WestminsterResearch

http://www.westminster.ac.uk/westminsterresearch

IoT Motion Detection Sensors for Monitoring in a Smart Campus

Ucar, C., Maloku, M., Yugay, O. and Budimir, D.

This is a copy of the author’s accepted version of a paper subsequently published in the
proceedings of the 13th Mediterranean Conference on Embedded Computing
(MECO’2024) & International Conference on Cyber-Physical Systems and Internet-of-
Things (CPS&IoT’2024), Budva, Montenegro 11 - 14 Jun 2024.

The final published version will be available online at:

https://doi.org/10.1109/meco62516.2024.10577922

© 2024 IEEE . This manuscript version is made available under the CC-BY 4.0 license
https://creativecommons.org/licenses/by/4.0/

The WestminsterResearch online digital archive at the University of Westminster aims to

make the research output of the University available to a wider audience. Copyright and
Moral Rights remain with the authors and/or copyright owners.
 IoT Motion Detection Sensors for
Monitoring in a Smart Campus

Cihan Ucar Muharrem Maloku Olga Yugay

School of Computer Science and School of Computer Science and Wireless Communications Research
Eng. Eng. Group
University of Westminster University of Westmnster University of Westminster
London, UK. London, UK. London, UK, and Business
Information Systems
Djurradj Budimir Westminster University in Tashkent
Wireless Communications Research Tashkent, Uzbekistan
Group, School of Computer Science
and Eng. University of Westminster,
London, UK, and School of
Electrical Engineering,
Belgrade, Serbia.
[email protected]

Abstract—This paper presents the study of the monitoring of As of 2025, more than 30 billion devices will be connected to
student attendance in a Smart Campus using the IoT Motion the internet. While 5G rolls out at neck-breaking speed, the
Detection Sensors, and the implementation of IoT technology in resulting evolution in communications will bring the world a
educational institutions. The proposed solution is far more secure faster, smarter future. By the estimation of 5G Americas and
than the existing RFID monitoring system. The IoT devices are
embedded into the existing campus environment for data to be
analyst firm Omdia, 5G is on track to hit 8 billion connections
collected, transmitted through Wi-Fi using MQTT protocol and by 2028, at its ten-year mark—exceeding LTE’s 10-year
store data in local server. The collected data is then accessible to connection total by more than 2.5 billion.
the management with real time insights and attendance pattern. With the help of the wide arrays of devices connected over the
The proposed solution was tested in a one UG course in real time internet, IoT has greatly transformed sectors such as
over a period of 8 hours (9:00 – 13:00 and 14:00 -18:00). agriculture, healthcare, entertainment, and education [1]. As a
result of the increased data connectivity, communication has
Keywords—Internet of Things (IoT), RFID, MQTT, classroom reached a significant level, which has had enormous practical
occupancy, smart campus. advantages. These devices can monitor and control their
I. INTRODUCTION environment with the help of real-time data inputs from several
other interconnected devices. The education sector has evolved
The recent growing advancements in technology and from traditional blackboards to projectors and online
increase in computational power pose a serious threat to the presentation classes; similarly, IoT can catapult the evolution of
security of user data. The Internet of Things can be the education sector. Even though IoTs are already being
implemented to eliminate cyber security risks to a certain level implemented in educational institutions, such as RFIDs to
and also to make better use of the available resources to monitor attendance and provide access to the facility, the
improve the quality of service. Universities are under pressure educational platform can still be improved by developing a
to make better use of their available resources due to the rising smart environment to monitor and control certain tasks, which
demand for higher education. Even though IoT system may increase security and make better use of resources. With
development and implementation are very popular in education the aid of IoT, academic staff may easily engage with remote
institutions, there is a concern that security is still not given students who are in different locations and control the
enough consideration. Universities are adopting creative instruction in smart classrooms using gestures, speech, and
solutions in today’s rapidly evolving technology landscape to facial recognition. Additionally, they may automatically gather
protect the security of data and replace traditional methods with data from sensors, analyze analytics on student behaviors,
more cost-effective IoT solutions. interest, performance, and engagement, and collect feedback
from students on their interests and the likeability of a teaching
session [1]. Certain schools in China have already started using TABLE I. COMPONENTS REQUIRED
IoT devices to monitor and assess the emotions of the students, Software components Hardware Components
including their mood changes and their attention level. Which Gateway Handset
are processed and sent as feedback to the students and their Docker and Docker containers Raspberry Pi 5
caretakers. This method of using IoT to assess student’s Mosquito (MQTT broker) Raspberry Pi Pico W IoT node
protocol
emotions and distress levels has been proven to improve their Influx DB Micro PIR Motion sensor
academic results. An Internet of Things (IoT) system enables Grafana Breadboard
data interchange across an extensive network and gives users Node Red
management over connected as well as wireless devices.
Through various communication interfaces, data may be
gathered as input, stored in a computer system, processed, and B. IoT System Implementation
utilized by other applications. Wirelessly linked devices As mentioned earlier, we have integrated the Raspberry Pi
provide users with the liberty to function over remote networks Pico W and Micro pir motion sensors together to collect,
with more control, which improves overall productivity [2]. process, share, and store data. The IoT system is equipped with
motion sensors integrated into a Pico W on a single-piece
There are number of available approaches to count the number breadboard such as shown in Fig.1. The sensors are used to
of students in a classroom, such as thermal imaging, ultrasound capture measurements such as student’s motion.
imaging, camera image processing, thermal imaging, and beam The first stage involves storing the student’s details, such as
counters affixed to entryways [3], WiFi-based approach [4], their allotted seat number, student registration number,
and RFID-based attendance system [5]-[9]. Each approach has scheduled classroom, and course timetable. Once the students
its own advantages and disadvantages across various are in their allotted seats, the motion sensor reading gets
dimensions such deployment, privacy, accuracy, cost, power, triggered. The sensor readings are then sent to the MQTT
communications, and operations. For example, using camera- broker and stored in the InfluxDB datastore. The data can then
based image processing is computationally expensive, cameras be sent to the institution's backend server. The architecture of
endanger privacy, thermal and ultrasound imaging and WiFi the IoT system is shown in Figure 1. The sensor data is collected
data have low accuracy. Therefore, understanding both by the first Pico W and is constantly published under the topic
challenges and benefits of various approaches in order to adopt "sensor_mo_data" using the MQTT protocol. In real time,
the most suitable methods is important for the real deployment multiple sensors will be employed to collect attendance from
of classroom occupancy monitoring system. multiple students i.e. desks. All the collected data will be
published under topics (one per IoT node) using the MQTT
This paper describes our experiences in adopting Internet of broker connected via gateway microprocessor.
Things (IoT) motion detection sensors to monitor the Temperature Sensor
attendance of lectures in courses at our university campus in
real time in a smart campus, process the data, present the IoT node
RasP. Pi 5
processed data in a graphical format for easy interpretation, and Rasp Pi Pico W
compare the obtained result with the expected result. To MQTT Broker
determine its accuracy, the data will be collected over a period WIFI
(2 x 4 hours) and then compared to the number of students Motion Sensor WIFI Data Analysis
enrolled in the classes. Section II. will describe the IoT system Gateway
design and architecture. Some experimental results will be
presented in Section III. Data Visualization
Humidity
Sensor Smartphone
II. IOT SYSTEM DESIGN AND ARCHITECTURE WIFI

A. IoT System Architecture

Web portal
The IoT based attendance system includes hardware
selection and setup, software design, integration of IoT required
components within the existing infrastructure, and testing the
Figure 1. System architecture of monitoring
proposed solution in real time. By integrating IoT components,
we have developed a smart automated attendance system. The The sensor data must be sent to a MQTT broker for distribution
aim of this paper is to propose an IoT-based solution that is to subscribers one of which is storing it in a database
more secure than the traditional RFID based system. The IoT (InfluxDB). As mentioned earlier, we have used the MQTT
solution is developed in a cost-effective way using the below protocol for communication as it allows publishing and
required components: subscription of topics over Wi-Fi. The MicroPython script
mentioned above publishes the collected data under
"sensor_mo_data". The MQTT (Eclipse Mosquitto) broker acts
as a middleman to exchange the messages. It is responsible for
streamlining the processing of incoming messages and sending
them to appropriate receivers (subscribers). The sensor data is
then transformed into a JSON string and marked with a
timestamp and sensor ID using a MicroPython script that we
created in order to standardise the data format.

We used the IoT Stack to build a Docker environment to install

and run the tools required for the project including the MQTT
broker (Mosquitto). The Docker environment is also used to
host a database to store the collected data for further processing
and visualisation. As mentioned earlier, we have used the
InfluxDB database as it adds a timestamp to every entry, which
is crucial when it comes to storing attendance data and provides
us with a timeline for attendance. The IoT system setup shown
in Figure 2. Includes handset (gateway), Rasp. Pi 5 (Docker),
Raspberry Pi Pico W IoT node. The data received can be
visualised for the students and the administrators to track their
attendance with ease. The Node-RED tool is used to integrate
all the nodes and connect MQTT to the database to data, also to
debug the output values. The Grafana tool is used to visualise
the output in user-interactive dashboards. The user can also
view the data graph of each individual motion sensor
measurement. The graphical interface of the tool can be
accessed by navigating to the Docker server’s IP along with the Figure 3. Python script
tool’s protocol and port number.
III. RESULTS
The micro PIR (Passive Infrared) motion sensors collect
data such as student’s motion in the classroom – more
specifically on their desk. The test was conducted in a simulated
environment, we collected data student allocated data for one
UG class and compared it with our simulated IoT results. We
used classroom-type simulations to conduct the study. The
number of students posed a challenge, as we required multiple
sensors and input nodes to record attendance at a time. To
eliminate the need for using multiple sensors for study, we
calculated the motion sensor data over a period of time (2 x 4
hours), calculated the average data collected, and then
compared it with the manually calculated attendance result. The
data is collected over tone UG course periods that are
approximately 4 x 60 minutes each.
Figure 2. System hardware
The motion sensor is used to record attendance. Each graph
represents real-time measurements from motion sensors
To make the controller with connected sensor functional, we without sacrificing time, student count, data readings. The
utilized the MicroPython language through the Thonny IDE, motion vs. time slots (9:00 – 13:00 and 14:00 - 18:00) graphs
and installed MQTT libraries in Thonny. The created code for for student 1 are shown in Figure 4a and 4b, respectively. Figure
Pico W is shown in Figure 3. 5a and 5b show the motion vs. time slots (9:00 – 13:00 and
14:00 - 18:00) graphs for student 2. The Grafana application
tool to provide data visualisation is used. The tool is used to plot
real-time data and allows users to switch between different
sensor readings. The tool allows the user to track participation
pattern in a course and usage rate over various time periods by
selecting the course dashboard tab. We saved the data we
received from Grafana as a csv file and turned it into IV. CONCLUSION
understandable graphs using the graph creation program, The monitoring of student attendance in a Smart Campus
created in Python. using the IoT Motion Detection Sensors, and the
implementation of IoT technology in educational institutions
has been presented in this paper. The IoT setup is composed of
IoT motion detection sensors, Raspberry Pi devices (Rasp. Pi 5)
and Rasp Pi Pico W. The IoT devices are embedded into the
existing campus environment for data to be collected,
transmitted through Wi-Fi using MQTT protocol and store data
in local server. The collected data is then accessible to the
management with real time insights and attendance pattern. A
one UG course in real time over a period of two time slots (2 x
4) hours (9:00 – 13:00 and 14:00 -18:00) was tested.

a)
ACKNOWLEDGMENT
This work was supported by a WIUT-UOW Research
collaboration funded project under grant RCF2022-003.

REFERENCES
[1] C. Harmer, “IoT: A study into the effectiveness of smart classrooms.
[online] Available at:
https://www.researchgate.net/publication/366274611_IoT_A_study_into
_the_effectiveness_of_smart_classrooms”, 2022.
[2] F. J. Perez Montalbo, “An IoT Smart Lighting System for University
Classrooms. International Symposium on Educational Technology
b)
(ISET). [online] doi:https://doi.org/10.1109/ISET49818.2020.00011”,
Figure 4. Student 1 Motion vs. Time slot a) 9:00 – 13:00 and
2020.
b) 14:00 to 18:00
[3] T. Sutjarittham, H. H. Gharakheili, S. S. Kanhere, and V. Sivaraman,
“Data-driven monitoring and optimization of classroom usage in a
smart campus,” in Proc. ACM/IEEE IPSN, Porto, Portugal, 2018, pp.
224–229.
[4] I. P. Mohottige and T. Moors, “Estimating room occupancy in a smart
campus using WiFi soft sensors,” in Proc. IEEE LCN, Chicago, IL,
USA, Oct. 2018, pp. 191–199.
[5] S. Mohite, G. Kulkarni, and R. Sutar,“RFID Security Issues.
International Journal of Engineering Research & Technology (IJERT),
ISSN: 2278-0181, Vol. 2 Issue 9, September - 2013.
[6] G. T. Bharathy, “Smart Attendance Monitoring System using IoT and
RFID. [online] 3(6). doi:https://doi.org/10.35629/5252-030613071313”,
2021.
a) [7] J. Dedy Irawan, “RFID and IOT for Attendance Monitoring System.
ResearchGate. [online]
doi:https://doi.org/10.1051/matecconf/201816401020”, 2018.
[8] S. Rajarshi, A. Ghosh, S. Sounak Dey, M. Das, and A. Chakrabarty, “IoT
& Cloud-based Smart Attendance Management System using RFID.
International Research Journal on Advanced Science Hub , [online]
5(03), pp.111–118. doi:https://doi.org/10.47392/irjash.2023.020”, 2023.
[9] P. Kuppusamy, “Smart Education Using Internet of Things Technology.
Advances in Data Mining and Database Management, [online] pp.385–
412. doi:https://doi.org/10.4018/978-1-5225-8446-9.ch017”, 2019.

b)
Figure 5. Student 2 Motion vs. Time slot a) 9:00 – 13:00 and b) 14:00 to 18:00