Available online at www.sciencedirect.

com

ScienceDirect
Procedia Engineering 145 (2016) 396 – 403

International Conference on Sustainable Design, Engineering and Construction

An Internet-of-Things (IoT) system development and

implementation for bathroom safety enhancement

Dan D. Kooa,*, John J. Leea, Aleksei Sebastiania, and Jonghoon Kimb

a
IUPUI, 799 W. Michigan St. ET 314J, Indianapolis, IN, 46202, USA
b
Arizona State University, 660 S. College Avel, Tempe, AZ, 85287, USA

Abstract

Statistics show that a bathroom is one of the most hazardous places especially for older people. Older people typically have
greater difficulties with mobility and balance, making them more vulnerable to fall and slip injuries in a bathroom and causing
serious health issues related to short and long-term well-being. Various components in a bathroom including shower, tub, floor,
and toilet have been re-designed, and independently upgraded their ergonomics and safety aspects; however, the number of
bathroom injuries remains consistently high in general. Internet-of-Things (IoT) is a new concept applicable to almost
everywhere and man-made objects. Wireless sensors detect abnormalities and send data through the network. A large amount of
data can be collected from multiple IoT systems and it can be utilized for a big data analysis. The big data may reveal a hidden
positive outcome beyond the initially intended purposes. A few commercial IoT applications such as wearable health monitoring
and intelligent transportation systems are available. Nevertheless, An IoT application for a bathroom is not currently known.
Unlike other applications, bathrooms have some unique aspects such as privacy and wet environment. This paper presents a
holistic conceptual approach of an Internet-of-Things (IoT) system development and implementation to enhance bathroom safety.
The concept focuses on the application in a large nursing care facility as a pilot testing bed. Authors propose 1) sensor selection
and application, 2) integration of a wireless sensor local network system, 3) design concept for IoT implementation, and 4) a big
data analysis system model in this paper.

©©2016
2015Published
The Authors. Published
by Elsevier by Elsevier
Ltd. This Ltd.
is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of organizing committee of the International Conference on Sustainable Design, Engineering
Peer-review under responsibility
and Construction 2015. of the organizing committee of ICSDEC 2016

Keywords: Internet-of-Things; Bathroom Safety; Wireless Sensor Network; Big Data

* Corresponding author. Tel.: +1-317-278-1957

E-mail address: [email protected]

1877-7058 © 2016 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the organizing committee of ICSDEC 2016
doi:10.1016/j.proeng.2016.04.004
 Dan D. Koo et al. / Procedia Engineering 145 (2016) 396 – 403 397

1. Introduction

Injuries associated with bathroom activities cannot be ignored. According to the Center for Disease Control and
Prevention [1], About 235,000 people visit emergency room per year because of injuries caused in bathrooms.
Injuries increase with age; especially people over 85 suffer more than half of their injuries near toilets. The
bathroom injury rate for women was 72 percent higher than that for men. Three most hazardous activities for all
ages are bathing, showering, and getting out of a tub. People who have degenerative joint disease and joint or other
senile disease suffer disability while using a toilet. Similarly, according to New York Times, over 30 percent of
total injuries associated with activities in bathrooms are head and neck parts. Main activities are bathing, using a
toilet, and leaving and entering a tub. The injury rate is significantly increasing as people become older over 60
years [2].
IoT development in healthcare industry has focused on wearable and in-body sensor devices such as digital pill,
smart bed, smart food, and smart band-aid. Wearable devices are typically attached in various body parts including
head, neck, wrist, waist, and leg to get specific metabolic data [3]. However, wearable devices are not suitable to be
used in bathroom environment and private activities because of wetting and undressing clothes. In addition,
wearable devices need constant re-charging of power and maintenance to maintain functionality and network
connectivity. Therefore, sensors shall be installed in objects in a bathroom such as toilet, bath tub, floor, ceiling,
light fixtures. These unique aspects of bathroom environments differentiate from other IoT applications.
IoT is an emerging IT technology. A main function of IoT is to collect data measured by sensors integrated with
short range wireless networks such as Bluetooth, ZigBee, or Wi-Fi, which again transmit data to larger networks
such as Internet network gateways. IoT sensors provide low cost, scalable, efficient, low power, and integrated data
through all sub-networks. As more sensors are incorporated and data collection period increases, the data becomes
significantly large and hence the name “Big Data”. Big data was introduced by Gartner Report in 2001 [8] and had a
threefold definition encompassing 3Vs: Volume, Velocity, and Variety. This definition has been reiterated by others
to include a fourth V: Veracity [9]. In summary, IoT provides means of data collection, detection of monitoring
events, algorithms to act on the events, and big data storage and analysis.
The main objective of this research is to provide a conceivable IoT system and to develop a big data model. This
paper presents a holistic conceptual approach of an Internet-of-Things (IoT) system development and
implementation to enhance bathroom safety. Authors propose 1) sensor selection and installation, 2) integration of a
wireless sensor local network system, 3) design concept for IoT implementation, and 4) a big data analysis system
model in this paper.

2. Sensors for Bathroom Applications

Most electronic sensors measure electric resistance or energy occurred by physical and chemical changes. Electronic
signals can be digitally processed in an on-board microcontroller or directly transmitted to other devices through
short-range wireless networks. Many sensors are commercially available, but require signal calibration, hardware
modification to fit to existing bathroom objects, and configuration to connect to sensor communication networks.
Below is a list of commercially available sensors.
x Leak Detection Sensor: It measures electronic resistance between two nodes. It can be installed on a bathroom
floor to detect water leakage and flooding especially when an injured person is unable to move.
x Digital Light / Lux Sensor: It measures intensity of light radiation. Light sensors detect most spectrum ranges
including infrared ray, visible ray, and ultra-violet ray. The sensor can be used to detect human-presence and
movement, total time spent in a bathroom, and identification and frequency of a person in a bathroom.
x Voice Detection Sensor: It detects acoustic and noise signals. It is very useful as it can translate pre-programed
messages (e.g., Help!) through a microcontroller in the case of medical emergency.
x Pressure Sensor: It detects magnitude of contact-pressure. A simple pressure sensor is able to detect a person’s
status of using toilet and bath tub. A digital load cell sensor is same kind but provides more precise pressure
readings that can be used for weight pattern analysis.
398 Dan D. Koo et al. / Procedia Engineering 145 (2016) 396 – 403

x Positional Sensor (Gyro/Accelerometer/Magnometers): It detects tilt, pitch, and inclination. These sensors can
be used to detect falling, tilting, and locational changes for installed objects such as bath tub, shower room, and
bathroom floor.
x Motion Sensor: It detects movement of people in a bathroom. Mobility in a bathroom can provide pattern of
bathroom usage and alarm for emergency situation.
x Water Flow Sensor: Water leakage detection and measurement of water usage are important data for saving
water and monitoring health conditions. Unusual water usage may indicate an emergency situation such as
unconsciousness before finishing a shower or unnoticed internal plumbing leakage.
x Energy Harvesting Sensor: One of key requirements for IoT sensors is low power consumption to minimize
maintenance of those sensors. Energy harvesting sensors generate small amount of electronic energy enough for
their operation from the sensor device or adjacent energy harvesting medium such as sunlight, pressure, and
hydraulic flow.
Although the aforementioned sensors are at the lowest hierarchy of an IoT system, their functions/capabilities can be
used in various ways for this study, which are summarized as follows:
x Water consumption and flushing sequence in a toilet
x Pressure distribution on a toilet cover
x Deriving medical conditions of human from feces and urine
x Water consumption, water temperature, event sequence in shower and bath tub
x Falling or slipping detection from impact and pressure sensors
Our preliminary sensor study demonstrates suitability of the off-the-shelf sensors in bathroom environment
conditions including, but not limited to, high humidity, wet and submerged in water, and confined and private space.
However, those sensors are designed for normal operational conditions that may not be suitable for specific
bathroom environmental conditions. Thus, appropriate modification is required.
A typical sensor module consists of four main components/functions including power supply, sensor part, actuator,
and communication port. Figure 1 shows layout example for sensors that can be installed in a toilet and a bath tub.

Figure 1. Layout Example for Toilet and Tub Sensors

3. Sensor Network and IoT Application

3.1 Sensor Network and Bathroom Design for IoT Implementation

Figure 2 illustrates a sensor network integrating multiple and heterogeneous (digital and analogue) sensor data
through a short-range network within a bathroom. The short-range network transmits data to a larger network
through a gateway or hub. Short-range networks are then integrated through multiple gateways and connected to a
virtual cloud serving as a database server.
 Dan D. Koo et al. / Procedia Engineering 145 (2016) 396 – 403 399

Such an IoT and big data system can be implemented in a pilot test bed. Basic functions and installations of a
bathroom are similar regardless of difference in architectural design. Sensors shall be embedded in the fixtures and
bodies of objects such as inside of floor, ceiling, and wall. Interior design of a bathroom pilot test bed focuses on
alleviating black spots of network communication, improving convenience of sensor operation and maintenance,
meeting design requirements, and enhancing well-being of older people. The proposed design for the pilot test bed is
shown in Figure 3.

Figure 2. Sensor Controller, Algorithm, and IoT Network Flow Diagram.

3.2 Design of a Pilot Test Bed
Our pilot test bed consists of multiple groups of sensors, IoT networks, and big data components. Data analysis
algorithms and user services shall be tested to ensure functionality and feasibility of the system. Utilities including
electric power, water supply, and wastewater drainage will be connected to sink, toilet, and shower booth to be fully
functional. All sensors and networks will be installed in several levels from short-range sensor networks to long-
range networks via gateways. A virtual cloud server will be connected and completes a full circle of data
transmission and storage. As shown in Figure 3, the pilot test bed plans to utilize Temboo, IFTTT, Arduino IDE,
XCTU [4], Microsoft Azure and Power BI. Temboo is used to automatically compose sensor as well as gateway
code/programs from function libraries. Arduino IDE is used to fine-tune the code. XCTU is used to test and build
short-range networks. Azure and Power BI are used to store and visualize sensor data. The programs are installed on
sensors and gateways. When integrated, IFTTT can be used to interface clouds services for data logging and
notifications. The pilot test bed provides a baseline of research outcome implementation to a business case. Testing
results and lessons from building a pilot test bed can be used to improve design and construction process of an IoT
bathroom.
3.3 Big Data System Development
One of main objectives of this study is to develop a big data concept for the data collection, analysis, and
implementation. Figure 4 illustrates the configuration of our big data analysis prototype system. It consists of four
parts: a sensor unit, a hub unit, a base station, and a main server.
The sensor unit is composed of various sensors and actuators, but is grouped for easier management and
geographical distance and connected to a hub in the region. The sensor unit and information gathering hub has
similar structure to that described in detail in Figure 2. The difference here is that we aim to have a more
standardized and bigger scale, but in terms of functionality, we plan to shrink a little because in this sub-task, data
processing will be done more on the primary server and the base station.
A hub is responsible for connecting between a group of sensors/actuators and the base station, and supports both
wired and wireless Machine-to-Machine (M2M) communication. In addition, hubs are used for setting sensor
configurations and the development and improvement of sensor program and management algorithms.
A building base station serves to connect the information acquisition target buildings and the main server, and serves
to provide a real-time analysis service for administrators, care recipients, and their guardians. The user information
400 Dan D. Koo et al. / Procedia Engineering 145 (2016) 396 – 403

module stores encrypted personal information of care recipients, and security management module provides a user
login and logout service and secure connection, and generates a session ID to maintain security [7]. The service
broker receives queries from users and is responsible for transmitting sensed values/information and processed
information to the users. The building base station includes a router and gateway that supports short-distance wired
or wireless communication. Main mission is to process real-time analysis commands of an administrator and user’s
service queries. The human resource information module stores information of allowed users, and the security
management module manages and controls their secure connection. The service broker module obtains data and
information from both sensors and the main server to fulfil user queries, and additionally sends periodic
notifications (for example, automatic safety check function is invoked when bath time is prolonged abnormally) or
recommendations. The real-time analysis module senses an accident (for example, falls) or event that requires
immediate action and sends a warning or processing feedback to the necessary components and recipients, and
provides alert to the administrator.

Figure 3. Pilot test bed design and installation layout.

The main server is a core intelligent computer used to efficiently process and analyze large sensor data, and store the
results for research and technology development of comprehensive and intelligent processing, management, display,
and empirical evidence. Here we develop methods and algorithms that comprehensively handle a variety of sensor
events, knowledge database construction, implement a decision-support expert system, and support the work of the
administrator through the administrator interface. To effectively classify, store, and handle diverse and continuous
stream of sensor data, this study will use an open source Oliot EPCIS [5] event-based sensor data processing
middleware that shows excellent scalability and flexibility. If we are unable to intelligently and smartly process,
display, and manage the stream data that are collected from a variety of sensors, it will only add confusion.
Therefore, the technology research, development, and demonstration for a structural abnormality monitoring and
management are also important. The main server provides storage service for the sensor data coming from all
managed bathrooms, and provides data analysis and its results to the administrator group. Administrators can
 Dan D. Koo et al. / Procedia Engineering 145 (2016) 396 – 403 401

develop and run a wide variety of algorithms for analyzing and processing data through the administrator interface.
A knowledge storage, an expert system, and a decision support system serve to aid in the various operations of
administrators. If needed, the system takes immediate action by sending messages or leaving voicemails to care
recipients’ guardians, physicians, and emergency crew. For this purpose, we plan to use a fault-tolerant, scalable,
open-source Apache Cassandra that supports column-oriented storage, and is best for schema flexibility and
efficient big data storage and management.

Figure 4. Big data system data and control flow diagram, modified from Ahmed et al. [6]
3.4 Big Data Acquisition, Analysis, and Application
Big data analysis and application focus on the utilization of collected data from bathrooms. The data we collect form
bathrooms are water consumption in toilets, the frequency of toilet usage, pressure distribution on toilet covers,
toilet usage and time, toilet flushing interval and sequence, water consumption in shower and bath tubs, temperature
and humidity for shower and bath tubs, shower and tub usage and time, impact and pressure sensors for falling or
slipping, as well as motion and scream voices.
The big data collected via a bathroom IoT system can be used for various purposes. One of the most important
purposes is to analyze data in order to develop an algorithm used to derive individual personalized data patterns. In a
variety of modern living lives, life patterns for individuals differ significantly from others, and thus, it is not
appropriate to issue a warning based on criteria that are established with uniform standards. Rather, first set
personalized criteria by observing the lifestyle of each individual over time, and then, in case a situation differs from
the individually set criteria, an emergency or appropriate warning should be sent. As many participants need to be
observed and various sensor data exist, the study is to derive complex logic by analyzing and processing
individually tailored data patterns. Figure 5 illustrates an example of a complex logic flow chart. For example, from
the flow chart, falling possibility could be raised when several conditions are met such as screaming and large area
of pressure on the floor detected while no pressure detected both on toilet and shower tub for a specified period.
An early-warning capability is another big data application example. Once individual criteria are established, by
using big data predictive technology, when anticipated events do not occur or unanticipated events occur, alarm can
be raised and thus important action can take place. The following four steps are proposed as development guidelines
for individually tailored data pattern analysis and utilization.
402 Dan D. Koo et al. / Procedia Engineering 145 (2016) 396 – 403

1) Development of algorithm for personalized criteria generation, storage, management, and abnormality
detection.
2) Development of customized warning system algorithms with respect to alarming stage and magnitude of injury
3) Development and application of algorithms for the efficient management and use of big sensor data obtained
from IoT network.
4) Customized service and rule creation for system administrators and customers

Figure 5. The flowchart of bathroom IoT system operation.

4. IoT Toilet Test

A small-scale IoT application has been tested using two round Force Sensitive Resistor (FSR) pressure sensors and
one CdS photocell sensor. A set of electroluminescent (EL) wire was placed under a toilet cover to provide light in
dark night. The hardware components of this system include an open-sourced development board and other open
hardware sensor breakout boards.

Data are collected from the analog sensors by a Wi-Fi-enabled microcontroller (Arduino Yun) and algorithms were
devised to act on appropriate events. For example, if a “lights on” event is triggered from a low light level in the
bathroom, the toilet interior lighting is powered on. The real-time data obtained from sensors are sent via Wi-Fi to a
service called Temboo, which will pass them to Microsoft Azure’s SQL database as Yun cannot communicate
directly with Azure. Azure functions as a data storehouse for the big data created by many nodes. For data
visualization, Microsoft Power Bi polls the SQL database and graphically displays the sensor data in real-time as
shown in Figure 6. Power Bi visualizes user-designated parameters, weather graphs, spreadsheets, or pie charts, with
any set of manipulation that could normally be performed in Microsoft Excel. In Figure 6, the display is showing a
brief segment of the recorded time and specified sensor data. The green line represents light levels and the black line
 Dan D. Koo et al. / Procedia Engineering 145 (2016) 396 – 403 403

represents pressure on the toilet seat. The large spikes in the waveforms are corollary to the discrete events of force
applied to the toilet seat and lights being turned on (when no force detected) or off (when force detected).

Figure 6. The outputs of the sensors visualized using Power Bi (provided by Microsoft).

5. Conclusion

Information technology advances in theG development of IoT and big data toward various aspects of human
activities. Bathrooms are one of the most dangerous places especially for older people. The IoT and big data concept
introduced in this paper proposes four objectives: 1) sensor selection and application, 2) integration of a wireless
sensor local area network system, 3) design concept for IoT implementation, and 4) a big data analysis system
model. Sensor study reveals that off-the-shelf sensors are available for IoT applications, but many of them require
hardware modification and software configuration to integrate into small range network systems. Several small
range wireless sensor network protocols are available and are able to connect to larger network systems such as
Internet through various levels and types of hubs/gateways. The study proposes a design concept for a pilot test and
big data system. Various software applications and open source platforms are reviewed and will be tested to consist
of large-scale big data, which is based on the data acquired from sensors detecting bathroom activities. As an
example, a test was conducted using an IoT application in a toilet. Several sensors including pressure sensors are
tested to verify the concept in a small-scale example. Potential impact of this study is not only to timely inform
special events but also to reduce hazardous conditions in bathroom environment and probability of severe
implication from bathroom injuries especially for older people. Moreover, this study demonstrates tremendous
potential of IoT and big data applications for improving user sustainability in various aspects.

References

[1] Center for Disease Control and Prevention (CDC) (2008), Nonfatal Bathroom Injuries Among Persons Aged over 15 Years in United States,
June 10, 2011, accessed from: http://www.cdc.gov/mmwr/preview/mmwrhtml/mm6022a1.htm
[2] Bakalar, N. (2011), Watch Your Step While Washing Up, The New York Times, Aug. 15, 2011, accessed from:
http://www.nytimes.com/2011/08/16/health/research/16stats.html
[3] Park, J., Chun, S., and Ko, S. (2014), Trend for IoT based Healthcare Service and Platform, Information and Communication Magazine, The
Korean Institute of Communications and Information Sciences, Vol. 31, No. 12, accessed from:
http://society.kisti.re.kr/sv/SV_svpsbs03VR.do?method=detail&cn2=GCSHBA_2014_v31n12_25&menuid=1&subid=11
[4] Digi International (2016), XCTU – Next Generation Configuration Platform for XBee/RF Solutions, accessed from: http://www.digi.com/xctu
[5] Byun, J. and Kim, D. (2015), Oliot EPCIS: New EPC Information Service and Challenges towards the Internet of Things, 9th Annual IEEE
International Conference on RFID 2015, San Diego Conventional Center, April 15 – 17, 2015, San Diego, CA.
[6] Ahmed, M.U., Björkman, M., and Lindén, M. (2015), A Generic System-Level Framework for Self-Serve Health Monitoring System through
Internet of Things (IoT), Studies in Health Technology and Informatics, Vol. 211, pg 305-307.
[7] Financial Institutions and Customer Information: Complying with the Safeguards Rule, https://www.ftc.gov/tips-advice/business-
center/guidance/financial-institutions-customer-information-complying
[8] L. Douglas. (2001), 3D data management: Controlling data volume, velocity and variety. Gartner. Retrieved, 6, 2001.
[9] IBM (2013), What is big data? - Bringing big data to the enterprise. http://www-01.ibm.com/software/data/bigdata/