Page 1 of 6
Transactions on Engineering and Computing Sciences - Vol. 11, No. 4
Publication Date: August 25, 2023
DOI:10.14738/tecs.114.15155.
Sargsyan, S., Hovakimyan, A., & Antonyan, S. (2023). Creating a Security Assurance System Using IoT and ML Technologies.
Transactions on Engineering and Computing Sciences, 11(4). 51-56.
Services for Science and Education – United Kingdom
Creating a Security Assurance System Using IoT and ML
Technologies
Siranush Sargsyan
Yerevan State University, Yerevan, Armenia
Anna Hovakimyan
Yerevan State University, Yerevan, Armenia
Sveta Antonyan
Yerevan State University, Yerevan, Armenia
ABSTRACT
The work is dedicated to the development of security systems using the Internet of
Things and Machine Learning technologies. The need for such systems use is
associated with the increase in crimes, as well as with the increase in number of
attacks on special areas. The authors of the article have created a security system
that is located in a guarded area. The security system reacts in real-time and
informs about the expected danger through a picture or video. The security device
senses the movement of any object in the area through sensors, and the camera
connected to the device captures the moving object. A Siamese neural network
examines the image. If it is from the class of unwanted objects, the app sends it to
the Telegram app of the guard of the monitored area. The authors used Raspberry
Pi, various sensor components, and a Siamese neural network to create the security
system. The system can be implemented in bank vaults, offices, and special-purpose
buildings. We, being the professors of the university, attach more importance to the
use of the created system on the premises of the educational institution.
Keywords: Internet of Things, Machine Learning, security system, Siamese Neural
Network, classifier, Raspberry Pi.
INTRODUCTION
Nowadays, one of the primary problems in various fields of human activity is ensuring the
safety of the person and the given area. In the world, including Armenia, the use of security
systems become very necessary. People install personal security systems in their homes and at
the entrances of residential buildings. The operation of service centers, shops, banks, and hotels
is impossible to imagine without a security system. As a result of the expansion of the
consumption of security equipment, there is a need to improve such systems that are already
in operation, making them safer and more affordable.
The security system described in the article is intended to make the special purpose areas more
secure and exclude any kind of illegal movement there.
Page 2 of 6
52
Transactions on Engineering and Computing Sciences (TECS) Vol 11, Issue 4, August - 2023
Services for Science and Education – United Kingdom
To create the security system that uses the Internet of Things (IoT) technologies, the most
suitable option from the types of single-board computers was selected, and the structure
scheme of the device was modeled, taking into account the possibilities of its use in various
fields.
We used a Siamese neural network to solve the face recognition problem using Machine
Learning (ML) technologies. The security system software is written in the Python
programming language. The created security system can be used in special-purpose areas of
the educational institution, such as computer classrooms, laboratories, archives, etc.
STATEMENT OF THE PROBLEM
The purpose of the article is to describe the creation of a reliable and affordable security system
using IoT and ML technologies.
IoT devices are designed to collect, analyze, process, and transmit data [1]. Thanks to this
technology, it becomes possible to collect data in real-time, quickly respond to changing
environments and make decisions without or with minimal human participation. First, devices
collect data, for example, about the temperature in an apartment or a person’s heart rate, then
this data is sent to the cloud. The software handles them. If, for example, the temperature is too
high or there is a burglar in the house, the system notifies the user about this or performs
further actions itself, for example, calling the police. IoT systems operate in real-time and
usually consist of a network of intelligent devices and a cloud infrastructure to which they are
connected via WiFi, Bluetooth, or other forms of communication [1, 2, 3].
IoT devices collect data from various sources, supporting the learning process of an artificial
intelligence (AI) system so that it knows how to properly automate the desired operations.
Thanks to the IoT system, they can learn and make decisions about data management and
analysis [9, 10].
Thus, IoT and ML technologies jointly create intelligent connected systems. IoT combined with
the most dynamic AI technology can make the IoT system itself smarter and more easily mimic
human activity [12, 13]. The use of ML algorithms with IoT in different applications depends on
the field of application. These are often called smart IoT applications.
Analyzing data generated by smart IoT devices has applications in agriculture, healthcare,
autonomous vehicles, wearable devices, etc.
In the presented work, a smart IoT device is created using the Siamese neural network (SNN)
of ML [15].
HARDWARE AND SOFTWARE COMPONENTS OF A SECURITY SYSTEM
Raspberry Pi
A developed security system has hardware and software components. The hardware
component consists of a Raspberry Pi single-board computer, which has USB, Ethernet, CSI
Camera, Micro USB inputs, HDMI, Audio Jack outputs [4, 5].