Proceedings of the Third International Conference on Smart Systems and Inventive Technology (ICSSIT 2020)

IEEE Xplore Part Number: CFP20P17-ART; ISBN: 978-1-7281-5821-1

A Novel IoT based Accident Detection and Rescue

System
Pranto Karmokar1 , Saikot Bairagi1 , Anuprova Mondal1 , Fernaz Narin Nur2 , Nazmun Nessa Moon1 , Asif Karim3 , Kheng
Cher Yeo 3*
1
Department of Computer Science and Engineering, Daffodil International University, Bangladesh
2
Department of Computer Science and Engineering, Notre Dame University, Bangladesh
3
College of Engineering, IT and Environment, Charles Darwin University, NT, Australia
1 1
{pranto15-6002, bairagi15-5642, mondal15-5992}@diu.edu.bd, moon@daffodilvarsity.edu.bd,
2
fernazcse@ndub.edu.bd, 3 {asif.karim, charles.yeo*}@cdu.edu.au

Abstract—In S outh-East Asian cities such as Delhi, Dhaka due to legal complexit ies, people do not want to involve in
road accidents are a very common occurrence which brings any kind of rescue operation post-accident, resulting in tragic
disaster to human lives as well as infrastructures. S ometimes consequences for the victims. So met imes bystanders do not
people cannot reach hospitals prompt after an accident even have enough time to help that victim. All these
because of the traffic jam, deficit of ambulance, lack of a bottlenecks obstruct the quick transfer of the victim to the
mechanism to timely propagate information to the appropriate nearby hospital.
authority. To ensure the safety of lives, this paper proposes an
automated IoT based effective accident detection system. The acquaintances are often notified long after the
Immediately after an incident, the data information is sent to incident. The overall situation is worsened by the fact that
the webserver, instant S MS is forwarded to the victim’s the tracing and reachability of the accident spot in a
acquaintances and also to the relevant authorities such as metropolitan area is a substantially difficult task. Thus in this
traffic control room, nearby police station, ambulance service. paper, an automatic accident detection and rescue system is
To evaluate the performance of the system, a simulated road focused on using IoT by which the kin of the victim will be
scenario has been designed. The result obtained after a notified on time. Real-time notification to the nearest police
thorough integration and system testing demonstrates that the station and traffic control server will also be transmitted.
proposed system not only achieves the stated goal of the Nearest ambulance service will also be alerted through the
research but also can deliver the expected outcome in a rather
internet with the location of the accident. The complete
cost-effective way.
system will significantly improve the prompt co-ordination
Keywords—IoT, load cell GPS, GSM, Wi-Fi module, rescue of the necessary actions after an accident. In the field of IoT,
system. the objects communicate and exchange information to
provide advanced intelligent services to the users.
I. INT RODUCT ION The rest of this paper is organized as follows. Section
Owing to the intensive development in technology, II reviews some of the related works in this field. The
various intelligent and autonomous systems have been proposed methodology is introduced in Section III. The
invented in recent years [2-6]. No wadays quite dependent on system design and implementations are presented in Section
technologies which operate as an automated or semi- IV. In Section V, the results and finding are discussed.
automated way to make things easier and comfortable for us Finally, Section VI presents the concluding remarks and
[6-10]. future direction of this research.
In South Asian cities, the number of vehicles is on a II. RELAT ED ST UDY
constant rise, consequently, road accidents are increasing
alarmingly. The accidents main ly occur for rush driving and Khan et al. [2] described a system for accident detection
overtaking. In 2017, the number of death by road accidents using a three-axis accelerometer sensor to the cloud server.
in Bangladesh reached 20,607 (2.62% of total death). In The framework auto matically initiates the process of
Bangladesh, even though it will soon be recognized as one of dispatching the nearest ambulance by processing the Global
the developing countries, both economically and socially Positioning System (GPS) coordinates of the incident and
[17], the death rate is 15.56 per 100,000 of population. The providing a specific route to the accident spot in question. By
total number of registered motor vehicles in Bangladesh is using the related android application, ambulance driver can
3.4 million as of March 2018. The nu mber of drivers is 3.8 reach the spot in a rather efficient manner. Saga et al. [13]
million, 60% of which are illiterate [11]. Since numerous designed a framework where soon as the accident occurs, the
smart systems aiming to make provide us with a better life airbag will open and auto-lock the breaks. Then a buzzer will
has been developed [1-10], there should be a technological be switched on. The system incorporates a GPS module,
solution in cases related to an accident. through which a microcontroller determines the coordinates
of the location.
No latest technology can avert accidents completely and
do not even know where and when an accident may occur. In Then the GSM (Global System for Mobile
its current state, most of the time the life of the victim Co mmunication) module sends this accident information to
depends upon the responses of people nearby. Oftentimes the victim’s family for urgent attention. Khan et al. [14]
proposed a smart rescue system based on an Android

978-1-7281-5821-1/20/$31.00 ©2020 IEEE 322

Authorized licensed use limited to: Middlesex University. Downloaded on November 11,2020 at 07:41:53 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of the Third International Conference on Smart Systems and Inventive Technology (ICSSIT 2020)
IEEE Xplore Part Number: CFP20P17-ART; ISBN: 978-1-7281-5821-1

application. Under this system, the user will be automatically

monitored at all times. However, the user can turn the system
off if the need be. This system can detect any kinds of a jerk
and generate an alert, a false alert can be cancelled by the
driver. If the driver does not cancel this alert within 15
seconds, then it is assumed that a substantial accident has
taken place and victim’s location is sent via an API to the
nearest emergency responder so that emergency rescue
operations can be initiated. Ramya et al. [15] described an
intelligent traffic light system (ITLS) where an ambulance
can move easily without traffic light stopping it on the way.
Through the GPRS (General Packet Radio Serv ice, a packet-
oriented mobile data standard on the 2G and 3G cellular
communication networks) 3G modem, the traffic controller
will get the location of the ambulance. If the ambulance is
near to the traffic junction, then the corresponding signal will
be green. Automated systems in traffic detection are also on
the rise [27].
Fernandez et al. [16] put forward a system of accident
detection that uses GPS, GPRM C and M CU through a speed
monitoring algorithm. GPRMC is the most common
sentence transmitted by almost all GPS devices, this sentence
or line of data contains nearly everything a GPS application
requires. If an accident occurs then the place of accident,
speed before accident, time and date will be readily available
through SMS using GPRS.
Ali and Eid [25] have considered the number of factors
such as sudden alteration in acceleration, rotation and an
impact force at the rear into their automatic accident
detection system, the system uses fuzzy logic to finalize a
decision and attained an accuracy of 98.67%. However, it
only works if the accident occurs from behind.
Another recent proposition by Yee et al. [26] uses GPS
and Accelerometer-based accident detection system. The Fig. 1. System flowchart of the proposed scheme.
system not only detects accidents but also provide the degree
of severity of the crash. A limitation of the system is that it IV. SYST EM DESIGN AND IMPLEMENT AT ION
can only detect a service provider within 10 KM of the In this section, our practical approach is going to be
radius. Also, it is not clear how the system calculates many
discussed that has been taken to solve the problem. In the
injured onboard passengers.
system, a sensor named load sensor has been used which
III. PROPOSED M ET HODOLOGY detects the mass of the car and finds changes in weight.
In Fig. 1, the flo wchart for the system is illustrated, When an accident occurs, the detection will start by the
which starts with the sensor check collision. Then it Load Cell. Load Cell sensor is used to capture dynamic
evaluates whether the accident has occurred. If it has not, pressure and impact factor exerted on an accident. A fter any
then it does not execute any operation. But in case an type of accident, the sensor sends the data. Consequently,
accident does occur, then the sensor determines the location the authority will know the state of the accident. The
and sends it to the microcontroller. The microcontroller will
authorized entity will then be able to track the driver of the
direct an SMS to a previously stored number belonging to
the kin of the victim. The SMS will have full details of the ambulance and send offline SMS. The SMS contains the
location. location of an accident and the driver will receive necessary
directions to go that spot. On the receiving end, the user
Fig.2 v isualizes the entire framework. In the proposed
whose number is stored in the database will get the SM S
methodology, accident detection will be detected with a
collision sensor. At first, when an accident occurs, the with the information on the accident. They can see the place
collision sensor detects the collision. After the detection, the of accident and use alternative way if necessary. Wi-Fi
GPS sensor sends the accident location to a microcontroller. module will provide the necessary internet connectivity
Following which the microcontroller read numbers fro m facility.
database server via WiFi module. At last, the
microcontroller sends SMS to the Smartphone to transmit
the information on the accident.
As can be seen from Fig.1, the proposed accident
detection system consists of some sensors, microcontroller
and WiFi module in a single development board.

978-1-7281-5821-1/20/$31.00 ©2020 IEEE 323

Authorized licensed use limited to: Middlesex University. Downloaded on November 11,2020 at 07:41:53 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of the Third International Conference on Smart Systems and Inventive Technology (ICSSIT 2020)
IEEE Xplore Part Number: CFP20P17-ART; ISBN: 978-1-7281-5821-1

3) Traffic light – The traffic light is attached in the road

scenery which is located in a common point where several
roads are connected.
4) Collision sensor – Many collision sensors are attached
in the road. The sensors check whether any accident has
occurred or not.

Fig. 2. Proposed system architecture.

A. System Design
This study was focused on creating a smart rescue system
to instantly facilitate post-accident co-ordination. An
experimental prototype is designed to better demonstrate the
usefulness of the overall framework. Location coordinates of
a vehicle are manually measured using GPS sensor and this
location data is stored in the database through a
microcontroller. These coordinates consist of latitude and
longitude value; it will pinpoint the vehicle’s position in the
area. In Fig.3, the proposed IoT based prototype has been
demonstrated where a road scenario can be seen and all Fig. 4. A smart car with a load cell and GPS module.
modules connected to detect any incident. The materials are
chosen for the prototype based on their availability, cost and
performance. Hence, for the real-time application, these
materials might be replaced with more rigid materials. A
roadside scenario is made to test the prototype application
where the following devices and materials have been used-

Fig. 5. Collision Detection using Load Cell.

2
In Fig.4, the car will be equipped with few Load Cells in
3 front and rear. There will be a GPS sensor inside the car.
This GPS sensor is linked with the Load Cells through which
4 the accident will be detected and identification and
pinpointing of the accident location will take place instantly .
Every moment is important after a crash takes place.
B. Implementation of the Proposed System
In this section, Fig.5, Fig.6 and Fig.7 represent the
overall imp lementation and demonstration for the proposed
system. In Fig.5, Load Cells are used to detect the collision.
Fig. 3. Proposed rescue plan.
When any vehicles push with reasonable force fro m the back
or front of the car, then the Load Cells identify it as a
1) Smart car – A smart car is designed with Load cells collision. A Load Cell is a type of force measuring
and GPS to detect an accident. instrument that gauges the amount of pressure exerted on it.
2) Road scenery – A road scenery with roadside It has a transducer that generates an electrical signal whose
buildings is made for the prototype. magnitude is directly proportional to the force being applied
[20].

978-1-7281-5821-1/20/$31.00 ©2020 IEEE 324

Authorized licensed use limited to: Middlesex University. Downloaded on November 11,2020 at 07:41:53 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of the Third International Conference on Smart Systems and Inventive Technology (ICSSIT 2020)
IEEE Xplore Part Number: CFP20P17-ART; ISBN: 978-1-7281-5821-1

In Fig.6, the SMS sub-system of the framework has been TABLE I. I NTEGRATION T EST RESULT
shown. The SMS is sent via the GSM module to the number
Test Case Expected Observed Test
that is already stored in the database. The message will
contain detailed information on the accident location. When Result Result Result
the system collecting the stored contact numbers of users, the When the accident
occurred the Load Can detect
system will send SMS of accident location link to the users Can detect
Cell should be able an accident Pass
by GSM Module. GSM (Global System for Mobile to detect the accident.
Co mmunication) is an architecture used for mobile accident
communication in almost all of the countries now a day [21]. GPS module of this
system should be Location
In Fig.7, a map is visualized to the user. When a user Location is
able to detect should Pass
clicks on the link of the point of accident, the location will be vehicle location be exact.
exact.
displayed on google map. By knowing the location, any user correctly
or the rescue team can dynamically take the shortest route to GSM M odule of
reach the destination, an ambulance will also head towards the system should SM S SM S has
the accident location to provide emergency medical will be been Pass
be able to send
assistance. sent. sent
SM S
The
V. RESULT AND DISCUSSION microcontroller
The result and testing part is divided into two sections should be able to Can send
which represent the uniqueness of the framework and send data to a and
Can send
accuracy of the results. server using the retrieve Pass
Can retrieve
Wi-Fi module. The data to data
A. Integration Test microcontroller and from
Integration testing is a type of testing where individual should be able to the server
retrieve data from
modules are comb ined and tested as a group. It occurs after the server
the unit testing phase. Purpose of integrated testing is to
expose faults within the interaction between integrated units
[28]. Table I shows test cases, expected results and observed B. System Test
results for each module of the system. Fro m the results
acquire fro m Tab le-I all the modules can be seen working System testing of software or hardware is conducted in
successfully in the system. a complete and integrated environment to evaluate its
compliance with the specified requirements. System testing
takes all integrated modules that have passed integrated
testing as its input. The testing then aims to detect any
inconsistencies between the units integrated [29]. System
testing can detect bugs with the interaction between different
modules of a framework. Oftentimes, these interactions
glitches may not have a huge impact on a modular level, but
when the impact is measured on the overall performance of
the system, such minor modular level impacts may become
more serious than anticipated initially [22]. System testing
helps us to identify any such inconsistencies at an early stage
of the development. Table II. Shows test cases, expected
results and observed results of system testing.

TABLE II. SYSTEM T EST RESULT

Test Case Expected Observed Test

Result Result Result
Users should be
able to get SM S SM S
SM S should
from this received Pass
arrive
system
successfully
Accident Accident
User should see
location location
accident Pass
should be
location can be seen
seen
Ambulance
should be able Ambulance Gets the
to get the should get required Pass
direction to proper notification
reach the notification
accident spots
Authority Data in the
Can update
should be able server can Pass
Fig. 6. Sending SMS after an accident. server data
to update server

978-1-7281-5821-1/20/$31.00 ©2020 IEEE 325

Authorized licensed use limited to: Middlesex University. Downloaded on November 11,2020 at 07:41:53 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of the Third International Conference on Smart Systems and Inventive Technology (ICSSIT 2020)
IEEE Xplore Part Number: CFP20P17-ART; ISBN: 978-1-7281-5821-1

data be updated Innovation in Information Systems (ICRIIS), Langkawi, 2017, pp. 1 -

6.
VI. COMPARISON BET WEEN EXIST ING SYST EMS AND [6] H. U. Zaman, J. Khisha, N. Zerin and M. H. Jamal, "Speech
PROPOSED FRAMEWORK responsive mobile robot for transporting objects of different weight
categories," 2017 18th International Conference on Advanced
In this portion, in Table III, the difference between Robotics (ICAR), Hong Kong, 2017, pp. 395-400.
features available in some tradit ional systems and our [7] A. H. Siddique, B. Barkat and M. Poshtan, "Smart electrical
proposed IoT based smart system is tabularized by which the protection method for industries operations," 2013 IEEE Electric Ship
advancement of our research work can be demonstrated. T echnologies Symposium (ESTS), Arlington, VA, 2013, pp. 394-398.
[8] M. Hasan, M. H. Anik and S. Islam, "Microcontroller Based Smart
TABLE III. P ROP OSED FRAMEWORK I N COMP ARISON TO EXISTING Home System with Enhanced Appliance Switching Capacity," 2018
SYSTEMS Fifth HCT Information Technology Trends (ITT), Dubai, United Arab
Emirates, 2018, pp. 364-367.
Test Case Existing System Proposed System [9] S. Chowdhury, F. R. Wasee, M. S. Islam and H. U. Zaman, "Bengali
Handwriting Recognition and Conversion to Editable Text," 2018
M ost other system Second International Conference on Advances in Electronics,
Detection do not use the Load Sensors Computers and Communications (ICAECC), Bangalore, 2018, pp. 1 -
M echanism functionality of have been applied. 6.
Load Sensors [10] M. A. H. Sakib, F. Islam, S. S. Haque and H. U. Zaman, "Doctor
Providing locator: A web application to improve online doctor directories in
instant Current systems do Our system Bangladesh features and functionalities of a better online solution to
direction to not have this provides this find a doctor," 2016 5th International Conference on Informatics,
the accident feature feature Electronics and Vision (ICIEV), Dhaka, 2016, pp. 100-104.
spot [11] M. Hasan, M. H. Anik, S. Chowdhury, S. A. Chowdhury, T. I. Bilash,
In-expensive GSM S. Islam, “Low-Cost Appliance Switching Circuit for Discarding
and WiFi modules Technical Issues of Microcontroller Controlled Smart Home,”
M ost of the have been International Journal of Sensors and Sensor Networks, vol. 7, no. 2,
prevalent pp. 16-22, 2019.
implemented,
Cost frameworks use giving a [12] Daily Sun (2018), ”Bangladesh: Road Traffic Accident”, Retrieved
effectiveness expensive from: https://www.daily-
significant edge in sun.com/printversion/details/327901/2018/08/08/Bangladesh:-Road-
mechanism such as
terms of T raffic-Accidents.
Bluetooth
deployment [13] Khan, A.R., Suri, P., Tejaswani, S.P., Paigude, T.A. “Automatic
expenses. Accident Detection”. International Journal for Engineering Science
and computer-, Volume 8, Issue March 2018.
[14] Bhoite, P.A., Gopal, K., Saga, W., Sisodiya, T ., Satish, P. “ Accident
VII. CONCLUSION Detection System Using Arduino”, International Science and
technology journal- Volume 7, Issue 4, 2018.
The current study proposes and implements a smart
[15] Khan, A., Bibi, F., Dilshad, M., Ahmed, S., Ali, Z.H. “ Accident
accident detection and rescue system for busy metropolitan Detection and Smart Rescue System using
cities. This system ensures smart accident detection and
[16] K. Sangeetha, P. Archana, M. Ramya, P. Ramya. IOSR Journal of
rescue mechanism that can provide the exact location of the Engineering (IOSRJEN), Vol. 04, Issue 02(February. 2014)|| V5|| PP
accident. It also aids any ambulance to get direction. 53-57.
Ambulance driver can take the shortest path to the accident [17] S. G. Fernandez, R. Palanisamy, K. Vijayakumar “GPS & GSM
spot and thereby rescue the victims in the least possible Based Accident Detection And Auto Intimation”, Indonesian Journal
timeframe. Besides, our system puts a minimu m of financial of Electrical Engineering and Computer Science. Vol.11, No.1, July
strain on the user to implement than many other systems, but 2018, pp 336-361.
with a more reliable output. In future machine learning [18] Daily Tribute (2018)” Developing Bangladesh : A new identity ”,
Retrieved from:
techniques [24] can be implemented to get even better http://www.dhakatribune.com/bangladesh/development/2018/03/10/d
insights into the situation in an automated manner. evelopi ng-bangladesh-new-identity/.
[19] Learn about IntegrationT esting, Retrieved from:
REFERENCES https://en.wikipedia.org/wiki/Integration_testing.
[1] Hasnat G.N.T ., Kabir M.A., Hossain M.A. (2018) Major [20] Learn about test ing, Retrieved from:
Environmental Issues and Problems of South Asia, Particularly https://en.wikipedia.org/wiki/System_testing
Bangladesh. In: Hussain C. (eds) Handbook of Environmental [21] Makoto, M. and Harada, S. “Research of Protection Equipment for
Materials Management. Springer, Cham the Load Cell Receiving a Large Impact Load.” 2006 SICE-ICASE
[2] M. Hasan, P. Biswas, M. T. I. Bilash and M. A. Z. Dipto, "Smart International Joint Conference, 2006, doi:10.1109/sice.2006.314738.
Home Systems: Overview and Comparative Analysis," 2018 Fourth [22] M. Y. (2013). Wireless Facsimile Connectivity Using Gsm Module.
International Conference on Research in Computational Intelligence International Journal of Research in Engineering and Technology,
and Communication Networks (ICRCICN), Kolkata, India, 2018, pp. 02(05), 836-840. doi:10.15623/ijret.2013.0205018
264-268.
[23] Alkhalid, A. (2018). Bridging the Gap Between GUI Functional
[3] A. H. Siddique, B. Barkat and M. Poshtan, "Experimental setup of System T esting and Functional System Logic Testing. Carleton
wide area monitoring using Zigbee IEEE 802.15.4 technology and RF University Research Virtual Environment Search Whole Repository.
FM technique," World Congress on Sustainable Technologies doi:10.22215/etd/2018-13306.
(WCST -2012), London, 2012, pp. 44-48.
[24] Saleh, A.J., Karim, A., Shanmugam, B., Azam, S., Kannoorpatti, K.,
[4] M. Hasan, A. Z. Dipto, M. S. Islam, A. Sorwar, S. Islam, “A Smart Jonkman, M., Boer, F.D. An Intelligent Spam Detection Model Based
Semi-Automated Multifarious Surveillance Bot for Outdoor Security on Artificial Immune System. Information 2019, 10, 209.
Using Thermal Image Processing,” Advances in Networks, vol. 7, no.
2, pp. 21-28, 2019. [25] A. Karim, S. Azam, B. Shanmugam, K. Kannoorpatti, and M. Alazab,
“A Comprehensive Survey for Intelligent Spam Email Detection,”
[5] S. Paul, P. D. Nath, N. M. Abdus Sattar and H. U. Zaman, "rTraffic - IEEE Access, vol. 7, pp. 168261–168295, 2019.
a realtime web application for traffic status update in the streets of
Bangladesh," 2017 International Conference on Research and [26] Ali, A., & Eid, M. (2015). An automated system for Accident
Detection. 2015 IEEE International Instrumentation and Measurement

978-1-7281-5821-1/20/$31.00 ©2020 IEEE 326

Authorized licensed use limited to: Middlesex University. Downloaded on November 11,2020 at 07:41:53 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of the Third International Conference on Smart Systems and Inventive Technology (ICSSIT 2020)
IEEE Xplore Part Number: CFP20P17-ART; ISBN: 978-1-7281-5821-1

Technology Conference (I2MT C) Proceedings. doi: Vehicles Detection To Maintain Traffic By Image Processing”,
10.1109/i2mtc.2015.7151519 International Journal of Scientific & T echnology Research, 2020,
[27] Yee, T. H., & Lau, P. Y. (2018). Mobile vehicle crash detection 9(2), pp. 2921-2928.
system. 2018 International Workshop on Advanced Image
T echnology (IWAIT). doi: 10.1109/iwait.2018.8369671.
[28] F. M. J. M. Shamrat, I. Mahmud, A. K. M. S. Rahman, A. Majumder,
Z. Tasnim, N. Islam, “ A Smart Automated System Model For

978-1-7281-5821-1/20/$31.00 ©2020 IEEE 327

Authorized licensed use limited to: Middlesex University. Downloaded on November 11,2020 at 07:41:53 UTC from IEEE Xplore. Restrictions apply.