhood and not by city), dense and
multi-centralized)
• Growth and internal improvement, not its expansion
• Rational and work to make better use of available resources
• Work better on issues related to waste
• Reduced pollution rates due to better management of urban mobility
Social • Community directly involved and active with city life
• Intimately committed to the development of the knowledge and knowledge of
its inhabitants
• Valuing the culture of knowledge and intelligence, to strengthen the
competitive advantages of the territory
• Creativity is the DNA of the city
• They can identify their identity, develop them, and stimulate their growth
toward differentiation and competitiveness
• Incentives to creation, providing access to new knowledge and diversity
• Ideas that are connected to new technologies in a constant and growing act
within the territory, improving the quality of life of its inhabitants
• Communities in which it exercises the articulation, cooperation, tolerance, and
multiculturalism around the objectives and common responsibility in the city
Digital • Massive Internet access
• Open data and transparency as a strategy for citizen engagement
• The citizen involvement of all levels of social class
• Collaborative digital platforms, increasing the likelihood of innovation
• The citizens are oriented to provide ideas, through a participatory plan of
innovation, independent learning, feedback
• Information is provided in real-time to the citizen, giving him greater
participation in decision making, and therefore more conscious, participatory
and collaborative
564 K. C. P. Jordão et al.

5 The Natural Inhibitor of Corruption and Illicit Markets

Potentialized Through the Smart City Concept

The intangibles aspects have become the major economic asset of countries, whether
they are developed, emerging, or underdevelopment. The intangibles aspects are
everything that will create knowledge of its inhabitants, in which it values the culture
of intelligence, to strengthen the competitive advantages of the territory, stimulates
growth toward differentiation and competitiveness, encourages creation by providing
access to the new, diversity. It will result in research, innovation, developing goods
and services, which allow adapted throughout several sectors of an economy and
society, being appointed as one factor to avoid the prosperity of corruption orga-
nization and illicit markets through the decrease possibility of manipulation and
fraudulent operations and the decrease customer interest channel of illicit products.
How much more a nation is intellectually developed, more your citizens will look for
practices structured that reflect order, discipline, organization, and prosperity [23].
The nations that historically invested in intangible aspects are naturally nations
that have overcome different adversities and are located as the most developed coun-
tries globally. The intellectual knowledge is one of three pillars of a Smart City model,
where is directly related to the social and cultural aspects of citizens. According to
the Smart City model, the intellectual knowledge is responsible for promoting the
city creativity, developing characteristics and stimulate its potential toward differ-
entiation and competitiveness even in the face of globalized competition, and mini-
mizing the social discrepancies, being, therefore, an important and natural inhibitor
for corruption process and illicit markets [21].
Another aspect observed with the use of a Smart City model is the rescue of the
city to your citizens. A common characteristic observed—independent of localiza-
tion, economic, or development level—is the citizens are more and more mistreating
their cities, aggravated by an exponential increase of cars, people, pollution, noises,
insecurity, unhealthy, unsustainable, among others; that result in a city without a
quality life and citizens that do not care and want to be there. Barcelona City during
the process of transformation that led it to become one of the most respected intel-
ligent cities in the world, understood that it needed to rescue the history of the
neighborhoods where its citizens lived. This was done by sectorizing the city into
12 subcenters. For each subgroup was identified its advantages and disadvantages of
infrastructures, it was mapped the characteristics that rescued the identity of each of
the sub-centralities mapped. Its culture, the DNA interconnects the citizens and the
region. It was mapped the regions that had urban voids, taking advantage of these
regions to be populated, using the available infrastructure. Each subcenter should be
self-sustaining in employment, housing, schools, hospitals, culture, leisure, super-
markets, to reduce the daily displacement between the regions of the city. These
actions rescued the interest of the citizens by the place where they live, in the city.
Citizens started to have idle time, which of course was occupied by the diversity of
cultural and social options that the subcenter provided. The consumer practices were
Smart City: A Qualitative Reflection … 565

reduced, and the interest in the diversity of offers of leisure that the city offered was
increased [8].
Finally, natural inhibition is related to digital aspects, directly related to the diver-
sity and dynamism of technology used by a Smart City model. The interconnec-
tion between urban systems promote the efficiency of the urban services (infor-
mation about goods and services), improving the businesses efficiency (helping
locate sellers), when enlarge the market scope (facilitating convenient delivery),
lower operate barriers and cost (allows goods to be purchased easily), intensifying
competitions and reducing cost (facilitating price comparisons), process improve-
ment, promoting a agility and quality of services and reducing traditional commerce
channels [9]. In order to support all kinds of technology adopted by a city, it is manda-
tory which has workflows structured and process well done defined. Without it, any
technology (for more efficient and promising it can be) will fail by the technology
itself. Additionality of this, workflows structured, and process well done defined
naturally has the power to inhibit the corruption network formulation.
Another positive aspect promoted by the Smart City model is related to the diver-
sity of data collected and analyzed never did before. The blockchain technology is
an example of it, being studied their use to mitigate government corruption process,
supply chain management (combating the pirated), smart contracting and financial
transactions, through the decentralized validation of the numerous transactions in
process flow (blocks) [24]. With all the city being monitored by digital technologies
(applications, sensors, citizen services, among others), the number of data collected
and mainly the information provided by this data certainly will show new ways
to solve old issues. According to ACDE, information is still an essential key for
designing and implement effective policies and measures to combat illicit markets
[25].
However, data will not fix the problems existing in a city by itself. It is necessary
to have new models of administration, that supported by information never observed
before and can signal for simple and efficient solutions.

6 An Ethic Governance Proposal

Today’s administrations bring problem-solving models that need to be reviewed

against the new possibilities offered by future Smart Cities. According to Wind et al.
[26]: “When the world changes significantly, there is a risk of a model completely
irrelevant to the new situation.” In this way, the context offered by new management
tools based on information and communication technologies (TIC) that incorporate
ad hoc networks, powerful databases and efficient data monitoring results in the
possibility of innovating in the investment of new forms of energy generation, in
the dynamics of education, health, safety, etc. All of this leads to a change in the
mental models that structure managerial decisions. However, it is worth noting the
observation [27] of the risk that exists to a world that, depending so much on collecting
566 K. C. P. Jordão et al.

large amounts of information and desire for quick decisions, loses the context of its
sources.
Therefore, it is important to do not rely exclusively on the power of the new tools
offered by TICs, as regards the new models that need to be created. Just as the current
bureaucracy was not enough to overcome the problems seen at the beginning of this
work, the simple fact of giving more efficiency to technological resources ends up
reproducing “controls” similar to those that generated the problems described. In
this context, Smart Cities should not only emphasize technological solutions in the
information field and its processing, which would lead to a technocracy. The risk is to
give life to all kinds of data, which postman alert and classify [27] as garbage infor-
mation, that no more responding to human concerns and it being little useful to give a
coherent direction to the solution of worldly problems. In this context is proposed—
as a methodological way to solve the problems pointed out, which can’t increase
in these new environments of the Smart Cities—a new concept of Governance, as
presented by Melo, apud Santos [28]: “The ‘modus operandi’ of government poli-
cies which includes, among others, issues related to political-institutional format of
decision processes, an appropriate definition of public/private mix policies with a
participation decentralized, and new mechanisms for financing policies programs
with a global reach”. However, this “modus operandi” needs to discuss the ethical
dimension from a different perspective, a little bit singular. This ethical, in a prag-
matic and realistic way, must rescue the right and fair conduct to be applied to any
organization or society, driving for what is the correct to do in the cities in order to
transform them in intelligent and sustainable, but fundamentally human places.
The ethics inclusion, as an effective dimension in a Smart City governance, would
elevate the city into the condition of true human abode. The Smart Cities could be the
carry of principles and values that would guide citizens to establish a new civilization
paradigm, which Boff refers to the future with the look of “caution.” And who would
be responsible for “care” in a Smart Cities model? It is necessary to overcome the
actual model where public administration has the responsibility to monopolizing the
directives to regulate the citizen’s life [1].
The Brazilian Cities Statute Guideline, for example [29], must be followed by
public administration to elaborate on the urban city policy. In this document, it
is stressed that all actions must be focused on ensuring a fair city to all citizens,
where the poor and rich can enjoy the benefits of urbanization. In order to follow
this aspiration, Rezende explains [22] that cities depend on their public managers to
municipal planning and city management, but also is not limited to them. Any kind of
strategic, multiannual, fundamentally or master plan of the city must be attended by a
committee of citizens, private and public administration. However how to ensure that
all involved (citizens, public and private administration) commit, feel responsible and
accountable, in the greater objective of achieving effective quality in the governance
of CISH, and the necessary “care” with the population, which encompasses the many
situations of vulnerability and risk that exist? It is in this space of accomplishment
that the numerous problems described at the beginning of this work find space for
existence, and only more efficiency in data collection and mining would not be
Smart City: A Qualitative Reflection … 567

enough, it becomes essential to be explicit as the data can be used by those who
possess it, its limits and possibilities.
Thus, in practice, Hans Jonas’s warning [30] when referring to ethics should be
rescued here, and that: “Ethics must exist because of men act, and ethics exist to order
their actions and regulate their power to act. Its existence is all the more necessary;
therefore, the greater the powers of action which it has to regulate.” In this context, it
is then necessary to clarify the field of action in which ethics will become one of the
forces that will concretize a Smart City that will guide its decisions through ethics.
Firstly, it is necessary to define the ethical guidelines that will guide the modus
operandi according to the values and principles that point to the good of all. Within this
framework, you can consider t [31] the approach that points out some general ethical
principles that result in ethical attitudes committed to society and thus consistent with
the goals desired here. They would be personalized principle, in which the valuation
of the human person is practiced; principle of guidance for the common good, as
a stimulus for each one to go beyond private interests, and to search for what is
good for society as a whole; principle of order of responsibility, in order to prioritize
actions that, depending on the responsibility of each one, be guided by what is most
achievable and within its reach; principle of managers’ prudence, where decision-
making does not distance itself from a safe analysis in which ethical, conviction
(deontology, treaty of duties), and responsibility (teleology, study of human ends)
are elements of reflection given the complexity of the real.
Once the guiding principles have been defined, it is necessary to create conditions
so that they can manifest themselves in the decision process, in the face of practical
problems, when people are confronted with the need to guide their behavior through
norms that are more appropriate or worthy of being fulfilled. From these assumptions,
it is important to continuously develop ethical awareness through educational actions.
In what concerns the governance of a Smart City, it is suggested at this point that
the guidance should consider the categorical imperative of Kant, which according to
Vázquez [32] is the supreme expression of the commandment of reason and is thus
prescribed: “Age so that you may want the motive that led you to act to be a universal
law.”
Finally, acting in a way that establishes a management model guided by ethics
include practical actions such as the creation of communication centers, citizen-
ship, and ethics, whose dynamics can establish a dialogue between managers, busi-
nessman, and citizens, creating ways to encourage citizen participation. And, in
particular, this citizen participation that is not limited to the Civil Councils linked to
political positions, and, with this, it prevents that it happens to be effectively away
from the population in a general way [1].
The presence of ombudsmen capable of establishing fair paths to blind spots in
administrative processes as a whole; external ethical audits, which can go beyond
the techno-administrative visions, but can enlarge the look to what differs from the
established principles; foster a culture of ethical and transparent management. A
transparent government where all citizens can have access not only to legislation in
each area of government, but also to their flows and processes, freeing this citizen
568 K. C. P. Jordão et al.

from being held hostage by verbal information from servants or from written and
slow protocols and inefficient [1–6].

7 Conclusions

Regardless of all challenges involved to combat corruption schemes and illicit

markets around the world, it is notorious the need to reinforcing designing policies
to address those issues. Even in the face of this challenging scenario, the Smart City
is opening one more channel to mitigate the corruption and illicit markets through
citizen involvement, considering to be the key to drive this chain.
Those positives opportunities that can be exercised, like the economy is stimulated
by the drop unemployment rate (when each subgroup exercise to be self-sufficient
with all kind or service and product to supply your subgroup), by the use of data (big
data) and new technologies to mitigate the corruption scheme, by the government
transparence and the population involvement to decide and supervision the main
investments and decision to the city; by decrease, the number of accidents in transit
when reduce the number of trips daily, reduce the congestions and increasing the
accessibility and use of alternative modes, and consequently, the air quality is signif-
icantly improved; by the new city mobility will win bike and pedestrian adepts, that
naturality improve the city space; by the decreasing number of accidents, it means that
reduce the cost with hospital care and disability insurance, reducing the economy’s
expenses with these aspects; by the society is more interactive with the governments,
and the governments with the society (bottom up/ top down), promoting an exchange
of feedbacks about the efficiency of the services provided, process improvement,
agility, quality, within the territory.
Finally, the citizens understanding that it is part of the city, as responsible to keep
the system operating and the order established and will work to keep this perception
in each subgroup that they are inserted.

Acknowledgements The authors would like to thank the Coordination of Superior Level Staff
Improvement (CAPES), and National Council for Scientific and Technological Development
(CNPq).

References

1. Lopes, N. V. M.: Smart Governance for Cities: Perspectives and Experiences. Springer Nature
(2020)
2. Mehmood, R., Katib, S. S. I., Chlamtac, I.: Smart Infrastructure and Applications. Springer
International Publishing (2020)
3. Chizzotti, A., Chizzotti, A., Trevisan, A., Verillo, J., Ianhez, J.: The fight against corruption in
prefectures in Brazil. Ateliê Editorial, Sao Paulo (2003)
Smart City: A Qualitative Reflection … 569

4. Shan, M., Le, Y., Chan, A. P., Hu, Y.: Principal causes of corruption in the public construction
sector. In: Corruption in the Public Construction Sector, pp. 49–77. Springer, Singapore (2020)
5. Shan, M., Le, Y., Chan, A. P., Hu, Y.: Corruption in construction: a global review. In: Corruption
in the Public Construction Sector, pp. 9–22. Springer, Singapore (2020)
6. Ryder, N., Pasculli, L.: Corruption, Integrity and the Law: Global Regulatory Challenges (2020)
7. Atkinson, C. L.: Theme-Based Book Review: Public-Sector Corruption (2019)
8. FIESP.: Transnational illicit markets in São Paulo: the transnational criminal economy/FIESP.
FIESP, São Paulo (2017)
9. Hanna, R.: Technology beats corruption. Science 355(6322), 244–245 (2017)
10. Jordão, K. C. P.: Intelligent Cities: A Proposal for the Transformation of Brazilian Cities (2016)
11. Panhan, A. M., Mendes, L. D. S., Breda, G. D.: Building Smart Cities, vol. 1, 1st edn, p. 149.
Appris, Curitiba (2016)
12. Song, H., Srinivasan, R., Sookoor, T., Jeschke, S.: Smart Cities: Foundations, Principles, and
Applications. Wiley (2017)
13. Padilha, R., Iano, Y., Monteiro, A. C. B., Arthur, R., Estrela, V. V.: Betterment proposal to
multipath fading channels potential to MIMO systems. In: Brazilian Technology Symposium,
pp. 115–130. Springer, Cham (2018)
14. Monteiro, A. C. B., Iano, Y., França, R. P., Arthur, R., Estrela, V. V.: A comparative study
between methodologies based on the hough transform and watershed transform on the blood
cell count. In: Brazilian Technology Symposium, pp. 65–78. Springer, Cham (2018)
15. Monteiro, A. C. B., Iano, Y., França, R. P., Arthur, R.: Methodology of high accuracy, sensitivity
and specificity in the counts of erythrocytes and leukocytes in blood smear images. In: Brazilian
Technology Symposium, pp. 79–90. Springer, Cham (2018)
16. França, R. P., Iano, Y., Monteiro, A. C. B., Arthur, R.: Potential proposal to improve data
transmission in healthcare systems. In: Deep Learning Techniques for Biomedical and Health
Informatics, pp. 267–283. Academic Press (2020)
17. Kneib, É.: Project and City: Centralities and Urban Mobility, vol. 2, p. 324 (2014)
18. Crainic, T. G., Gendreau, M., Potvin, J. Y.: Intelligent freight-transportation systems: assess-
ment and the contribution of operations research. Transport. Res. Montréal Part C, 541–557
(2008)
19. Komninos, N.: The Age of Intelligent Cities: Smart Environments and Innovations-for-all
Strategies, vol. 1, 1st edn, p. 298. Routledge, London (2014)
20. Fachinelli, A. C., D’arisbo, A. E. D. M.: The importance of social innovation and the creative
economy as an indicator for sustainable development. Int. J. Knowl. Eng. 3(5), 276–293 (2014)
21. Landry, C.: The Creative City, A Toolkit for Urban Innovators, vol. 1, 2nd edn, p. 352. Comedia,
London (2012)
22. Rezende, D. A.: Planning of Municipal Information and Strategies for the Digital City: Guide
for Projects in Prefectures and Public Organizations. Atlas, São Paulo (2012)
23. Fernandes, R., Gama, R. A.: Territorial Creativity in Portugal: From the Indicators to the
Creative Territories, pp. 1–12. FLUC Geografia, Coimbra (2012)
24. Puiu, T.: How Bitcoin’s Blockchain Could Mark the End to Corruption. ZME Science (2015)
25. OECD Workshop.: Assessing the Harms Posed by Illicit Trade and the Illegal Economy
for Global Supply Chain Integrity, Economic Growth and Market Security (March,
2018). https://www.oecd.org/governance/risk/Illicit%20Trade%20Workshop%2026%20Octo
ber%202012-rev.pdf
26. Wind, Y. J., Crook, C., Gunther, R.: The power of Mental Models. Bookmark (2009)
27. Postman, N.: Technology: The Surrender of Culture to Technology (1994)
28. Santos, M. H. C.: Governance, and Democracy: Creation of Governance Capacity and
Executive-Legislative Relations in Post-Constitutional Brazil (April, 2018). http://egov.ufsc.
br/portal/conteudo/governabilidade-governan%C3%A7a-e-democracia-cria%C3%A7%C3%
A3o-de-capacidade-governativa-e-rela%C3%A7%C3%B5es-execut
29. Federal Senate: Statute of the City. (April, 2018). https://www2.senado.leg.br/bdsf/bitstream/
handle/id/70317/000070317.pdf?sequence=6
570 K. C. P. Jordão et al.

30. Jonas, H.: The Principle of Responsibility: a Test of an Ethics for Technological Civilization
(2006)
31. Pineda, E. S., Cárdenas, J. A.: Ética en las Organizaciones. McGraw-Hill (2007)
32. Vázquez, A. S.: Ethic (2006)
A Discussion of the Challenges Small
Towns Face in Reaching the Promising
Scenario of Electronic Government
Intelligent Cities

Valéria Sueli Reis , Yuzo Iano , Gabriel Gomes de Oliveira ,

Telmo C. Lustosa , Mauro A. S. Miranda , Odair dos Santos Mesquita ,
Ana Carolina Borges Monteiro , and Reinaldo Padilha França

Abstract This work aims to total the reality of cities through some Brazilian elec-
tronic data e-Gov, and other research conducted and counteracted with the latest
assumptions proposed by smart cities governance models (CI) and intelligent gover-
nance (IG) increasingly imposed by the information society and network, realizing
the new social engineering of collaboratively and ICT. Data were collected through
reports of their research and also the theoretical literature review concerning the issue.
This paper presents a discussion in order to clarify how the small municipalities still
need to mature to achieve a reality of smart city governance.

V. S. Reis · Y. Iano · G. Gomes de Oliveira (B) · T. C. Lustosa · M. A. S. Miranda ·

O. dos Santos Mesquita · A. C. Borges Monteiro · R. Padilha França
School of Electrical and Computer Engineering (FEEC), University of Campinas—UNICAMP,
Av. Albert Einstein-400, Barão Geraldo, Campinas, SP, Brazil
e-mail: [email protected]
V. S. Reis
e-mail: [email protected]
Y. Iano
e-mail: [email protected]
T. C. Lustosa
e-mail: [email protected]
M. A. S. Miranda
e-mail: [email protected]
O. dos Santos Mesquita
e-mail: [email protected]
A. C. Borges Monteiro
e-mail: [email protected]
R. Padilha França
e-mail: [email protected]

© The Editor(s) (if applicable) and The Author(s), under exclusive license 571
to Springer Nature Switzerland AG 2021
Y. Iano et al. (eds.), Proceedings of the 5th Brazilian Technology Symposium,
Smart Innovation, Systems and Technologies 202,
https://doi.org/10.1007/978-3-030-57566-3_56
572 V. S. Reis et al.

Keywords Innovation in public service · Service design · E-governance · E-gov ·

Smart city

1 Introduction

Forms of public–private cooperation and its strengthening are perhaps one of the
best advantages arising and improving the success and knowledge about smart city
models (CI). Therefore, there is no governance without innovation, and innovation
emphasizes collaborative elements, partnership, integration, new arrangements for
solutions to the old problems of municipal governance.
The evaluation of public policies is prevalent factor for evolution to occur in the
efficiency of government, and as a result of advances in e-government, we used in
this article some references in the latest research carried out as Tic e-Gov operated by
CGI.br, ENAP service search data public, and IEGM of São Paulo Court of Auditors,
on the municipal, state and federal government initiatives, and we will explore some
aspects revealed, as well as other researchers at the national level dealing with the
municipal electronic governance.
In the early 2000s, the governance of Brazilian cities was still under strong influ-
ence of the precepts of NPM (theory of New Public Management) that brought to
public administration its load control and efficiency, inspired by the private manage-
ment, where the electronic government had their space and growth in the moderniza-
tion of information systems and general computerization of departments and munic-
ipal departments. Innovation was not feared or reflection in this period was disen-
franchised almost or totally, the values that focused in this phase are the quality as
far as possible, and often questionable, the economic viability that much focus still
on the bidding process, and immediatism efficiency.
But cities still in need of solutions to their problems of governance, we can score
the areas of finance, human and regulatory resources, secondly, but no less prob-
lematic, the negative impact on key areas of provision of services to the population:
health, basic education, infrastructure, social welfare, urban mobility.
In the way of governance/e-government to think of smart city, opened the scope of
possibilities that are still strongly supported by the use of Information and Commu-
nication Technology (ICT), but with intelligence element that requires smart govern-
ment, smart servers, active and socially intelligent citizens, and intelligent technolog-
ical solutions, we will discuss intelligent governance (GI) who resort, for example,
autonomous systems, sensing and Internet of Things (IoT), knowledge manage-
ment and innovation, intelligent platforms, big data, analytics, service design, health
services with quality and low cost, telecommunication systems with low memory
consumption. All this must be done for that the objectives and interests should be
aligned between the people and resources that make up the city [1–6]. This definition
of [7], among several, and several authors who defined smart cities, contemplates
broadly: Smart cities are those that well realize the vision of the future in various
aspects—economy, people, governance, mobility, environment, and quality of life
A Discussion of the Challenges Small Towns … 573

and are built on the intelligent combination of decisive action, independent, and
aware of the actors who work in them and generate public value.
In the municipality of context, we would enter in your smartest stage when the
management follows an integrated, real-time manner, exposure data, and relevant
information and value-added services to citizens and businesses. It is based on shared
intelligence by all actors of the city and the city’s platform a facilitator of collaborative
solutions [8].

2 Methodology

This survey carried out a bibliographic review focusing on data that were collected
through reports of academic research, governmental data, and also the theoretical
literature review concerning the issue.

2.1 Policy Brief Analysis of the Municipalities

to the Challenges for CI and GI

We go forward with the legislation in some aspects such as LAI1 [9] and Law
139/20092 [10, 11]. That enabled access from several municipalities toward trans-
parency and look at its governance, providing information, but this legal/political
framework must evolve more, and we justify with data later.
The base decisions should be distributed, shared, transparent, and democratic,
within the representative democracy we live, expanding participatory democracy with
the presence of citizen, government relationship expansion with the local community,
with business, with institutions that contribute municipal management.
More consultations with society, more ears. It is not for the smart city closed
management, individual decisions, solutions that do not add public value or without
social interest or generate waste, by a requirement of the information society itself,
although it is obviously the exception to the willingness of data, open data, and infor-
mation security, which is a big challenge, among others, to municipal IT governance,
and obligation of social welfare and the city’s quality of life that must be shared and
liable for all and the intelligent management cannot do without this support, even as
local governments increasingly suffer from the scarcity of resources, and increase in
expenditure.
Brazil’s cities suffer from the limitation of income, many survive only with two
recipes, namely the property tax and the annual transfer of National Fund of Munic-
ipalities (FNM), being held hostage for most of the period of funding parliamentar-
ians and transfers from the federal government having difficulties to have qualified

1 Law No. 12,527 of November 18, 2011, known as the Law of Access to Public Information, (LAI).
2 Complementary Law No. 131 commonly called Transparency Law (2000), as amended by Law.
574 V. S. Reis et al.

human resources, qualifying servers and a tract of continuous training, and that
directly suffers the impact of change management (managers) which compromises
the continuity of the plans and projects.
Having a recurring difficulty of working with deadlines and schedules of projects
and deliveries, showing the absence of a culture of deadlines. Smart cities plans
represent a set of actions that stir the political and economic structures established
in the municipality and strongly questioned its internal organization.
According to [8], the need to revisit the federal pact in Brazil is illustrated, in terms
that balance the investment needs and funding and access to financial resources. At
the same time propose new forms of organization for more horizontal municipal
administrations, more integrated, more open to the participation of society, and new
work processes. In this scenario, the ICTs should play an important role in making
these changes [2, 8].

2.2 Smart City and Implies a New Insight

into the Governance of the City

So, it is necessary to think about the theme of governance of cities with integrator
based, sustainable and efficient, and also in tune with the needs of communities and
citizens, that meets the improvement of the value of services and quality of urban
life, where it is necessary to present a brief historical tour, especially with regard to
technology applied in municipal management, being relevant to deal with the impact
of e-government on small municipalities.
With the democratization process in Brazil from the Citizen Constitution of 1988
[12], the Brazilian public administration is moving forward with parallel legislation
to the maturing of the democratic principles of management and incorporating the
principles of private management (NPM) in Brazil clearly demonstrated from the
90s [12–16]. More detail about the evolution of public administration models, the
phase that is characterized today in management is the bureaucratic model Webe-
rian that is characterized by strong bureaucracy and government inflexibility; it is
a regimented government marked by formality, impartiality, and professionalism;
this phase it is important to stress that we inherited over control and attachment
information. The next phase recorded in theory comes the era of managerialism.3
Because there, in 2000, already we have the LRF Fiscal Responsibility Law [16].4
However, there is criticism of this influence, it does not include a vision of partici-
patory citizen able to citizenship, but as a user/client, and in this respect there is the
theory of NPS—New Public Service that favors all public commitment to the fact that

3 Relatedto transparency, accountability of rulers, as well as the accountability of rulers for their
actions.
4 LRF Fiscal Responsibility Law (LRF)—(Complementary Law No. 101 of 04 May 2000.
A Discussion of the Challenges Small Towns … 575

investment e-government, where the government has to be open to citizens to exer-

cise participation. In these last two phases, the concepts of accountability, account-
ability (accountability), transparency, citizen engagement, democratic governance,
and network governance or digital, now have strength, the 2000s to the present.
As well as some federal laws that encouraged good practice in municipalities
(LAI, LRF, Law transparency, Cities Statute, this one suggests various instruments
of governance, underused and unknown to many managers/cities), TCU’s initiatives
and other ministries, and finally on the efforts and growing more isolated initiatives,
the need a national plan for smart cities, would be important.
Despite some initiatives by the federal government, such as Dec. 9319, the
Brazilian Strategy for Digital Transformation, E-Digital, should also reflect the
medium term in the cities in order to make investments and technological innovation
and development in cities.
Moving forward in our analysis, in the first phase of e-government in the munici-
palities, there is a focus on improving internal processes and subsequent investments
for relations with citizens, according to ENAP [17] are legal frameworks reflecting
in governance actions e-government and municipalities.

2.3 The Facts to the Data: Some Aspects of Electronic Data

Governance in Small Towns

To the specifics of municipalities with up to 100,000 inhabitants, covering 97% of the

cities contingent in Brazil, we noted the lack of research studies and more compliant
models to this city’s profile, so relevant if it does, this article explicitly within the
logic evaluation of public policies, recent survey data, such as Tic gov 2017 that may
show aspects that demonstrate the use of ICT and to reflect in greater transparency,
citizen participation, open data, sustainability, and security.
The first information is that only 23% of cities between cities up to 100 thousand
inhabitants have smart cities plans. According to the results of research, Tic gov
only 18% of all Brazilian municipalities had a project or smart city project. The
municipality must tow your regular planning mechanism as a master plan, multi-
year plan (PPA), and make adjustments in the legislation with a view to regulation
that allows GI CI and actions.
The second base and as a gateway that is 92% of these municipalities have a
website; Only 52% of these municipalities have their own IT industry, demonstrating
the high degree of outsourcing of IT services, and only one in four municipalities
with up to 10,000 inhabitants also has a department focused on ICT [18, 19].
About infrastructure data networks, there has been increased use of cable and fiber
optic connections to 87% of these connections in the range of municipalities with up
to 100,000 inhabitants and 77% those with up to 10,000 inhabitants. Regarding what
is available information and services on the portals and municipal sites, according
to Table 1 (derived from the websites Tic and Gov 2017).
576 V. S. Reis et al.

Table 1 Services on the websites of municipalities with up to 100,000 inhabitants

City halls, by type of service available on the More than 10 thousand to 1000 thousand
website, location, and size (2017) inhabitants (%)
Download documents or forms 84
Fill or send forms on the website 58
Electronic invoicing issue (e-Invoicing) 61
Consult ongoing administrative or judicial 47
processes
Issuing documents such as licenses, certificates, 44
permissions, and others
Register or enroll, for example, in competitions, 33
courses in schools
Schedule appointments, consultations, services, 21
among others
Make payments such as fees and taxes 25

There was growth in the provision of some services in the municipal sites, however,
they are not those configured as transactional, payment services, the opening of new
enterprises, fines payment, or requisition of personal documents. Other transactional
services such as possibilities for payments of fees, taxes, and registration or enroll-
ment in competitions, courses, and schools were less frequent on the websites of
municipalities, except the capitals with the highest proportion of the supply of these
resources.
The more common in municipal websites, for example, is the possibility of down-
loading documents or forms. Of the nine services indicated in the survey only three
were offered for more than half of the municipalities that have a site, they are:
download documents or forms with 84%; fill and send a form with 58%, and issue
electronic invoices with 61% incidence. In this scenario, there is the difficulty of
municipalities to provide online services and potentiating the digitization of public
services.
For this information, have posted positive indicators, most public accounts or
rendering of accounts with 98%, purchasing, procurement, or electronic procurement
to 93%. Some content, however, as documents with plans and goals and results, and
even public service catalog are less publicized, which can be configured as an obstacle
for society to access information and track the progress of public policies and actions
of the public sector organizations.
The catalog services, quoted at least 59% should improve this ratio with that
service user code provides that demand more imperative charter services in its Art.
7, starting next year. This instrument is important for citizens to know what to look
for and have available the services provided by the city.
About half of the municipalities (49%) claimed to have done electronic auction and
this most common practice in the capital (88%) and in cities with over 500,000 inhabi-
tants (90%). Therefore, the potential benefits of electronic trading, as the expansion of
A Discussion of the Challenges Small Towns … 577

competition and increased transparency and social control, are not being incorporated
into part of Brazilian municipalities.
Another important feature, 47% of the municipalities in this size have an online
ombudsman. The ombudsman is one of empowering instruments of citizen partici-
pation, described in the main initiatives of the Federal Government as the Access to
Information Act, 2011, and most exalted in the recent public service user code.
The ombudsman is an instrument important for citizens can demand their rights
with regard to public services, and the expansion of institutional channels for partic-
ipation via the Internet is another important challenge of smart governance. One of
the dimensions associated with digital government policies is a government to adopt
measures that consider the demands of society in times of definition and evaluation
of public policies [20].
Addressing solely on the municipalities of SP, it exposed the initiative of the court,
in the state of São Paulo is a ruler marking out the municipalities in its governance
[21, 22] and measures the knowledge and use of Information Technology resources
in favor of society.
This index collects information about computer usage policy, information security,
staff training, and transparency. The average rating of municipalities was in 2017 with
the average +C on a scale of A, B + B, C+, and C.
Classification:
A Highly effective
B Effective
B+ Very effective
C+ In adaptation phase
C Low level of adjustment.
Where IEG-C + M is greater than or equal to 50% and less than 60% of the
maximum score. St. Paul of the municipalities is still undergoing adjustments to
its electronic services and e-governance [23–29]. The Municipal Index of Tech-
nology Governance i-Gov Information IT/TCESP makes up the Effectiveness Index
of Municipal Management—IEGM IEGM/TCESP [30, 31], as shown in Fig. 1.
The Munich research–Profile of municipalities conducted by IBGE and published
with 2017 results [31–36], we highlight results that address sustainability and secu-
rity. These illustrate how natural disasters have affected the Brazilian cities such as
over 48% of the municipalities have been affected by drought, (2706 municipalities).
1726 were affected by flooding; 1515 by floods; 1093 by accelerated erosion and 833
by landslides. And yet, 59% of the municipalities do not have any kind of natural
disaster prevention tool and only 14.7% of Brazilian municipalities, 821 municipali-
ties have some contingency plan or protection. It makes us reflect on the potential of
emerging technologies and advances in tools for prevention and assistance to these
problems. Obviously, there is an important context to be faced with a smart and
challenging vision, which is not only social inequality, but also regional disparities,
political and economic that befall our country.
578 V. S. Reis et al.

Fig. 1 Municipal index of governance of information technology I-Gov IT/TCESP [27]

3 Results and Discussion

Despite the different levels of appropriation of intelligent governance between munic-

ipalities, these are in the intensive process of computerization/digitalization and all
feature some website. Most acts of transparency are performed so that it meets the
legal instruments (laws), and there are few services geared to citizens that they can
expand their arch action or intervention in public administration, with the majority of
services refer to the collection of taxes (Building urban land tax, electronic invoice,
tax on services).
It also points out the importance of governance models that includes the subjective
nature of the project actions as leadership and culture change within the municipal
government. With the same focus on service, design can be applied to the empow-
erment of this governance and alignment of goals together stakeholders, citizens,
and generate public value. The initial plan, the evaluation, and correction of routes,
deployment of smart governance is not an end in itself, be a major challenge, the
discontinuities of the municipal administrations.
However, e-government is an eminent contemporary need to municipalities with
their local conditions, economic and social needs in the face of opportunities and
advancement in technology. The theoretical references, models and empirical expe-
riences and new Tic’s, denote that there are great possibilities for the deployment of
smart e-government systems or maturation of existing initiatives in municipalities,
citizens, and generate public value.
The initial plan, the evaluation, and correction of routes, deployment of smart
governance is not an end in itself, be a major challenge, the discontinuities of the
municipal administrations. However, e-government is an eminent contemporary need
to municipalities with their local conditions, economic and social needs in the face
of opportunities and advancement in technology.
A Discussion of the Challenges Small Towns … 579

The theoretical references, models and empirical experiences and new Tic’s,
denote that there are great possibilities for the deployment of smart e-government
systems or maturation of existing initiatives in municipalities, citizens, and generate
public value. The initial plan, the evaluation, and correction of routes, deployment
of smart governance is not an end in itself, be a major challenge, the discontinuities
of the municipal administrations.
However, e-government is an eminent contemporary need to municipalities with
their local conditions, economic and social needs in the face of opportunities and
advancement in technology. The theoretical references, models and empirical expe-
riences and new Tic’s, denote that there are great possibilities for the deployment of
smart e-government systems or maturation of existing initiatives in municipalities.
With discontinuities of the municipal administrations. However, e-government
is an eminent contemporary need to municipalities with their local conditions,
economic and social needs in the face of opportunities and advancement in tech-
nology. The theoretical references, models and empirical experiences and new Tic’s,
denote that there are great possibilities for the deployment of smart e-government
systems or maturation of existing initiatives in municipalities—the discontinuities
of the municipal administrations.
However, e-government is an eminent contemporary need to municipalities with
their local conditions, economic and social needs in the face of opportunities and
advancement in technology. The theoretical references, models and empirical expe-
riences and new Tic’s, denote that there are great possibilities for the deployment of
smart e-government systems or maturation of existing initiatives in municipalities.

4 Conclusions

Smart cities are urban spaces capable of promoting economic development and
increasing people’s quality of life. Given this, it is possible to see that smart munic-
ipalities have a lot to offer to society, since the use of technologies capable of
making cities more intelligent directly influence the way society lives and behaves
in urban environments, considering the applications and other government digital
platforms that make urban management more efficient and still carrying out planned
development projects, with the aim of making cities increasingly humanized.
It is important to highlight that the concept of smart cities cannot be an end in
itself; it must allow society to develop in a more educated way, rescuing local culture
and generating jobs and growth opportunities in the most varied sectors of life.
Smart cities are based on the union of public, private, and civil society efforts to
create a planned infrastructure, providing for the use of technology and the effective
participation of the community in decision making, based on the union of efforts of
the public, private, and civil society sectors. Smart cities use digital technologies to
optimize resources, reduce costs, connect people and, mainly, improve the quality of
public and private services, always focusing on users.
580 V. S. Reis et al.

Thus, it is possible to transform large cities into smart urban centers. Just prior-
itize people and the quality of living in the spaces. Since the need for investment
in technologies that allow the development of projects, with a focus on bringing
communities and the environment closer together, connecting people and opportu-
nities, strengthening the economy, and promoting business, are crucial factors that
overcome the challenges faced for cities that are not yet considered smart.

References

1. Caragliu, A., Del Bo, C., Nijkamp, P.: Smart cities in Europe. J. Urban Technol. 18(2), 65–82
(2011)
2. França, R. P., Iano, Y., Monteiro, A. C. B., Arthur, R.: Improvement of the transmission of
information for ICT techniques through CBEDE methodology. In: Utilizing Educational Data
Mining Techniques for Improved Learning: Emerging Research and Opportunities (pp. 13–34).
IGI Global (2020)
3. França, R. P., Iano, Y., Monteiro, A. C. B., Arthur, R.: Intelligent applications of WSN in the
World: a technological and literary background. In: Handbook of Wireless Sensor Networks:
Issues and Challenges in Current Scenario’s (pp. 13–34). Springer (2020)
4. Padilha, R., Iano, Y., Monteiro, A. C. B., Arthur, R., Estrela, V. V.: Betterment proposal to
multipath fading channels potential to MIMO systems. In: Brazilian Technology Symposium
(pp. 115–130). Springer, Cham (2018)
5. Monteiro, A. C. B., Iano, Y., França, R. P., Arthur, R., Estrela, V. V.: A comparative study
between methodologies based on the hough transform and watershed transform on the blood
cell count. In: Brazilian Technology Symposium (pp. 65–78). Springer, Cham (2018)
6. Monteiro, A. C. B., Iano, Y., França, R. P., Arthur, R.: Methodology of high accuracy, sensitivity
and specificity in the counts of erythrocytes and leukocytes in blood smear images. In: Brazilian
Technology Symposium (pp. 79–90). Springer, Cham (2018)
7. Giffinger, R., Gudrun, H.: Smarter cities ranking: an effective instrument for the positioning
of cities? ACE: architecture. City Environ. 12, 7–25 (2010)
8. Cunha, M. A., Przeybilovicz, E., Macaya, J. F. M., Burgos, F.: Smart cities [electronic resource]:
digital transformation of city. Program Public Administration and Citizenship-PGPC, St. Paul,
p. 161. ISBN: 978-85-87426-29-1 (2016)
9. Brazil: Law on Access to Public Information. Law No. 12,527, of 18 of November 2011.
Presidency. Civil House. Subchefia for Legal Affairs. Brasília (2011)
10. Brazil: Decree No. 6932 of 11 August 2009. Brasilia: Official Gazette [of] the Federative
Republic of Brazil, 2009 Brasil. Law No. Law 11.079/04 of 30 December 2004. Brasilia:
Official Gazette of [the] Federal Republic of Brazil (2004)
11. Brazil: Transparency law. Complementary Law No. 131 of 27 May 2009. President of the
Republic. Civil House. Subchefia for Legal Affairs. Brasília (2009)
12. Brazil: the Presidency. Constitution of the Federative Republic of Brazil 1988 to October 5
(1988)
13. Denhardt, R. B.: Theories of Public Administration, 6th edn, p. 367. Cencage Learning, São
Paulo (2012). ISBN 978-85-221-1081-0
14. Perry, J.L.: Democracy, and the new public service. Am. Rev. Public Adm. 37(1), 3–16 (2007).
https://doi.org/10.1177/0275074006296091
15. Tsukumo, A. N., Pepper, M. F., Teracine, E.: Information technology providing new directions
for Public Administration. III Seminar, Democracy, Law and Public Management Theme III—
New Directions for Public Management. Brasilia (24–25 November, 2011)
16. Brazil: Federal Law No. 13 460 of 26 June 2017. Provides for participation, protect and defend
the rights of the user of public services of the public administration (2017). Available in: http://
www.planalto.gov.br/ccivil_03/_ato2015-2018/2017/lei/l13460.htm
A Discussion of the Challenges Small Towns … 581

17. Brazil: Fiscal Responsibility Law. Complementary Law No. 101 of 04 May 2000. President of
the Republic. Civil House. Subchefia for Legal Affairs. Brasília (2000)
18. ENAP: Research on utility services from the federal government. Brasilia: Enap, pp. 10–11,
p. 75. Yl. (Enap notebooks, 55) (2018). Accessed September 10, 2019. http://www.enap.gov.br/
19. Studies Center of the Technologies of Information and Communication (Cetic.br). ICT
Government Electronic (2017). Available at: https://cetic.br/tics/governo/2017/prefeituras/E2/
20. Organization for Economic Cooperation and Development—OECD. Review of the digital
government of Brazil: Towards the transformation type of the public sector (2018)
21. Bolivar, M. P. R., Meijer, A. J.: Smart governance: using a literature review and empirical
analysis to build a model research. Social Science Computer Review 0894439315611088, first
published on (29 October, 2015). http://doi.org/10.1177/0894439315611088
22. Castells, M.: The network society. The information age: economy, society, and culture, vol. 1.
Paz e Terra, São Paulo (1999)
23. Dunleavy, P., Margetts, H.: Design Principles for Digital Essentially Governance (2015)
24. Margetts, H., Dunleavy, P.: The second wave of digital-era governance: a quasi-government
paradigm is on the Web Phil. Trans. R. Soc. A 371(1987), 20120382 (2013)
25. Meijer, A., Bolívar, M.P.R.: Governing the smart city: a review of the literature on smart urban
governance. Int. Rev. Admin. Sci. 82(2), 392–408 (2016)
26. Meijer, A. J., Gil-Garcia, J. R., Bolivar, M. P. R.: Smart city research contextual conditions,
governance models, and public value assessment. Soc. Sci. Comput. Rev. 0894439315618890
(2015)
27. SavoldellI, et al.: Understanding the e-government paradox: learning from literature and
practice on barriers to adoption. Gov. Inf. Q. 31, 63–71 (2014)
28. Meijer. A., Bekkers. V.: The meta theory of e-government: creating some order in the
fragmented
29. Letaifa, S.B.: How to strategize smart cities: Revealing the SMART model Elsevier. J. Bus.
Res. 68, 1414–1419 (2015)
30. www.governodigital.gov.br/documentos-e-arquivos/digital-gov-review-brazil-portu-gues.pdf/
view
31. TCESP, IEGM/TCESP: Effectiveness Index of Municipal Management (2017). https://iegm.
tce.sp.gov.br/respostas.html
32. ECA-SP. https://iegm.tce.sp.gov.br/
33. IBGE: Profile of Brazilian municipalities: 2017/ IBGE, Coordination of Social indicators popu-
lation. IBGE, Rio de Janeiro (2017). http://portalods.com.br/wp-content/up-loads/2018/07/per
fil-dos-municipios-brasileiros-2017-IBGE.pdf
34. Weiss, M. C., Bernardes, R. C., Consoni, F. L.: Smart cities as a new practice for the management
of services and urban infrastructure: the experience of Porto Alegre. Metropolis, Rev. Bras.
Gest. Urban 7(3), 310–324 (EPUB) (18 September, 2015). ISSN 2175–3369. http://dx.doi.org/
10.1590/2175-3369.007.003.AO01
35. Brazil, Decree No. 7724 of 16 May 2012. Regulates Law No. 12,527, of November 18, 2011,
which provides for access to information under item XXXIII of art. 5, item II, § 3 of art. 37 and
§ 2 of art. 216 constitution. Official Gazette of [the] Federative Republic of Brazil. Brasília,
DF, Extra Edition, vol. 94, p. 1 (16 May, 2012)
36. Ministry of Science, Technology, Innovation, and Communications—MCTIC: Brazilian
Digital Transformation Strategy—E-digital (2018). http://www.mctic.gov.br/mctic/export/
sites/institucional/estrategiadigital.pdf