International Journal of Scientific & Engineering Research, Volume 7, Issue 4, April-2016 71

ISSN 2229-5518

A survey on Road Accident Detection and

Reporting
1
Amrutha Madhusan, 2Lavanya Viswanathan, 3Vaishnavi Ravindran,4Dr.Shanta Rangaswamy
E-Mail: [email protected], [email protected], [email protected],
1
4
[email protected]
Department of Computer Science and Engineering,
R.V.College of Engineering, Bengaluru-560059

ABSTRACT: On road accident is a major issue of concern. Even with all modern developments in the field of vehicle
design, road lane design and management, accidents do occur. Timely accident detection and taking immediate action
with respect to emergency health care of victims by informing an emergency center such as a hospital or a police station
about the accident on time plays a vital role inhuman safety androad traffic management. Accident detection can be done
under various domains. Most of the papers surveyed use application of sensor technology, besides trying to detect
accidents automatically using machine learning and computer vision from surveillance systems. Any kind of accident
detected is automatically sent as an alert to the required destination. Each of these methods has different percentages of
accuracy and their own limitations. This survey paper discusses various approaches to detect the occurrence of traffic
accidents on a road supervised under surveillance camera.

Keywords: Accident monitoring, Accident detection, Accelerometer, GSM, GPS

——————————  ——————————

IJSER
Introduction: vehicles. The incidents of accidental deaths
have shown increasing trend during the year
According to the Global status report on road
2000-2015 with an increase of 50 percent in the
safety 2015 [1], the total number of deaths
year 2010 as compared to the year 2000.
caused due to road accidents has levelled out at
According to Planning Commission of India, the
1.25 million a year. India faces the highest
total annual economic loss is 2.5% of India’s
number of accidents and accidental fatalities in
GDP due to rising number of road fatalities.
the world. In India, there are many kind of
According to National Crime Records Bureau,
places like hilly area plateaus, and due to
Ministry of Road Transport & Highway, Law
improper road facilities accidents are more and
commission of India, one serious road accident
death rate due to this accidents are more [29].
in the country occurs every minute and 16 die
The maximum number of accidents are reported
on Indian roads every hour. According to the
in the transport sector, that is, on road as well as
“WHO Report 2015: Data Tables”[2] the total
railways. Some approximations claim that
number of fatalities in India in 2013 is
Indian roads alone accounted for approximately
238,562 and reported number of road traffic deaths is
105,000 accidental fatalities in 2010. This is
137,572 with the estimated road traffic death rate per
almost 15 percent of the global road fatalities 100,000 population being 16.6. The leading cause of
when India has just 1% of the total global

IJSER © 2016
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 7, Issue 4, April-2016 72
ISSN 2229-5518

death is road traffic injuries especially among young Electro Mechanical System (MEMS) sensor and
people and it costs countries about 1-3% of the gross vibration sensor. MEMS sensor also helps in
domestic product [3].One of the major factors that is measuring the angle of roll over of the car. A
increasing this number is the delay in reporting of
32-bit ARM controller is used as the main high
the crashes to emergency centers like near-by police
speed data-processing unit. The vibrations are
stations and health care centers. According to [1],
sent from the vibrating sensor to the controller
in 2015 the 2030 Agenda for Sustainable
after passing through an amplifying circuit.
development was launched which aims at
Similarly the roll over angle is sent from the
reducing the number of deaths and injuries
MEMS sensor to the controller.
arising due to road crashed to half its number by
the year 2020. In regard to this, authors provide Advantages and results

a review on the existing technologies that aim to Both the accident and the accident location can
detect accidents automatically and alert the be detected as opposed to only one in the other

IJSER
emergency centers without much delay. approaches. There is also a method to stop
sending the alert message and hence save time
Literature Review:
of the rescue time. The use of GPS adds to the
The proposed system in [4] deals with an
advantage of the system being cost-effective,
automatic accident detection system involving
portable and detecting the accurate location [6]
vehicles which sends information about the
and the time taken for the entire detection
accident including the location, the time and
process and sending of the message is greatly
angle of the accident to a rescue team like a first
reduced as compared to other methods. Overall
aid center and the police station. This
the system is portable, has a small size, and is of
information is sent in the form of an alert
low cost and expandable.
message. But in the cases where there are no
casualties a switch is provided which can be In [7], traffic accidents have been detected
turned off by the driver to terminate sending the successfully using an Internet of Things (IoT)
alert message. A GSM module is used to send and cloud computing framework. Traffic
the alert message and a GPS module is used to accidents have been detected using Support
detect the location of the accident. The GPS and Vector Machine (SVM) that hasbeen improved
GSM module are interfaced to the control unit with the Ant Colony Algorithm (ACA). ACA is
using serial communication [5].The accident used for the parameter selection of SVM which
itself is detected using two sensors- Micro plays an important role in the accuracy that can

IJSER © 2016
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 7, Issue 4, April-2016 73
ISSN 2229-5518

be achieved by SVM. The IoT sensors used here The proposed system in [9] aims at reducing the
for monitoring the vehicles are the highly loss of lives due to traffic accidents and
sensitive magneto resistive sensors. In fact performs three main tasks – (1) detecting an
multiple sensor modules are employed to detect accident and sending the location to the nearest
the presence of vehicles including sensors that hospital, (2) controlling traffic light signals in
detect thechanges in the magnetic field on the the route taken by the ambulance [10] [11] and
road,the sound signals from brakes and (3) monitoring vital parameters of the patient
collisions and two different sensors that help inside the ambulance itself and sending this
detect the direction of the vehicles.SVM is information to the hospital. These three tasks are
trained with historical traffic information and achieved by the working of four units into
tested on future traffic data. The algorithm tries which the system is divided:
to find a decision plane that separates the class The vehicle unit: This unit consists of a
of ‘traffic accident’ from the class of “no traffic microcontroller, sensors, GPS, GSM module

IJSER
accident’. This is improved by using ACA and an accelerometer. The sensors detect the
which is an optimization algorithm [8]. accident, the GPS gets the location and the
GSM module conveys this information to the
Results
main server unit. The accelerometer can help
The metrics that have been used to detect the avoid accidents by notifying the driver when the
efficiency of detection include: (1) False Alarm position of the vehicle is deviated from the
Rate (FAR): ratio of error alerts to all detected normal. The entire vehicle unit must be installed
events,(2) Detection Rate (DR): ratio of in the vehicle.
detected events to the real world accidents and The control unit: This unit contains the database
(3) Average Time to Detect (ATD): time of hospitals and is responsible for
average between detected and happened. communicating messages between all the units.

Results show that the IoT based method The ambulance unit: This unit has a patient

outperforms the traditional sensor based method monitoring system to constantly measure and

and the improvised SVM model with ACA convey the patient’s temperature and pulse rate

outperforms the traditional SVM model and also to the hospital.

has a faster convergence speed and the Mean The traffic junction unit: This unit turns the

Square Error (MSE) is lower compared to the signal to green when the ambulance is about 10

traditional SVM. meters away so that the path is clear for it to

IJSER © 2016
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 7, Issue 4, April-2016 74
ISSN 2229-5518

move quickly. This is achieved through RF The proposed system in [15] also uses an
communication. Thus, this system has overcome Android application where the smartphone must
many drawbacks of the existing accident be placed in a holder attached to the vehicle.
detection systems with respect to time. The Accident Detection Algorithm detects an
accident based on three kinds of events: (1) A
The proposed system in [12] is in the form of an collision is detected if the accelerometer shows
Android application which detects an accident a reading above the threshold which is 4g
using an accelerometer which is built in the (g=9.8m/s2) and the approximate severity of the
smartphone. The phone must be docked inside accident is determined by a metric called
the vehicle and not held by any person. The Acceleration Severity Index (ASI). (2)
working of this application is as follows: Rollovers are detected using a gyroscope and a
When the device is tilted above a certain magnetometer. If a rotation greater than 45
threshold and is detected by the accelerometer, degrees occurs and if the instantaneous speed is

IJSER
the application waits for 15 seconds. Here, three found to be less than 5km/h, it is considered as a
kinds of input can be received. (1) If the user is rollover. (3) And airbag deployed signal
active, he can press “cancel” if the device was indicates an accident as well.
tilted by mistake. (2) If the user is active, he can
If one of these three events is detected, an alert
press “send” if an accident has occurred. (3) If
is sent through three different sources - a
the user is inactive and no button is pressed after
Decentralized Environmental Notification
15 seconds, an accident is assumed to have
Message (DENM) message containing a Road
occurred.
Hazard Warning, by performing an e-Call to an
In case of (2) and (3), the current location is Emergency Medical System (EMS) and finally
fetched by GPS and a pre-recorded voice by SMS.
message along with the location is sent to the
This system controls false positives by sending
108 ambulance emergency response service. A
accident notifications only if a countdown
study on GPS services provided by Android has
sequence is not interrupted by any of the
been thoroughly conducted [13] [14]. Thus,
passengers.
through the use of just a smartphone without
any extra hardware components, efficient The proposed system in [16] consists of several

accident detection and notification has been components with different functions. First, the

achieved. vehicles should have embedded in it an On-

IJSER © 2016
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 7, Issue 4, April-2016 75
ISSN 2229-5518

Board Unit (OBU) responsible for detecting periodically is integrated using the rectangle
accidents and communicating information about rule and with the help of old data, and by setting
dangerous situations. The notification of the threshold, they are classified as accidents.
detected accidents is made through a
The internal structure of CU unit receives
combination of both Vehicle-to-Vehicle (V2V)
notification from the vehicle; it classifies the
communications and Vehicle and the Roadside
accident based on the severity and decides on
Infrastructure (V2I) communications, while the
the needed resources for the particular event.
destination of the information is the Control
The various actions include storing the accident
Unit (CU) which will handle the alert
information in the database, notifying the police
notification, estimating the severity of the
and emergency services with the help of
accident and communicating the incident to the
information in database.
appropriate emergency services which is done
through the internet service. If the internet is The accident severity is estimated using Data

IJSER
unable to be used, then an infrared module Mining classification algorithms like Bayesian

through RSU (Road side Units) can connect the Network and K2 algorithm which is run on

OBU with the CU. The layouts and street Weka platform for historical data from The

structure should be exploited for the National Highway Traffic Safety Administration

dissemination of the protocol. The makeup of which maintains the General Estimates System

every modules are as follows. The OBU [17] (GES), a database with information about traffic

consists of in-vehicle sensors to detect accidents that began its operation in 1988 in the

accidents, a data acquisition unit DAU unit to US. Based on number of attributes like time,

periodic data and an OBU processing unit to causalities and damage, the classification is

process and find the occurrence of an accident done.The improvements in this work is using

for which it makes use of an Accident Detection this system at high speed

Algorithm. The system in [19] detects an accident by

The two impactful events in an accident are Roll utilizing GPS and a Map Matching (MM)

over events and strong impacts. Roll over is algorithm. The GPS is used to send the position

captured by sudden strong horizontal tilt. and speed of the vehicle every 0.1 second to a

Applusand IDIADA [18] provides a lot of Micro Controller Unit (MCU) which compares

historical data, thus information on collision. the current speed to the previous speed every

The average tilt from the current vehicle taken o.1 seconds and if the speed has fallen below a

IJSER © 2016
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 7, Issue 4, April-2016 76
ISSN 2229-5518

threshold an accident is said to have occurred before sending an alarm out to the emergency
and it then checks for the location of the vehicle center. In these 5 seconds the driver can press a
using the map matching module and sends an button to cancel the alarm and help reduce the
alarm to an emergency center is the vehicle is number of false alarms. The speed, location,
found outside the road network. The location of time and the contact details of a relative of the
the accident is sent using GSM. vehicle occupant will be sent as GPRS data
using the GPRS module of the MCU and in
The braking distance is dependent on speed and
cases where there is no GPRS coverage an SMS
is proportional to its square and hence the faster
is sent with the same information. After this, a
a vehicle is going the more bleak are the
voice call is initiated to the emergency center if
chances of avoiding a crash [20]. The forces that
the occupant is in a position to describe the
help in bringing the vehicle to a stand-still after
accident.
the brakes are applied are frictional force and

IJSER
gravitational force. The maximum speed that Advantages
would be achieved after applying the break and
The system uses a GPS module which is now
when only these two forces are in play is
present in almost all vehicles and is also cheap.
calculated. If the speed sent by the GSM to the
The system also allows the emergency center to
microcontroller is however lesser than this
gauge the severity of the accident by sending the
maximum speed it is inferred as an accident in
previous speeds of the vehicle and moreover
which additional forces have played a role in
allows the vehicle occupant to manually turn off
decreasing the speed. These additional forces
the alarm and hence reduce false alarms.
come from the kinetic energy that is associated
with a body while in motion that gets converted The technologies used in [22] project is GPS,

into destructive forces during the collision [21] GPRS and Android. The main modules and their
functionalities include the hardware module for
The MM algorithm is used to find the location
accident detection which is used to limit the
of the vehicle on the road network. A
switches and Arduino. Then the Web based
Geographical Information System (GIS)
Central Registration System registers these
software is used for MM using which the
services. An Android Application is used for
vehicle is continuously positioned on a digital
controlling the circuit and User Interaction in
map. Finally, when an accident is detected a flag
the vehicle will raise alarm and ask the user for
is raised and the MCU waits for 5 seconds
the conduction of further outcomes [23]. A Web

IJSER © 2016
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 7, Issue 4, April-2016 77
ISSN 2229-5518

based Service for the Central Assistance Centre are integrated by the Kalman filter [26].
will receive notifications. Whenever the deceleration from the GPS is
available, IMU deceleration is updated by this
Working of the module includes registering of a
filter. The deceleration data from the IMU is
new vehicle and downloading the android app in
considered only in the GPS outage scenario. The
smartphone. Registering all the services such as
Razor INS can fill the gaps in case of GPS
ambulance and insurances services [24].When
outage and the IMU data are updated through
the vehicle collides with another object or
the integration of valid GPS data. The IMU
vehicle, the limit switch gets pressed and it
provides accelerometer, gyroscope and
establishes a circuit and the signal passes to the
magnetometer reading in all three axes. The
Arduino.The Arduino is connected to the limit
orientation of the vehicle relative to gravity can
switches and the Bluetooth module which is
be predicted by integrating gyroscope with the
further connected to the driver’s Android Smart
accelerometer and the direction of the vehicle

IJSER
Phone.The application determines the GPS co-
can be determined from the magnetometer. By
ordinates and transfers them to the central server
integrating these sensors in three axes, the
from where the location is sent as an alert to all
Attitude Heading Reference System
services registered.
(AHRS)[27] is built. The Direction Cosine
The different technologies used include GPS, Matrix (DCM) is used as a basis to integrate the
accelerometers and Kalman filters. The major accelerometer, gyroscope and magnetometer.
need for this algorithm aremore accuracy in This matrix is a form to represent a rotation
GPS systems as they suffer from the line of through a 3X3 R matrix. The DCM matrix
sight and poor update rate and as such, the described in [28] has been used as a model for
acceleration data derived from the GPS lacks in the IMU sensors to build the AHRS. The ARHS
instantaneous acceleration which is very data gets corrupted, which is fused with GPS
important in determining a sudden deceleration data using the Kalman filter. Vehicle decelerates
due to accident. High and instantaneous data can when the brake is applied. Any deceleration
be acquired from the IMU (Inertial more than 5Gs as obtained is considered as an
Measurement Unit) as opposed to GPS. accident situation by the proposed algorithm. In

The proposed system in [25] monitors the this situation, the system would raise an alarm

deceleration data from the GPS and IMU for the location detection module.The results

accelerometer sensors. These two types of data from the paper are based on the threshold

IJSER © 2016
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 7, Issue 4, April-2016 78
ISSN 2229-5518

maintained which is 5Gs, which when the http://www.who.int/violence_injury_preve

deceleration is below it, an alert is sent. ntion/road_safety_status/2015/GSRRS201
5_data/en/. (Accessed: 22- Mar- 2016).
CONCLUSION
[3] M. Ruikar, "National statistics of road
The papers provide various methods to detect
traffic accidents in India", Journal of
accidents using both hardwareand software
Orthopedics, Traumatology and
methodswhich provide good results. Most of the
Rehabilitation, vol. 6, no. 1, p. 1, 2013.
discussed methods also provide the driver with
the option of turning of the alarm in cases where [4] V. Goud, "Vehicle Accident Automatic
the accident is not serious or false detections of Detection and Remote Alarm
an accident.These methods are either mostly Device", International Journal of
dependent on some hardware like sensors that Reconfigurable and Embedded Systems
have to be present in the car or require a smart (IJRES), Vol. 1, no. 2, 2012.

IJSER
phone to be present within the car. While the
[5] GSM User Manual, SIMCOM Ltd, August
use of such hardware can prove to be a more
2006.
cost-efficient approach it has the drawback of
[6] GPS Beginner`s Guide, Garmin Ltd, 2008.
being destroyed in the accident and hence
[7] G. Liang, "Automatic Traffic Accident
giving spurious or no readings at all. Hence, an
Detection Based on the Internet of Things
approach that does not depend on any hardware
and Support Vector Machine", IJSH, Vol.
device or sensor that is associated with the car is
9, no. 4, pp. 97-106, 2015.
required for the detection of traffic accidents.
[8] S.L. Lin and Z. Liu, “Parameter selections
REFERENCES
in SVM with RBF kernel function”,
[1] "Global status report on road safety Journal-Zhejiang University of
2015", World Health Organization, 2016. Technology, vol. 35, no. 2, pp. 163, 2007
[Online]. Available:
[9] B. Prachi, D. Kasturi and C. Priyanka,
http://www.who.int/violence_injury_preve
"Intelligent Accident-Detection And
ntion/road_safety_status/2015/en/.
Ambulance- Rescue
(Accessed: 22- Mar- 2016)
System", International Journal Of
[2] "Data", World Health Organization, 2016. Scientific & Technology Research, Vol. 3,
[Online]. Available: No. 6, 2016.

IJSER © 2016
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 7, Issue 4, April-2016 79
ISSN 2229-5518

[10] Mr.S.Iyyappan and Mr.V.Nandagopal , Android Operating System”, IEEE

“Accident Detection and Ambulance International Conference on System,
Rescue with Intelligent Traffic Light Man and Cybernetics, San Antonio, pp.
System?”, International Journal of 2059-2061, 2009.
Advanced Technology and Engineering
[15] B. Fernandes, M. Alam, V. Gomes, J.
Research, 2013.
Ferreira and A. Oliveira, "Automatic

[11] AmneshGoel, Sukanya Ray, Nidhi accident detection with multi-modal alert

Chandra, “Intelligent Traffic Light system implementation for

System to Prioritized Emergency ITS", Vehicular Communications, Vol.

Purpose Vehicles based on Wireless 3, pp. 1-11, 2016.

Sensor Network”, International Journal [16] Pranav Dhole, Saba Shaikh, NishadGite,
of Computer Applications, Volume 40– Vijay Sonawane “Automatic Accident

IJSER
No.12, February 2012. Detection: Assistance Through

[12] K. Patel, "Utilizing the Emergence of Communication Technologies and

Android Smartphones for Public Welfare Vehicles”, IJETST- Vol. 02, Issue 04

by Providing Advance Accident Pages 2285-2288 April 2015

Detection and Remedy by 108 [17] OSGi Alliance, ‘‘OSGi service

Ambulances", International Journal of platform,’’ Core Specification, Release
Engineering Research & Technology 4, Version 4.1. San Ramon, CA: IOS
(IJERT), Vol. 2, No. 9, 2013. Press, Inc., 2007.

[13] Xinhua Shu, Zhenjun Du and Rong [18] Applus+ IDIADA Web Site. (2011, June
Chen, “Research on Mobile Location 9). IDIADA: Instituto de Investigacio ´n
Service Design Based on Android”, 5th AplicadadelAutomo ´vil. [Online].
International Conference On Wireless Available: http://www.idiada.es
Communication Networking and Mobile [19] M. Amin, M. SobhanBhuiyan, M.
Computing, Beijing, pp. 1-4, 2009. IbneReaz and S. Nasir, "GPS and Map
Matching Based Vehicle Accident
[14] J. Whipple, W. Arensman and M.S
Detection System", in 2013 IEEE
Boler, “A Public Safety Application of
Student Conference on Research and
GPS-Enabled Smart Phones And The

IJSER © 2016
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 7, Issue 4, April-2016 80
ISSN 2229-5518

Development (SCOReD), [24] Ohn Whipple., William Arensman.,

Putrajaya,Malaysia, 2013. Marian \Starr Boler. ,"A Public Safety
Application of GPS-Enabled
[20] A. Chandra, S. Jain, and M. A. Qadeer,
Smartphones and the Android Operating
"Implementation of location awareness
System", Proceedings of IEEE
and sharing system based on GPS and
International Conference on Systems,
GPRS using J2ME, PHP and MYSQL,"
2009.
in 3rd International Conference on
Computer Research and Development [25] M. Amin, M. Bin IbneReaz and S.
(ICCRD), pp. 216-220, 2011. Sheikh Nasir, "Integrated Vehicle
[21] M. S. Amin, J. Jalil, and M. B. I. Reaz, Accident Detection and Location
"Accident detection and reporting System", TELKOMNIKA
system using GPS, GPRS and GSM (Telecommunication Computing

IJSER
technology," in International Conference Electronics and Control), Vol. 12, no. 1,
on Informatics, Electronics & Vision pp. 73, 2014.
(ICIEV), Dhaka, Bangladesh, 2012, pp.
[26] Mahony R, Hamel T, Pflimlin JM.
640-643.
Nonlinear complementary filters on the
[22] P. Dhole, S. Shaikh, N. Gite and V.
special orthogonal group. IEEE
Sonawane, "Smart Accident Detection &
Transactions on Automatic Control.
Assistance System",International Journal
53(5): 1203-1218, 2008.
of Emerging Trends in Science and
Technology, Vol. 2, no. 4, pp. 2285-
[27] Building an AHRS/Head-tracker using
2288, 2015.
the "9DOF Razor IMU" or the "9DOF
Sensor Stick" by SparkFun. Available
[23] Chin Feng Lai, Chen Ying Liu, Sung-
online: https://dev.qu.tu-
Yen Chang and Yueh Min Huang,
berlin.de/projects/sf-razor-9dof-
Department of Engineering Science,
ahrs/wiki/Tutorial Accessed on 20
"Portable Automatic Conjecturing and
October 2013.
Announcing System for Real-Time
Accident Detection" National Chen
[28] Zaldivar J, Calafate CT, Cano JC,
Kung University, Tainan, Taiwan.
Manzoni P. Providing accident detection
in vehicular networks through OBD-II

IJSER © 2016
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 7, Issue 4, April-2016 81
ISSN 2229-5518

devices and Android-based smartphones.

Proceeding of Local Computer Networks
(LCN),pp. 813-819, 2011.

[29] Ramya Keerthi, G.Shanmukh, Dr. R.

Sivaram, Various Accident Detection
Technologies and Recovery Systems with
Victim Analysis, International Journal of
Advanced Trends in Computer Science
and Engineering (IJATCSE), Vol.2 ,
No. pp: 07-12, Special Issue of
ICCSIE 2013 .

IJSER

IJSER © 2016
http://www.ijser.org