Content uploaded by Swati Raut
Author content
All content in this area was uploaded by Swati Raut on May 15, 2015
Content may be subject to copyright.
Survey on Vehicle Collision Prediction in VANET
Swati B. Raut
Dept. of Computer Science & Engg,
G. H. Raisoni College of Engineering
Nagpur (M.S), India
surai.raut5@gmail.com
Dr L.G.Malik
Dept. of Computer Science & Engg
G. H. Raisoni College of Engineering
Nagpur (M.S), India
latesh.malik@raisoni.net
Abstract—Number of vehicles are increasing day by day,
adding to the existing traffic and so are the challenges concerning
it. Vehicle collision is one of such issues, which needs to be
addressed at priority. Many schemes have been put forth
recently, to detect vehicle collision probabilities, consequently
avoiding collisions and its adverse effects. These schemes in
general focus on the issues such as critical response time for
alerts, reliable data transfer, efficient broadcasting, security etc.
In this paper we have consolidated such recent and niche
techniques and its unique features. The survey done provides
different approaches for detecting collision and the parameters
considered for same. Discussion is also made on various
simulators used for simulation thereby providing details of
different scenarios and factors involved in VANET.
Keywords—Broadcasting; Collision Probabilities; Reliable
Data Transfer; Simulation Parameters; Vehicle collision;
Vehicular ad hoc networks (VANET)
I. INTRODUCTION
With the continuous increase in population over the globe,
problems concerning the same are increasing. Among these
problems, traffic related issues also need a major attention.
Despite of the advances in automobile and revised traffic
rules, it’s very crucial to align the things when it comes to
traffic on roads. The facts and figures coming from different
geographies have same updates when it comes to road
accidents. Every year, the numbers of traffic crashes are
increasing and the lives of millions people are cut short. In
addition to this millions of people are suffering from non-fatal
injuries [1]. In recent years, light vehicle crashes accounted for
majority of all motor vehicle fatalities. Furthermore, out of
these light vehicle fatalities, a certain percentage was from
side impacts, suggesting crashes at intersections or on
roadways close to and leading to intersection [2].
In view of above illustrated scenarios, prevention of
transportation accidents event ahead becomes an open issue
with an alarming vehicular traffic on highways, cities, and
urban areas. Different approaches for safety on road include
traffic monitoring and channeling, which work in accord with
technologies such as alert systems, digital maps, etc. On
similar note, collision warning systems are becoming an
important part of vehicle active safety. The accidents can be
prevented with VANET technology that uses inter-vehicular
communication (IVC). It can be incorporated with Intelligent
Transportation System (ITS) which works with the help of
Telematics technology. Telematics technology [3] solves the
issue of communication as; it supports the exchange of real-
time traffic information among vehicles. Typical telematics
techniques include: IVCs, VANETs, Car to Infrastructure
(C2I), Car to Car (C2C) [4], etc. Vehicle-to-vehicle (V2V)
and Vehicle-to-Infrastructure (V2X) communication has
gained an immense importance towards solving the problem
of vehicle collisions. VANET provides a platform for
implementing various techniques for improving road safety [5,
6]. For this reason, V2V and V2X communication is
considered an important part of future ITS. IVC concepts are
also used in crash mitigation, crash avoidance and Intersection
Collision Warning Systems which are ITSs applications [7].
In this paper survey on recent schemes which focus on
vehicle collision avoidance system is provided. Discussion is
also made on various factors involved in collision avoidance
system and the way they are derived. Simulation parameters
considered in VANET are also part of interest .In addition to
this, a summary of different schemes is also mentioned. The
paper is partitioned into following sections as follows : The
first section provides a basic introduction regarding the paper.
The next section involves related work done on collision
warning system. The third section describes factors invoved in
collision avoidance system. The fourth section describes the
design of vehicle collision prediction system. The fifth section
gives general simulation parameters. Finally sixth section
concludes the paper.
II. RELATED WORK
Stefan et al [2] comment on collision avoidance system at
intersections as ITS application. Two issues have normally
been overlooked in past. First issue is the networking-related
metrics, which do not expose the quality of ITS-based
solutions. Second, the gap between application requirements
and networking concepts, as it needs to be terminated. To
address above mention issue, a scheme is developed to
determined vehicle collision probability at intersections by
means of Inter-vehicular communication concepts. In addition
to this, a novel scheme is proposed to quantify future crash
probability at intersection, depending on the situation in which
a vehicle receives beacon message known as Basic Safety
Message (BSM). Presented scheme makes use of Intelligent
Driver Model (IDM), the car-following model. Safety metric
is validated by carrying out extensive simulation in veins
simulator .Veins couples the road traffic simulator-SUMO and
978-1-4799-3972-5/14/$31.00 ©2014 IEEE
1098
the network simulator OMNeT++. Typical suburban X-
intersection is considered in simulation, in which geometry
lane is imported from OpenStreetMap.
Ali et al [8], in his research, suggested Collisions Warning
Systems to reduce the causalities arising from road traffic and
accidents. These Systems are engineered by taking the driver
reaction times into consideration to send the warning and also
solution to address the problem of false alarm is proposed by
estimating the distribution of brake response times for a
particular driver based on VANET data. Brake response times
of individual drivers which can be used in collision warning
algorithms to reduce false alarm rates are also investigated.
Because of this, system will become more reliable and honest
for drivers. Safety applications could potentially take full
advantage of being customized to an individual’s
characteristics.
Chetankumar et al [9] incorporated VANET technology to
solve the problem of the communication between vehicles.
The accidents can be avoided if proper information can be
communicated to the vehicles which are probable to collide in
stipulated time frame. The proposed system used estimation
model to predict collision events. Estimation model performs
two functions, first, the estimation of parameters and secondly,
forecasting of potential risk. Kinematic model in [10] is used
to estimate vehicle parameters. The Kinematic model is
embedded in all vehicles of VANET and estimates the
parameters of the subject vehicle along with its neighbors.
Estimated parameters are used in advance to generate alerts in
case potential risk. Data estimated by estimation model takes
as input to the forecasting module to calculate potential risk,
which is integrated with communication model. This model
can be used to estimate the real time parameters in all types of
network [10] [9]. Communication in VANET is carried out by
Dedicated Short Range Communication (DSRC) protocol to
broadcast the alert message. The communication channels
provided by DSRC are one-way or two-way short-range to
medium-range wireless communication designed specifically
for automotive. It also provides a corresponding set of
protocols and standards. The overview of various Collision
prediction schemes is as given in Table I.
III. FACTORS INVOLVED IN COLLISION WARNING SYSTEM
This section discusses basic factors involved in Vehicle
Collision Probability Calculation. The factors such as
A. Vehicle Information
Vehicle Information involves vehicle speed, acceleration,
heading, latitude and longitude position .This information can
be obtained from V2V communication or by V2R
communication.
B. Trajectory Prediction
All possible behavior of driver is calculated with the help
of vehicle information .The trajectory is determined with the
help of mathematical formulae. Mathematical formulae
changes as per the scenario.
C. Collision Probability Calculation
Probability of collision is calculated by considering some
threshold value, which is set according to user choice.
D. Alert Messages
Alert messages will be broadcast, multicast and unicast
according to the situation detected by the system. System
should possess property of reliable messages transfer.
Fig. 1 General steps for vehicle collision prediction
IV. VEHICLE COLLISION PREDICTION SYSTEM DESIGN
This section mainly focuses on various approaches and
techniques used to predict vehicle collision with respect to
steps discussed in section III.
A. Vehicular Communication
As a beginning of the Collision prediction system, vehicles
need to be aware of other vehicles in their vicinity. IVC and
V2R communication makes vehicle aware of other vehicle in
the locality with the help of Cooperative Awareness Messages.
Cooperative awareness is achieved by beaconing concept
namely, Static beaconing and adaptive beaconing. Static
beaconing is used to generate the beacons which contains
information required for calculating the collision possibility in
every beacon interval and passes the message downward to the
MAC [2].The information includes position, speed,
acceleration, and heading of the vehicle. Adaptive beaconing is
achieved by adaptive traffic beacon, whose objective is to
maintain the wireless channel congestion free, during exchange
of information in knowledge bases. This is done by sending
beacons as frequently as possible. This is achieved via DSRC a
Star
t
V2V
\
V2R communication
Vehicle Data Collection
Vehicle trajectory prediction
Potential Collision Predictio
n
Warnin
g
broadcastin
g
En
d
Yes
No
2014 IEEE International Conference on Computational Intelligence and Computing Research
1099
standard provided by IEEE for short range communication
[11][2].
Inter-Vehicular Communication is responsible for
communication among the vehicles on road. It is a part of ITS
which assists in avoiding motor vehicle accidents [12].Vehicle
accidents mostly occur due to NLOS (Non line of sight)
situation. In [13], author presented an approach to address the
problem of NLOS using IVC. Very High Frequency (VHF)
signals have the capabilities of reaching through obstacles
hence the proposed system used this ability for communication
between vehicles in NLOS. The data packets are frequently
broadcasted by the vehicles containing location, velocity,
direction, and acceleration information. During the
broadcasting of packets if any situation for collision is detected
by the system, it sends a warning message to the probable
colliding vehicles. Heavy traffic causes congestion in
communication which leads to an issue of packet collision [12].
Transmission Control Scheme is proposed to overcome the
above mentioned problem. In Transmission Control Scheme, a
vehicle omits its next transmission when it successfully
receives a packet from the preceding vehicles.
B. Vehicle Trajectory Prediction
Following Vehicle Communication, trajectory is predicted
as discussed in this section. In general, vehicle motion and
trajectory prediction is estimated by classifying modeled
entities. It includes three levels of classification, with degree
of abstraction [1] discussed below.
1) Physics-based motion models
It is simple in nature. Trajectory of vehicle is
predicted by considering the vehicle motion that only depends
on physics law. Dynamic and kinematic models are used for
purpose of prediction. Physical parameters of vehicle such as
steering, acceleration, car properties and external conditions
are used to evaluate the vehicle status.
2) Maneuver-based motion models
It is advanced in nature as compared to physic-based
model. Trajectory of vehicle is predicted by early recognition
of maneuvers that drivers intend to perform.
3) Interaction-aware motion model
Trajectory of vehicle is predicted by considering the inter-
dependencies between vehicles maneuvers. It is based on
Dynamic Bayesian Networks.The overview of various
trajectory prediction schemes is as shown in Table II.
C. Collision Avoidance
This section highlights different techniques for Collision
Prediction System. Commercial CWSs using radar and camera
have been available on the market. Cooperative Driving
Concept is one of the emerging trends in the development of
CWS. Cooperative driving concept is based on vehicle and the
infrastructure communication to get attentive of location,
intention information to surrounding vehicles or nearby
infrastructure [9].
Vehicles in VANET have certain issues such as different
speed of vehicles, always vehicles do not follow the straight
path and GPS may provide position information with some
delay. In [15], author mainly focuses on enhancing the
accuracy of cooperative collision warning system by
overcoming the aforementioned three issues. To deal with the
above discussed challenges a Vector-based Co-operative
Collision Warning system (VCCW) is developed. VCCW
system is Omni directional protection system and uses
acceleration information to handle three issues; hence some
special cases of collision pre-warning can be handled, which
involves collision coming from separate lane and many more.
VCCW system has high accuracy’s warning rates in both
intersection and curve situations. And it is also tested on real
roads which results in sending alert messages to the driver
before collision.
Xing [16] provides a solution to broadcast scheme which
results in reasonable delay and high delivery rate, without
causing broadcast storm problem. Since VANET is an attack-
prone network and security is one of the most crucial part in
communication. Any kind of malicious behavior in VANET
might cause serious loss or even death in reality, so the privacy
of each vehicle must be obtained. The private information must
not be available to all as it can be manipulated. A multi-hop
broadcast scheme that makes use of the RSU and V2I
communication is therefore proposed. The proposed scheme
has a leading edge over the static stochastic broadcast scheme
in terms of delivery rate. The performance of the scheme is
analyzed by its Message delivery rate, network delay and
number of broadcasted packets. The recent studies show that
most of the problem faced by the drivers is mainly due to the
deficiency of the communication between vehicles. The
accidents can be avoided if the drivers are alerted within
time before the collision occurs [10][17].
Hafner et al [17] presented computationally efficient
decentralized algorithms to avoid collision of two vehicles at
intersection by using V2V communication. Distinctively, the
collision avoidance problem can be addressed by computing
the set of states called as the capture set which is estimated by
Kalman filter [18] .The capture set is independent of the input
choice [20]. The model and state estimation algorithms are
used to report the uncertainty and communication delay in
collision avoidance [17]. This algorithm utilizes formal control
theoretic methods to guarantee a safe system. Decentralized
algorithm uses V2V communication to determine whether
automatic control is needed to avoid collision. Automatic
control is done by actuating the brake and throttle, assuming
that the driver follows the nominal path established by driving
lane. The developed system assures collision free environment
by keeping the system state always, outside the capture set and
the limitations on the control efforts. In addition overrides are
also used to prevent crash when required. Experimental
validation of the above method on two instrumented vehicles is
the gist of the research .Implementation of above method is
done by engaging the two vehicles in an intersection collision
avoidance scenario in a test track. A number of parameters can
be chosen by the designer, including the maximal and minimal
brake and throttle efforts for automatic control, maximal and
minimal speeds and the size of the collision set (bad set)[20].
2014 IEEE International Conference on Computational Intelligence and Computing Research
1100
TABLE I. OVERVIEW OF ISUUES ,APPROCHES,METHODS AND SIMLATION FOR VEHICLE COLLISION AVOIDANCE
TABLE II . OVERVIEW OF VARIOUS TRAJECTORY PREDICTION METHODS
Sr .No Name Issue Discussed Approach And method Simulation
1 Stefan Joerer [2] Safety protocol and application
enhancement based on IVC
• Cooperative awareness messages and beacon messages
used to collect vehicle dynamics
• Mathematical model used to calculate future trajectories.
• Probability distribution function used to calculate
probability of vehicle collision at intersection.
Veins i.e. combination
of SUMO and
OMNeT++
2 T. Kim [25] Widen the coverage of crash
detection systems including various
crash modes, driver behaviors and
vehicle dynamics.
• Algorithm for detecting an imminent collision in general
road scenes consists of following factors:
• Monte Carlo simulations for the generation of crash
probability data.
• Three driver models to distinguish between normal,
dangerous road scenes, and a point of no return scene.
MATLAB
(using three scenario)
1. Rear end
2. Cut –in
3. T bone case
3 Michael R. Hafner
[17]
To leverage V2V communication to
implement computationally
efficient decentralized algorithm for
two-vehicle cooperative collision
avoidance at intersections.
• Control theoretic methods to guarantee a collision-free
(safe) system.
• State estimation algorithm to model uncertainty and
communication delay.
• Automatic control by only actuating the brake and throttle
depend upon state estimation set.
Experiment is done on
two instrumented
vehicles.
4 Ali Rakhshan[8] Reduced false alarm taken
consideration into drivers reaction
time.
• Real-Time estimation of the distribution of brake response
times for an individual driver using VANET.
• Using the estimated distribution to customize warnings in
order to minimize false alarms
Simulation is done in
individual driver’s
reaction time.
S.no Name Issue Trajectory prediction estimation method Parameters Features
1 Stefan
Joerer[2]
Prediction of all possible
drivers behavior
A probabilistic model for trajectories to
represent all possible future driver
behaviors.
Distance from
intersection and
speed.
• Estimation of driver’s behavior
considering uniform
acceleration and more realistic
behavior.
• Physics-based
2 Georges.S.
Aoude [21]
Improper path prediction due
to different sources of
uncertainty in dynamic
environment.
1. environmental
sensing
2. environmental
predictability
Models are Classified into Worst-case,
pattern based, dynamic based, CL-RRT
(Closed-Loop Rapidly-Exploring
Random Tree Algorithm) algorithm with
a GP (Gaussian processes) mixture
model to calculate sets of moving objects
in real-time.
Vehicle
acceleration, vehicle
heading, vehicle
speed
• Allow vehicle constraints and
world obstacles into trajectory
predictions of target agents.
• Capture the distribution over an
extremely wide range of
possible target agent motion
pattern.
3 Xinlu Ma [22] Omni-direction CCWS
(Cooperative Collision
Warning System) with curve
situation
Collision detection is based on
determining the least distance of vehicles
and future trajectories both in
combination of heuristic approach by
analyzing the movement of vehicles in
curve situation.
Steering Dynamics
with Ackerman
condition.
• Steering dynamics of vehicles
is a part of interest, known as
Ackerman steering which has
most generalized configuration.
4 Joshue Perez
Rastelli [23]
Smooth and safe path
generation using road and
obstacle detection inform-
ation.
Intelligent Trajectory Generator, which
considers infrastructure and vehicle
information.
dimensions of the
vehicle, speed limit
and comfort
acceleration
• Achieves an automatic method
to obtain a path.
• Recently used in project
CityMobil21.
• Physics-based
5 Yunpeng
Wang[24]
Trajectory prediction method
is based on vehicle to vehicle
(V2V) wireless
communication.
Kalman filtering is used to predict
vehicle trajectory in order to eliminate
the GPS errors and also dynamic vehicle
equations.
relative positions,
relative velocities,
and azimuth angle
• Method is generic and adaptive
to different types of
intersection.
• Physics-based
2014 IEEE International Conference on Computational Intelligence and Computing Research
1101
V. GENERAL SIMULATION PARAMETERS
This section describes about the general simulation
parameters used by different simulators to perform
simulation. These parameters are used to estimate the vehicle
collision probability. Parameters prominently used are
Relative speed, Position, Heading angle, Action time,
Number of vehicles, Acceleration, Number of lanes and
Period of safety message transmission.
VI. CONCLUSION
In this paper, we have analyzed various schemes and
techniques to overcome the challenges in VANET such as
Vehicle collision prediction and transmission of emergency
messages. Significant factors involved in the system are
vehicle information, trajectory prediction, collision
probability calculation, alert messages. The importance of
Inter-vehicular communication to cope up with the same is
also discussed. Issues such as collision avoidance, security,
acknowledgement also plays a crucial role in the system.
General simulations parameters used to detect collision and
avoid collision are also mentioned. An overview of types of
collision warning and their communication modes are also
focus of this paper. The necessity and importance of
VANET for the populated and developed countries to adopt
this methodology is also provided.
REFERENCES
[1] WHO Media centre, “Road Traffic Injuries,”Fact sheet N°358 March
2013, available at http:// www.who.int/media centre/factsheets/fs358
/en/.
[2] Stefan Joerer, Bastian Bloessl and Christoph Sommer, “A Vehicular
Networking Perspective on Estimating Vehicle Collision Probability
at Intersections,” IEEE Trans On Vehicular Technology, Vol. 63, No.
4, May2014.
[3] Chung-ming, Chung-ming and Yao-chung Chang,“Telematics
communication Technologies and Vehicular Networks: Wireless
Architectures and Applications,” Book ,2009 .
[4] K.Lee and H.Peng,“Evaluation of automotive forward collision
warning and collision avoidance algorithms,”Vehicle System
Dynamics: International Journal of Vehicle Mechanics and Mobility,
vol. 43, no. 10, pp. 735–751, Oct. 2005.
[5] S. Biswas, R. Tatchikou and F. Dion, “Vehicle-to-Vehicle Wireless
Communication Protocols for Enhancing Highway Traffic Safety”,
IEEE Communications Magazine, Volume 44 Issue 1, Page 74-82,
January 2006.
[6] Stephan Olariu and Michele C. Weigle ,“CAR-2-X Communication
in Europe,”in Vehicular Networks: From Theory to Practice,
Chapman and Hall/CRC, Old Dominion University , page 4-
12,March 17, 2009.
[7] Brian Cronin, “Vehicle-to-Vehicle Communications for Safety,”ITS
Joint Program Office Research and Innovative Technology,available
at http://www.its.dot.gov/research/v2v.htm#sthash.GCEQCvMo.dpuf,
June 2014.
[8] Ali Rakhshan,Hossein Pishro-Nik , Donald L. Fisher and M.
Nekoui,“Tuning Collision Warning Algorithms to Individual Drivers
for Design of Active Safety Systems,” IEEE Globecom Workshop
(GC Wkshps) ,page 1333 - 1337 , Dec 2013.
[9] Chetankumar B.Kamble,Dr H. R. Deshmukh and A.
Chandure,“Vehicle Collision Alert,” International Journal Of Pure
And Applied Research In Engineering And Technology,Volume 2
(9): 1061-1065, 2014.
[10] B. Djamel,G. Nacira and T.Cherif,“Forecasting Approach in VANET
based on vehicle collision alert,”International Conference on
Multimedia Computing and Systems (ICMCS), page 573 – 577,May
2012.
[11] Ozan K. Tonguz, Christoph Sommer and Falko Dressler,“Traffic
Information Systems:Efficient Message Dissemination via Adaptive
Beaconing,”IEEE Communications Magazine,Volume:49, Issue:5,pp
173 - 179 ,May 2011.
[12] S.Tsugawa,“Inter-vehicle communi-cations and their applications to
intelligent vehicles: an overview,”IEEE Intelligent Vehicle
Symposium, vol.2 ,pp 564 - 569 ,June 2002.
[13] Hidenori Suzuki ,Hidekazu Murata and Kiyomichi Araki
,“Performance of Inter-vehicle Communication Technique for
Intersection Collision Warning,”Fifth International Conference on
Information,Communi- cations and Signal Processing (ICICS),pp 603
- 606 , 2005.
[14] Stephanie Lefevre,Dizan Vasquez and Christian Laugier,“A survey
on motion prediction and risk assessment for intelligent
vehicles,”Springer ROBOMECH Journal,vol.1, July 2014.
[15] Chung-Ming Huang and Shih-Yang Lin,“Cooperative vehicle
collision warning system using the vector-based approach with
dedicated short range communication data transmission,”IET
Intelligent Transport Systems,Volume:8,Issue: 2 ,pp 124 - 134 ,March
2014.
[16] Xing Fan, Bo Yang , Ryo Yamamoto and Yoshiaki Tanaka ,“Road
Side Unit Assisted Stochastic Multi-hop Broadcast Scheme for
Instant Emergency Message Propagation,”16th International
Conference on Advanced Communication Technology
(ICACT),pp103-108,Feb 2014.
[17] Michael R. Hafner, Drew Cunningham, Lorenzo Caminiti, and
Domitilla Del Vecchio,“Cooperative Collision Avoidance at
Intersections: Algorithms and Experiments,”IEEE Transactions on
Intelligent Transportation Systems,Volume:14,Issue: 3,pp 1162 –
1175,April 2013.
[18] Liang Xue and Dongxiao Zhang,“A multimodel data assimilation
framework via the ensemble Kalman filter,”Water Resources
Research, Volume 50, Issue 5,2014,Available at http:// online library.
com/ doi/ 10.1002/2013WR014525/abstract.
[19] C. J. Tomlin, J. Lygeros and S. Sastry,“A game theoretic approach to
controller design for hybrid systems,”IEEE Proceedings,vol. 88, no.
7, pp. 949–970, July. 2000.
[20] A. Colombo and D. Del Vecchio,“Efficient algorithms for collision
avoidance at intersections,” Proceedings of the 15th ACM
international conference on Hybrid Systems: Computation and
Control, pp. 145–154,2012.
[21] Georges S. Aoude, Joshua Joseph, Nicholas Roy and Jonathan P.
How, “Mobile Agent Trajectory Prediction using Bayesian
Nonparametric Reach ability Trees”,American Institute of
Aeronautics and Astronautics,2011.
[22] Xinlu Ma,Yeren Liu and Jiang Cheng,“Analytic and Heuristic
Methods of Collision Detection for Curve-Moving Vehicle Based on
Cooperative Collision Warning System”,Journal of convergence
Information Technology(JCIT) ,Volume 8, Number 2, Jan 2013.
[23] Joshue Perez Rastelli ,Ray Lattarulo and Fawzi Nashashib,“Dynamic
Trajectory Generation Using Continuous-Curvature Algorithms for
Door to Door Assistance Vehicles,”IEEE Intelligent Vehicles
Symposium Proceedings, pp 510 - 515,June 8-11,2014.
[24] Yunpeng Wang,Wenjuan, E, Daxin Tian ,Guangquan Lu,Guizhen Yu
and Yifan Wang,“Vehicle Collision Warning System and Collision
Detection Algorithm Based on Vehicle Infrastructure
Integration,”IET, The Seventh Advanced Forum on Transportation of
China,pp 216 – 220, 2011.
[25] T .Kim and H. Y Jeong,“A Novel Algorithm for Crash Detection
Under General Road Scenes Using Crash Probabilities and an
Interactive Multiple Model Particle Filter,”IEEE Transactions on
Intelligent Transportation Systems, Volume:15 , Issue 6, 14 May
2014.
2014 IEEE International Conference on Computational Intelligence and Computing Research
1102