ISSN(Online) : 2456-8910

International Journal of Innovative Research in Applied Sciences and Engineering (IJIRASE)

Volume 4, Issue 4, DOI:10.29027/IJIRASE.v4.i4.2020.729-734, October 2020

Doorbell Automation using AI and IoT for old Aged and Differently
Physically Challenged People
Anmol Singh1 , AnuragSinha2
1
Department of Computer Science and Engineering, Research scholar, Amity University Jharkhand
Ranchi, Jharkhand, India
2
Department of Information Technology, Research scholar, Amity University Jharkhand
Ranchi, Jharkhand, India
Email: [email protected] [email protected]

Abstract- In this paper, I propose a smart door and doorbell system for home automation using the Internet of things
(IoT) and Artificial Intelligence (AI) for old aged and differently able people to make their life easy and secure. In this
model, I am using Raspberry Pi for managing various interfaces and communications. The model consists of multiple
interfaces, including sensors, actuators and multiple inputs and out interfaces. To implement my IoT solution I have used
AWS (Amazon web service) IoT to provide many cloud services to support IoT-based applications. The apps give access
to IoT devices and the features the cloud services offer to those connected. The tools use various technologies and
protocols such as Wi-Fi or cellular data to communicate with the cloud. The mode of communication is via MQTT
(Message Queuing and Telemetry Transport) process. I have also used various software, including device software,
control services, data derives and core services provided by the AWS IoT, making this model efficient and scalable. The
overall aim of this paper is to make appropriate use of the technology for the needy and coming generations to come
Keywords-Smart Doorbell System, Raspberry pi, MQTT, AWS IoT
I. INTRODUCTION

Today, about 962 million of the global population is aged 60 servers, database, analytics. Raspberry Pi is also connected
and above and 10 % of the world's population is differently- to 3 sensors, which are two motion sensors and a one-touch
abled people as predicted by WHO. Nearly in all families, sensor. It has a bell that rings when an intruder or member
some old aged people need care with time. Their mobility rings the doorbell. I have used AWS IoT Device SDK as a
decreases with age, but they want to live the same way they software kit that connects hardware devices or mobile
used to live. Another section of people is the differently- applications, allowing devices to connect, authenticate and
abled people who have difficulty in speaking, moving, synchronize messages via IoT using the MQTT process.
seeing, or hearing things around. When an intruder rings the Face recognition Machine Learning models can be installed
bell, they often panic or want to open the door, but it takes with the help of AWS IoT Greengrass ML inference on
time. It's a matter of concern for their family members to Raspberry Pi, which can perform face recognition locally.
leave them alone in their homes due to the above reason and
for security issues. Someone needs to be around them. This II. LITERATURE REVIEW
gives rise to the feeling of inferiority
Antunes and Grilo [1] described a web-based intrusion
Advancements in technology such as Artificial intelligence alarm system. It works on developing control and
and the Internet of Things, which has a vital role in home monitoring hardware of the distributed web-based intrusion
automation, are now making life more comfortable. But no alarm system. The alarm system should have at least one
such IoT projects have yet been developed specifically for control panel, rechargeable battery, backup power supply
the above group of people. Thus, in this paper, I am and internal or external keypad, plush interior and a
proposing a theoretical model that can solve this problem surrounding intrusion detector and an external siren.
when implemented practically. In this paper, I am using
In [2] I saw a short message service (SMS) intrusion alarm
Raspberry Pi as an application platform as it is widely used
system for motorcycles. This guy is Owners of a potential
in recent years; it also provides comfortable use support and
intrusion into motorcycles via alarm and SMS (Short
documentation. Raspberry Pi is connected to internet access,
Message Service) PIC microcontroller programming.
such as Wi-Fi. It uses python programming to connect and
Research work included the PIC16F877A microcontroller
communicate with the amazon Web Service Internet of
interface. With a nokia6100 mobile phone connected in
Things (AWS IoT ) as AWS allows Bi-directional
simple dual mode.
communication between the devices connected to Internet,
embedded devices and various AWS IoT such as cloud

Vol. 4 (4), October 2020, www.ijirase.com 729

 ISSN(Online) : 2456-8910
International Journal of Innovative Research in Applied Sciences and Engineering (IJIRASE)
Volume 4, Issue 4, DOI:10.29027/IJIRASE.v4.i4.2020.729-734, October 2020

From [3], the authors studied the security and privacy issues
of the IoT.Four perspectives. First, they point out the
limitations of applying security to IoT devices (e.g., Battery
life, computer Power) and the proposed solution (e.g. Light
Encryption schemes designed for embedded systems). They
summarize the classification of IoT attacks (e.g., physical,
remote, Area, etc.). Thirdly, they focus on mechanisms and
architecture. Designed and implemented for authentication
and authorization purpose. Finally, analyze security issues at
multiple levels. (Ex. Physical, network, etc.).

The authors of [4] also demonstrated the possibility of real-

time remote monitoring via IoT-based doorbells. All of
these reveal endless applications for IoT-based intelligent
security systems. Security systems are often breached by
smart thieves, requiring deterrence measures to keep up with
the criminals.

Intelligent Transport systems that use visual motion sensors

Proposed by M. Soga and K.Yamada [5]. Andrew Burkett
has come up with an excellent plan. Connect the Raspberry
Pi to the base of the 6-axis gyro sensor And on the
MPU6050 chip. Moreover, Installing the sensor interface
and SMBus module with Simple Python code to read the
data Proposal [6].

III. Data Collection and its Significance

Finding the problem faced by the old-aged and differently-

abled people, I have taken an online survey which justifies With 60 responses, 81.7% of people contact old aged, and
the use of technology such as IoT and AI for this section of 45% contact differently-abled people. Out of which 73.3%
the society can become a problem solving and make their have a problem when they are alone in their home.
lives easier

Vol. 4 (4), October 2020, www.ijirase.com 730

 ISSN(Online) : 2456-8910
International Journal of Innovative Research in Applied Sciences and Engineering (IJIRASE)
Volume 4, Issue 4, DOI:10.29027/IJIRASE.v4.i4.2020.729-734, October 2020

V. Model Description

I am using AWS IoT as it is the only cloud service that

integrates data management and rich analysis of data which
is easy to use with its multi-layered security and superior AI
integration AWS IoT, which has been proven a secure IoT
cloud platform. I also chose this cloud platform as it is
scalable. It offers different software and services like Free
RTOS and AWS IoT Green grass, a microcontroller OS that
makes low-power devices easy to deploy and manages local
messaging, synchronizing, and machine learning inference.
The connecting and control service like AWS IoT Core,
AWS IoT device defender and AWS IoT device
73.3% of security and safety is the primary concern, and management allows secure connection with the cloud, audit
83.3% agreed that doorbell automation could solve the old and monitor remotely. Analytics services like AWS IoT
aged and differently-abled people's above problems. analytics, AWS IoT Sidewise, AWS IoT events, AWS IoT
thing graph operates on the data produced by the things and
From the above analysis, I conclude that technologies like extract data from the IoT data stored in the cloud.
IoT and AI can solve the problems faced by this section of
society. Thus, in the next section, I propose a theoretical Mosquitto Broker
model that can efficiently solve these security problems
I have used Mosquitto broker as a lightweight open-source
when implemented.
message broker written in C language. It has a good
IV. Proposed Model Architecture collection of libraries to implement MQTT, suitable for low-
power computer board computers like Raspberry Pi to full
servers like AWS IoT. In this model, I am using this broker
for two-way communication between Raspberry Pi and
AWS IoT.

Python

With a clean syntax and robust programming, python

programming connects my project to the real world. Python
comes pre-installed on Raspbian and the Raspberry pi
foundation mainly selected this language as the primary
language because of its versatile nature and ease of use. It
can be used to read input from the physical sensors
connected to Raspberry Pi and send output to an external
component using Python scripts. Simultaneously, AWS IoT
Device SDK for Python enables developers to create and
write python scripts that will allow their devices to access
the AWS IoT cloud via MQTT.

The IoT things connected to AWS IoT are:

1) Raspberry Pi 4 -It is the latest low-cost Raspberry pi

computer that has Raspbian OS. It is connected to various
devices called 'things' and 'Internet Gateway' So the sensors
connected to raspberry pi provide things with the Internet
Gateway.

Figure 1: Main Model Raspberry pi 4 has a micro SD card slot for loading the
Raspbian operating system like other Raspberry computers.
Pi 4 models have a Broadcom BCM2711, quad-core Cortex-
Vol. 4 (4), October 2020, www.ijirase.com 731
 ISSN(Online) : 2456-8910
International Journal of Innovative Research in Applied Sciences and Engineering (IJIRASE)
Volume 4, Issue 4, DOI:10.29027/IJIRASE.v4.i4.2020.729-734, October 2020

A72 (ARM v8) 64-bit SoC at 1.5GHz .Memory is 4GB and navigate to IoT core. All the devices are registered to the
40pins – GPIO headers. AWS IoT core. Certificates are downloaded and installed
and policy is created. Mosquito MQTT is bridged with
2) Motion Sensor- The motion sensor is PIR Motion Raspberry Pi and AWS IoT. There is a standard android app
Sensor placed under the door lock. It senses human/object developed for the client. All the necessary python scripts are
when approaches it at a specific distance and activates the being created.
required task.
Implementation
3)Motor sensor-Motor sensor is a Servo motor, which acts
as a door lock. It rotates to a specific rotation angle when When an intruder stands in front of the door for more than 2
PWM (pulse width modulation) signals are sent to the minutes or when someone rings the doorbell, the sensors
motor. gets activated. The camera captures the photo/video. The
Raspberry pi gets notified and rings the doorbell and display
4)Touch sensor -There are two touch sensors: the the photo/video to the people inside the house. Raspberry pi
Momentary capacitive touch sensor, which is used as is connected to Wi-Fi, collects the data and sends it to the
Doorbell and Control 2, which will activate as long as broker. The AWS IoT further receives the broker's data and
someone is touching it. Control 2 will unlock the door from sends an alert message to the people who are not in their
inside. home via the android app notifying the presence of an
intruder in their home. Raspberry pi further receives the
5) Camera - The Raspberry Pi Camera Module v2 has Sony
IMX219 8-megapixel sensor that spectator humans via relevant information from the central control via AWS IoT.
video or image. The remote control gets activated and the people in the
house can open the door from 1 press of a button.

6) Control 1 & Remote control- Here, control 1 can be a The key model is shown in Fig1 because the customer wants
remote or an Android application with AWS IoT mobile it. It can be changed and it is possible to delete certain
SDK installed, which controls the door's opening and components and render sub-models.
closing. The remote control can be an Android app that
will notify the family members when an intruder enters the  For old aged people who don't live with their
house. family will not require an android app.
 People who have a problem in hearing may not
7) Bell - Here, a HiFi Berry DAC HAT + ADC for audio need a microphone and bell or speaker. External
and speaker can be installed, which will act as an alarm that vibrators can be installed with the Raspberry pi to
will notify the people inside the house. DAC+ADC provides feel the presence of an intruder.
perfect audio production by which we can record and play  For people who cannot see, LCD would not be
an analog input. required.

8) Display- It can be a small display screen LCD that can be

used to Espectate the outside the door.
Working Mechanism
VI. Configuration and Implementation
This model works on the MQTT (Message Queuing
Configuration Telemetry Transport) protocol for the exchange of data
between clients and servers. I have used MQTT because it is
Raspbian OS is being installed in Raspberry Pi 4 computers.
a light weight messaging protocol that uses TLS (Transport
All the input and output units are PIR motion sensor, Servo
Layer Security) encryption. It is a real-time low power and
motor,Raspberry Pi Camera Module v2, Momentary
network usage messaging protocol. The standard MQTT is a
capacitive touch sensor, One buttoned remote control, LCD
broker based protocol.
screen, and HiFi Berry DAC HAT + ADC installed with
speaker and microphone are being configured. The sensors
are being connected through the GPIO pins of the Raspberry
Pi. A fully secured door lock is built and the safety is
performed by servo motor. The next step is to log in with an
existing or new account to AWS official website. Then

Vol. 4 (4), October 2020, www.ijirase.com 732

 ISSN(Online) : 2456-8910
International Journal of Innovative Research in Applied Sciences and Engineering (IJIRASE)
Volume 4, Issue 4, DOI:10.29027/IJIRASE.v4.i4.2020.729-734, October 2020

Advantages and Limitations

Advantages:

 The above model can make life easier and efficient

for the old-aged and differently challenged people.
 Increases security and safety for the client.
 Raspberry pi being of low cost and massive
processing power, can support multiple interfaces
simultaneously, making it easy to build
applications as it supports languages such as
Python and Linux.
 MQTT provides Quality of Service Levels (QoS)
and can carry any data. It uses low sized packets
and can be used for low bandwidth applications
with low power consumption.
Figure 2: Working model  Easy monitoring, controlling, managing, adding
devices and data from and to AWS IoT cloud
From fig 2., when an intruder stands outside the door or service with best security practices is provided at a
rings the doorbell, the various sensors sense their presence scalable price, i.e., we pay less for the best we get.
 Controlling the entire system is easy and takes no
and send signals electrical signals to the Raspberry pi (rings
time to learn for the client.
the doorbell). The Raspberry pi is the sender cloud that Limitations :
publishes this data from the various IoT devices (such as
doorbell and door lock sensors) of a particular IP  It can create difficulty for blind people to adapt to
address/client ID to the Mosquito broker. The broker breaks the system compared to other people.
the message into topics.AWS IoT being the receiver cloud,  The platform sets AWS IoT services limits. There
has already subscribed to specific issues and receives is a technical support fee with general cloud
messages from the broker. AWS IoT then notifies the computing issues.
controller (mobile and remote control) of the relevant and  There are heating issues with the high performance
required message received from the broker, i.e., intruders in and processing speed of Raspberry Pi.
front of the door. Then the controller performs the necessary  MQTT operates over TCP and it often affects long-
action and notifies back to the AWS IoT. This time cloud term battery consumption. As networks grow large,
then publishes the related messages in the form of topics brokers can affect scalability.
back to the Mosquito broker. Raspberry pi further subscribes  Data transfer totally depends on Wi-Fi.
to those topics and receive information and perform the final
 This model, if implemented for ordinary people,
task (opening and closing the door).
can make them lazy.
AWS IoT also offers AWS IoT Green grass, which includes
VII. Conclusion
pre-built Amazon Sage Maker for devices like raspberry pi The paper is based on the design and implementation of
so we need not build or train machine learning models from Door and Doorbell automation. This is an entirely sensor-
scratch. AWS IoT Green grass offers ML inferences locally based IoT model. This model can help the old aged people
on the IoT devices. AWS IoT Green grass also provides pre- and differently-abled people in various aspects. With the
trained Machine Learning models of Amazon Sage Maker, help of a single button, they can open the door when
which we can directly deploy to the cloud. The inferences someone rings the doorbell and they get notified by their
presence. At the same time, in their android app, household
can be performed locally on devices, which reduces the cost
members who are not with them are alerted through the
of sending IoT device data back to the cloud for machine cloud when the door is opened or when someone rings the
learning inferences. It can detect the outliner and send the doorbell. This model can be modified as per the need and
data to the cloud and further; Amazon Sage Maker can requirement of the client. For example, a bulb can be placed
reclassify and restrain and improve machine learning for the people who have a problem in listening so that they
models. can get notified of the presence of an intruder by the
blinking of the bulb. For others who have a problem in
seeing get notified via a small microphone and speaker,
which gets activated when someone rings the doorbell.
Others can easily via a small LCD screen placed inside the
house. This model is secure as all the data are being stored
Vol. 4 (4), October 2020, www.ijirase.com 733
 ISSN(Online) : 2456-8910
International Journal of Innovative Research in Applied Sciences and Engineering (IJIRASE)
Volume 4, Issue 4, DOI:10.29027/IJIRASE.v4.i4.2020.729-734, October 2020

in the cloud. One can access it at any time. Thus the above
model is cheap concerning the functionalities it provides to
the client.

This can become a life-changing tool for the needy and with
its demand after testing and manufacture, it can become a
source of employment for many people worldwide. In the
future sub, models can also be developed for ordinary
people for everyday use.

References
[1] Autunes, R.M., and Grilo, F.L. (2009) unwelcome
person Alarm System: The Road Ahead. Advanced
Technologies, Kankesu Jayanthakumaran (Ed.), ISBN: 978-
953-307-009-4, InTech.

[2] Rafiz, M. (2010) "Intruder Alert System via SMS for

Motorcycle". unpublished Thesis work, University of
Malaysia

[3] Yang Y , Wu L , Yin G , Li L , Zhao H . A survey on

security and privacy problems in internet-of-things. IEEE
net Things J 2017 .

[4] Ping W, Guichu W, Wenbin X, et al. (2010) Remote

Monitoring Intelligent System supported Fingerprint Door
Lock, International Conference on Intelligent Computation
Technology and Automation.

[5] Yamada, K. and M. Soga, A compact integrated visual

motion device for ITS applications. Intelligent
Transportation Systems, IEEE Transactions on, 2003.

[6] Azrina Abd Aziz, Y. Ahmet S¸ekercioglu, Paul

Fitzpatrick, and Milosh Ivanovich, "A Survey on Distributed
Topology Control Techniques for Extending the period of
Battery power-driven Wireless device Networks", IEEE
communications surveys & tutorials, VOL. 15, NO. 1, half-
moon 2013.

Vol. 4 (4), October 2020, www.ijirase.com 734