2024 Sixth International Conference on Computational Intelligence and Communication Technologies (CCICT)

Development and Execution of an IoT-Based

2024 Sixth International Conference on Computational Intelligence and Communication Technologies (CCICT) | 979-8-3503-7462-9/24/$31.00 ©2024 IEEE | DOI: 10.1109/CCICT62777.2024.00027

Accident Detection and Alert System with Location

Tracking
Prof. Somnath R. Wategonkar Dr. Kantilal P. Rane Prof. Surekha A. Khot
Assistant Professor at Bharati Associate Professor at Bharati Assistant Professor at A.C. Patil
Vidyapeeth College of Engineering, Vidyapeeth College of Engineering, College of Engineering, Kharghar,
Navi Mumbai Navi Mumbai Navi Mumbai
Mumbai University Mumbai University Mumbai University
Maharashtra, India. Maharashtra, India. Maharashtra, India.
[email protected] [email protected] [email protected]

Navin Ashok Yadav Saish Anand Vilankar Shyam Akashe

Bachelor’s in Bachelor’s in Professor ITM University,
Electronics & Telecommunications Electronics & Telecommunications Gwalior, India
Engineering Engineering [email protected]
Mumbai University Mumbai University
Maharashtra, India. Maharashtra, India.
[email protected] [email protected]

Abstract—Emergencies, by their very nature, are indicative of an accident. Upon detecting an accident, the
unpredictable and demand immediate action. Every second system’s GPS module pinpoints the vehicle’s exact
counts when lives are on the line, and rapid response is the location and transmits an alert to emergency responders,
cornerstone of effective emergency management. accompanied by the vehicle’s coordinates. This real-time
Unfortunately, around the world, countless lives are lost information enables emergency responders to swiftly
annually due to delays in rescue efforts. Automatic accident reach the accident scene, providing timely medical
detection and alert systems with live location tracking have assistance and potentially saving lives.
the potential to revolutionize road safety. By utilizing sensors
to detect accidents and GPS modules to pinpoint the The Implementation of automatic accident detection
vehicle’s location, these systems can rapidly notify and alert systems with live location tracking holds
emergency responders, minimizing response time and immense potential to transform road safety. By expediting
maximizing the chances of survival for accident victims. The emergency response times, these systems can significantly
benefits of these systems extend beyond saving lives, as they reduce the severity of injuries and fatalities, alleviating the
can also reduce property damage, healthcare costs, and burden on healthcare systems and reducing associated
provide peace of mind to drivers. While the cost of these costs. Additionally, these systems can instill peace of
systems can vary depending on their features and mind among drivers, knowing that timely assistance is
capabilities, the potential benefits far outweigh the readily available in the event of an accident.
investment.

Keywords—Accident detection, GPS Module, Sensors,

Blynk IoT, Road safety, emergency medical services, Vehicular
communication.

INTRODUCTION
Every year, 1.35 million lives are tragically lost and
20-50 million more are injured in car accidents globally,
often due to delays in response and secondary accidents
[1]. (Image of a car accident scene with emergency
vehicles arriving) In India alone, over 461,000 accidents
were recorded in 2022 [2]. Imagine a stadium filled with Figure 1. Type of collision in total accident in 2022
loved ones lost - that's the devastating human cost each
year. Beyond the immediate impact, delayed response As Figure 1 shows every 3 out of 2 accidents
strains healthcare systems and inflicts a heavy economic happened due to a collision between two vehicles or to a
burden. The development of automatic accident detection fixed object therefore the goal is to design the device
and alert systems with live location tracking has emerged considering these situations in mind [2].
as a promising solution. These innovative systems harness This paper presents the development and deployment
the power of sensors and GPS technology to promptly of an Automatic Vehicle Accident Detection and Alert
identify accidents and notify emergency responders, System using IoT, employing a TTGO TCALL SIM800L
thereby minimizing response time and potentially saving module from LILYGO as shown in Figure 2. The system
lives. is relatively straightforward to deploy and adaptable for
By embedding sensors within vehicles, these systems use across diverse vehicles. In numerous instances of
can continuously monitor vehicle dynamics and detect vehicular accidents, victims may find it challenging to
sudden changes in speed, acceleration, impact, and seek assistance actively. In such scenarios, a strategic plan
is essential, involving the prompt notification of relevant

979-8-3503-7462-9/24/$31.00 ©2024 IEEE 91

DOI 10.1109/CCICT62777.2024.00027
Authorized licensed use limited to: INSTITUTE OF ENGINEERING & MANAGEMENT TRUST. Downloaded on November 23,2024 at 10:24:52 UTC from IEEE Xplore. Restrictions apply.
 authorities and the swift provision of medical aid to those into vehicles are constructed using Raspberry Pi (RPi).
affected. Conventional accident detection methods often These RPi devices incorporate sensors for immediate
rely on manual reporting, leading to potential delays or accident detection and a camera to assess the severity of
inaccuracies. Automatic vehicle accident detection and the incident. The SVMs are utilized as an image
alert systems utilizing IoT present a promising resolution classification model grounded in machine learning.
to this challenge. These systems leverage various sensors,
including accelerometers, gyroscopes, and GPS modules, In a separate investigation, Praharsha Sarma et al.
for real-time accident detection. Upon detecting an (2020) [5] proposed Automatic vehicle accident detection
accident, the system can autonomously dispatch alerts to and alert systems utilizing IoT. This system relies on an
emergency services and/or notify the vehicle owner or ultrasonic sensor for obstacle detection and incorporates a
designated contacts, concurrently disabling the engine vibration sensor with an accelerometer to identify any
ignition. unusual movements. When an anomaly is detected, the
system utilizes an Arduino UNO to send a message
through the GSM module.
These studies demonstrate the feasibility of
developing automatic vehicle accident detection and alert
systems utilizing IoT. Such systems hold the capacity to
significantly decrease the occurrences of fatalities and
injuries resulting from vehicle accidents. Although
advancements have been achieved, there exist several
challenges that necessitate attention in the development
and deployment of IoT-based automatic vehicle accident
detection and alert systems.
PROPOSED APPROACH
Figure 2. Block Diagram of Project
This section delineates the methodology employed in
the proposed system which consistently observes sensor
data during the vehicle's operational state. Upon detecting
an accident, the system initiates the alert mechanism and
transmits a message to emergency responders. The
message encompasses the vehicle's location, accident
particulars, and timestamp.
As depicted in Figure 3, when an accident occurs, the
GPS module retrieves the precise location of the vehicle.
Subsequently, these coordinates are transmitted via the
GSM modem from the device to the emergency responder
or designated contact person, who can also monitor the
location via desktop/Laptop and also from their
Smartphone with the help of the Blynk App.
Upon receiving the alert message, emergency
Figure 3. Communication during an accident responders can send rescue services to the accident scene.
As shown below in Figure 4 the user can also track the
Literature Survey vehicle at will and also driver can send an SOS to the
user/operator [6].
Several studies have been conducted on the
development and implementation of Automatic vehicle
accident detection and alert systems utilizing IoT. For
example, a study by Nevonprojects (2023) [3] proposed
Automatic vehicle accident detection and alert systems
utilizing IoT using a Node MCU ESP8266
microcontroller. The system uses an Accelerometer,
gyroscope, tilt sensor, and GPS module to detect
accidents. Upon detection of an accident, the system
issues an alert message to the Blynk cloud platform. The
user can then view the alert message on their mobile app
and/or take appropriate action, such as calling Emergency
services.
Kumar Reddy Mallidi and V.V. Vineela (2018) [4]
introduced a smart Vehicle Monitoring System based on
IoT, aiming to identify accidents and prevent vehicle Figure 4. Communication between Device/Driver and User/Operator
theft. The system employs an IoT-driven support vector
machine (SVM) algorithm, and IoT devices integrated

92

Authorized licensed use limited to: INSTITUTE OF ENGINEERING & MANAGEMENT TRUST. Downloaded on November 23,2024 at 10:24:52 UTC from IEEE Xplore. Restrictions apply.
 Since one of the constraints was that the device needs
to be compact and small in size therefore TTGO TCALL
SIM800L LILYGO is used considering its small size and
less power requirement with great performance.
b) Sensors: Various sensors are used like 1)
Accelerometer and gyroscope: MPU6050 sensor module
for detecting sudden changes in speed, direction, and
vehicle orientation. 2) Impact sensors: Strategically placed
piezoelectric sensors to detect direct collisions, 3) GPS
Module: UBLOX NEO-6M or similar module for
accurate and reliable location data transmission, 4)
Cellular Modem: SIM800L or equivalent modem with
appropriate cellular network compatibility for alert
transmission.
c) Sensor Data Acquisition: Sensor data is
continuously collected at designated sampling rates. Data
pre-processing involves filtering noise and calibrating
sensor outputs for accuracy. Time synchronization across
sensors ensures accurate accident analysis.
d) Algorithm Development: Machine learning
algorithms trained on real-world accident data to identify
patterns indicative of various accident types. Supervised
learning methods such as Support Vector Machines
(SVM) or Random Forests demonstrate elevated accuracy
in accident classification. Algorithm optimization focuses
on minimizing false positives and maximizing true
Figure 5. Logic for accident verification positives for reliable accident detection as shown below
(Figure 5). Supervised learning methods such as Support
Hardware Selection Vector Machines (SVM) or Random Forests demonstrate
a) Microcontroller: The core unit, capable of elevated accuracy in accident classification. Algorithm
handling sensor data, running algorithms, and triggering optimization focuses on minimizing false positives and
alerts. Options include Arduino Mega, Raspberry Pi Pico, maximizing true positives for reliable accident detection
TTGO TCALL SIM800L LILYGO, or STM32 as shown below (Figure 4).
microcontroller. e) Communication Protocols: Standardized protocols
like NMEA 0183 for GPS data and AT commands are
used for cellular modem communication. Development of
a custom alert message format containing vehicle
location, accident type, timestamp, and additional relevant
details. Encryption and authentication mechanisms ensure
secure data transmission and prevent unauthorized access
as shown in next Figure 5. Alert messages also include the
specified data and voice clip messages to multiple
numbers
f) System Validation and Testing: Rigorous testing in
simulated and real-world environments is used to evaluate
accuracy, reliability, and performance under various
conditions.Collaboration with emergency responders for
feedback and refinement of alert messages and response
protocols [7]. Compliance testing with relevant
regulations and ethical considerations regarding data
privacy and security.
g) System Deployment and Scalability: Development
of user-friendly interfaces for vehicle integration and
potential driver interaction. The primary steps are
developed under User Interface (UI) (Figure 7)
Figure 6. Communication logic of the system
development.

93

Authorized licensed use limited to: INSTITUTE OF ENGINEERING & MANAGEMENT TRUST. Downloaded on November 23,2024 at 10:24:52 UTC from IEEE Xplore. Restrictions apply.
 TABLE 1. PREVIOUS WORK INTRODUCED IN THE FIELD

Title Year Problem Solution Limitation

The task proposed methodology only
Trauma Accident Detection and An accident detection mechanism is introduced handles falling scenarios. Also, they do
2018
Reporting System [8] to reduce the rate of fall-related deaths not handle any traffic accident-related
scenarios.
The development involves a smartphone
Smartphone-Based Accident The system is less accurate and gives
application that utilizes Android sensors for
Detection and Rescue System with 2015 multiple false results.
accident detection and offers real-time tracking
Real-Time Location Tracking [9]
of the accident location.
An automatic accident detection and alarm
Automatic Accident Detection and system which uses user profiling to identify There is no emergency response system.
2017
Alarm System [10] whether the user is distracted or not and then Limited to distracted walkers
takes appropriate action
Traffic Accident Detection System
An intelligent traffic system that utilizes edge The smartphone utilizes existing
with Intelligent Features Using 2018
computing to detect accidents using smartphones. datasets and is prone to errors.
Mobile Edge Computing [11]

1) User Research and Analysis: Conducted surveys a) Precision: Ensuring the system can precisely
and interviews with drivers to understand their identify accidents with minimal instances of false
preferences. Recorded the data. positives and false negatives is crucial.
2) Define User Requirements: Based on research b) Dependability: The system must exhibit
findings, user personas are created such as "Busy reliability and functionality across diverse environmental
Commuter" or "Tech-Savvy Driver" to represent different conditions.
user groups. Requirements like minimal distraction, easy 1) Modular design and open-source software to
access to essential functions, and compatibility with facilitate customization and adaptation for different
smartphones are defined. vehicle types and regions.
3) Conceptual Design: A dashboard display with 2) Cost-effective hardware selection and collaboration
large, easily readable icons for common functions like with manufacturers for wider implementation and
navigation, music, and climate control is sketched as an affordability.
initial interface concept.
3) Miniature battery backup is incorporated to provide
4) Wire-framing and Prototyping: Developed low- power even after the damage of the main battery or
fidelity prototypes using tools like Sketch. Prototypes are damage to the wiring system. This led to the
tested with drivers to gather feedback. Certain controls development of an emergency efficient system.
discovered are too small to operate easily while driving.
4) As all the hardware used is as per the standard of
5) Visual Design: The brand's color scheme and IoT 4.0, the size of the IoT processor and the
typography across the interface are used to create high- sensors are compact. Its operating voltage is 3.3v
fidelity mockups with visually appealing graphics and instead of 5v, and the whole system is designed
intuitive layouts by ensuring consistency with the with less power for its operation and data
vehicle's brand. transmission. Therefore, power consumption is less
6) Interaction Design: Interaction patterns are and the system can be fitted into a small size
designed such as swiping gestures for navigation through cabinet so that it may be hidden at any corner of a
menu options or tapping for selecting items. Clear vehicle.
feedback is provided for user actions, like visual or 5) Data captured from the accident is important to be
auditory cues when a button is pressed. preserved and is performed using HASH coding
These investigations illustrate the viability of creating [13]. Data to be stored is coded and whenever
and deploying Automatic Vehicle Accident Detection and required, it is required to be decoded with the
Alert Systems Utilizing IoT [12] Top of Form. These public and private key of HASH coding. But, as the
systems hold the promise of substantially diminishing the data of all the accidents captured is important, it is
occurrences of fatalities and injuries resulting from sent to the Indian emergency response helpline
vehicle accidents. system ethically.
While notable strides have been taken, there remain This proposed methodology provides a comprehensive
several challenges to confront in the advancement and framework for developing and deploying the automatic
execution of Automatic Vehicle Accident Detection and accident detection and alert system. By combining
Alert Systems Utilizing IoT. robust hardware, advanced algorithms, and secure
communication protocols, this system holds immense
The measuring parameters considered in the system
potential to revolutionize road safety and save lives.
are Precision, Dependability, and Affordability and are
discussed below.

94

Authorized licensed use limited to: INSTITUTE OF ENGINEERING & MANAGEMENT TRUST. Downloaded on November 23,2024 at 10:24:52 UTC from IEEE Xplore. Restrictions apply.
 TABLE 2. TOTAL NUMBER OF ACCIDENT FROM 2018 TO 2022

Year Accident % change over Fatalities % change over Person Injured % change over
previous period previous period previous year

2018 4,70,403 0.2 157593 5.1 4,64,715 -0.6

2019 4,56,959 -2.9 1,58,984 0.9 4,99,360 -3.3

2020 3,72,181 -18.6 1,38,383 -13.0 3,46,747 -22.8

2021 4,12,432 10.8 1,53,972 11.3 3,84,448 10.9

2022 4,61,312 11.9 1,68,491 9.4 4,43,366 15.3

Figure 7. Blynk Website Dashboard & Mobile dashboard

Figure 8. Type of Road Accident 2022

Figure 9. Alert message received to the emergency Responder

95

Authorized licensed use limited to: INSTITUTE OF ENGINEERING & MANAGEMENT TRUST. Downloaded on November 23,2024 at 10:24:52 UTC from IEEE Xplore. Restrictions apply.
 c) Affordability: The system must remain cost-
effective, catering to a broad spectrum of vehicles. User-
Friendliness: The system should be straightforward to
install and operate [14].
The Automatic Vehicle Accident Detection and Alert
Systems Utilizing IoT. Top of Form holds significant
potential for substantially mitigating the toll of deaths and
injuries resulting from vehicle accidents. Nonetheless,
there persist challenges in the development and execution
of these systems. By tackling these challenges and
Figure 10. Comparison of Standard Commercial Solution and enhancing the systems in the aforementioned aspects, we
Our Solution can establish Automatic Vehicle Accident Detection and
Alert Systems that contribute to enhanced road safety for
all.
Figure 7 shows the Dashboard of the Blynk website
and for Smartphones on which the user can track and see
the necessary details and data of the vehicle such as speed,
longitude, and latitude.
Figure 8 shows that the majority of accidents that
happen are fatal accidents that can cause the death of the
driver, the driver can survive if medical help is given to
them in time because stopping excessive blood loss is a
major factor in their survival [15].

Figure 11. Comparison of Standard Commercial Solution and II. RESULT

Our Solution Accuracy
As illustrated in Figure 9, when an accident occurs, an
alert message containing the longitude and latitude will be
dispatched to the mobile number of the emergency
responder.
A. Test Environment
Hardware: Arduino Mega 2560, MPU6050 sensor
module, impact sensors, UBLOX NEO-6M GPS module,
SIM800L cellular modem.
Software: Custom-developed C++ code, Machine.
B. Emergency Response Improvement
Simulated response time is reduced by 40-50%
compared to traditional methods. Real-world testing
demonstrated faster dispatch and arrival of emergency
Figure 12. Comparison of Standard Commercial Solution and responders. Positive feedback from emergency personnel
Our Solution for false
regarding accurate location and accident details.
Existing emergency helpline numbers were defined by
the Government of India (Table 3). In the proposed
system, provision is made to add multiple helpline
numbers calling using voice clip messaging for the
integration with an existing system.
TABLE 3: INDIAN RESPONSE HELPLINE NUMBERS
National Emergency Number 112
Police 100 or 112
Fire 101
Ambulance 102
Disaster Management Services 108

C. User Feedback
Participants in the testing phase reported increased
Figure 13. Comparison of different solution for various
peace of mind knowing about the real-time alert system.
Drivers felt a sense of safety and potential reduction in
anxiety due to the system's capabilities Figure 10 shows
the comparison of the standard commercial system and
our work in which we tested our work multiple times.

96

Authorized licensed use limited to: INSTITUTE OF ENGINEERING & MANAGEMENT TRUST. Downloaded on November 23,2024 at 10:24:52 UTC from IEEE Xplore. Restrictions apply.
 Figure 11 shows the comparison between Standard Conf. Automat., Qual. Test., Robot. (AQTR), pp. 1–5, May
Commercial Solution and Our Solution Accuracy. Where 2018.
[9] G. R. Shete, D. A. Shah, A. R. Gaidhani, S. D. Shinde, and S.
our solution has better results than the commercial
R. Sharma, ‘‘Smartphone based vehicle tracking and accident
solution. Figure 12 shows the comparison of the standard prevention system,” Int Journal of Computer Applications
commercial system with our work in which the false Information Technology, vol. 8, p. 172, 2015.
Reporting has been compared. Figure 13 shows the [10] Z. Wei, S.-W. Lo, Y. Liang, T. Li, J. Shen, and R. H. Deng,
comparison of the customized system commercial system ‘‘Automatic accident detection and alarm system,’’ in Proc.
and the proposed System the comparison is made based 23rd ACM Int. Conf. Multimedia, pp. 781–784, 2015.
on the cost accuracy and false Reporting. [11] C. Liao, G. Shou, Y. Liu, Y. Hu, and Z. Guo, ‘‘Intelligent
traffic accident detection system based on mobile edge
Overall the results indicate the project's success in computing,’’ in Proc. 3rd IEEE Int. Conf. Comput. Commun.
developing a reliable and efficient automatic accident (ICCC), pp. 2110–2115, December 2017.
detection and alert system with live location tracking. The [12] D. A. Whitney and J. J. Pisano TASC, Inc., Reading,
Massachusetts, "AutoAlert: Automated Acoustic Detection of
high accuracy, minimal latency, and positive impacts on Incidents", IDEA project, [Online]. Accessed on 05 November
emergency response and user experience demonstrate the 2023, Available: http://pubsindex.trb.org/view.aspxid= 481489
potential of this technology to revolutionize road safety. [13] F. E. De Guzman, B. D. Gerardo and R. P. Medina,
"Implementation of Enhanced Secure Hash Algorithm Towards
III. CONCLUSION a Secured Web Portal," IEEE 4th International Conference on
Automatic accident detection and alert systems with Computer and Communication Systems (ICCCS), Singapore,
pp. 189-192, 2019, doi: 10.1109/CCOMS.2019.8821763.
live location tracking hold immense potential to transform [14] Dr. C. K. Gomathy, K Rohan, Bandi Mani Kiran Reddy, Dr. V
road safety by revolutionizing emergency response and Geetha, “Accident Detection and Alert System,”, Journal of
reducing the devastating consequences of vehicle engineering, computing, and architecture, Vol. 12, Issue 3,
accidents. By harnessing the power of sensors, GPS March 2022.
technology, and cellular communication, these systems [15] R Elvik, P. Christensen, A. Amundsen, “Speed and road
can promptly identify accidents, provide real-time accidents: an evaluation of the Power Model,” TOI Report,
information to emergency responders, and optimize [Online], Accessed on 04 December 2023, Available:
http://www.trg.dk/elvik/740-2004.pdf.
resource allocation, leading to faster response times,
reduced fatalities and injuries, and improved outcomes for
accident victims. The implementation of these systems
has the potential to save lives, minimize injuries, reduce
property damage, increase peace of mind among drivers,
optimize emergency services, and alleviate the burden on
healthcare systems. As technology continues to advance,
the capabilities and effectiveness of these systems are
expected to further improve, making them an
indispensable tool for enhancing road safety worldwide
REFERENCES
[1] World Health Organization. (2013). Info-graphics on Global
Road Safety 2013. Accessed: Jan. 05, 2024.
http://www.pbs.gov.pk/content/traffic-accidents-annual,
Accessed on 02 December 2023.
[2] Ministry of Road Transport and Highways
https://morth.nic.in/sites/default/files/RA_2022_30_Oct.pdf,
Accessed on 02 December 2023.
[3] Nevonprojects, IoT-Based Automatic Vehicle Accident
Detection and Alert System. Retrieved from
https://nevonprojects.com/iot-based-automatic-vehicle-accident-
detection-and-rescue-system, Accessed on 02 December 2023.
[4] Kumar Reddy Mallidi, S., & V.V. Vineela, K. “Intelligent
vehicle monitoring system based on IoT.” International Journal
of Engineering and Technology (UAE), 7, 4.19, 110-113, 2018.
[5] Praharsha Sarma, S., Sreedhar, J., & Pavan, K., “IoT-Based
Automatic Vehicle Accident Detection and Messaging System,”
International Journal of Innovative Research in Electrical and
Electronics Engineering, 9(3), 1-5, 2020.
[6] L. Chuan-zhi, H. Ru-fu, Y.E. Hong-wu, “Method of Freeway
Incident Detection Using Wireless Positioning,” in Proceedings
of the IEEE International Conference on Automation and
Logistics, pp. 2801 – 2804, 2010.
[7] C. Thompson, J. White, B. Dougherty, A. Albright, and D. C.
Schmidt, "Using smart-phones to Detect Car Accidents and
Provide Situational Awareness to Emergency Responders," in
3rd International ICST Conference on MOBILe wireless
Middleware, Operating Systems, and Applications (Mobileware
2010), 2010.
[8] A. F. Pop, A. Puscasiu, S. Folea, and H. Vălean, ‘‘Trauma
accident detection and reporting system,’’ in Proc. IEEE Int.

97

Authorized licensed use limited to: INSTITUTE OF ENGINEERING & MANAGEMENT TRUST. Downloaded on November 23,2024 at 10:24:52 UTC from IEEE Xplore. Restrictions apply.