Prevent collisionavoidance system like ship, cycle etc. uses sensors


Prevent Vehicle Collision by Different Methods

Joseph Saji                                                                    
 Basil C Sunny
               Department Of Computer
Department Of Computer Science                                             

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!

order now

Shankara Institute Of Engineering and Technology             Adi Shankara Institute Of
Engineering and Technology
Kalady 683574
[email protected]                                                               [email protected]            

Abstract The topic is related to the vehicle collision avoidance system.
All the other papers of vehicle collisionavoidance system like ship, cycle etc.
uses sensors that only avoid collisions and was only effectivewithin certain
speed ranges. In this paper, a scooter collision avoidance system is proposed
that canavoid accidents at intersections and also warns the other scooter
vehicles. The relevance of thispaper is that here infrastructure based
solutions such as those utilizing radar or camera are notconsidered. It
advances by using a smartphone that provides high penetration rate. The paper
hasmany advantages like low cost, not affected by obstructions, functions at
every intersections.


Keywords: Machine learning, collision avoidance, red-light
runner, scooter, motorcycle.




cooter is a type
of motorcycle which is the most important transportation means within several
countries. The fuel efficiency of scooters is at least two times better than
that of cars and its sale price is only about one-tenth of that of regular
passenger cars .Due to their smaller sizes, it is much easier for scooters to
pass through traffic congestion, hence shortens the time to reach the destination.
This helps the people who live in urban areas and suburban areas to prefer the
use of scooters over cars.   But also
there are problems caused by the high density of scooters. In particular, road
traf?c safety has been the main concern for the scooter riders. Statistics show
that, most of deaths in traf?c accidents are caused by those involving
scooters. It also showed that the accidents were much less for one travelling
in the cars than the scooters. This is because when the scooter is used one needs
to maintain his/her balance while riding but the car provides a shelter within
inside. Many advanced safety features has also resulted in the favor of
passenger cars. Even though the safety measures can be used for scooters the
problem is the cost for sensors, such as cameras and radar sensors, and costly
processing circuits. Examples include Blind Spot Information System (BLIS),
Line Departure Warning System (LDWS), pedestrian detection system, and forward
collision warning system. When the problem was considered as a serious issue
they came to the conclusion that the red light runners are the cause for the
scooter collision.

          Scooter collision avoidance system
can identify red-light runners (RLRs) at intersections .The system would advise
the RLR to slow down immediately and warn nearby vehicles on the intersecting
road in real time. This scooter collision technique is done using smart phones
carried by scooter riders. RLR classifier is used for learning and predicting
RLR behavior




1.Red Eye

                  Fig 1 : Red Eye Architecture


2.Data collector

data collector periodically obtains data from the GPS, the accelerometer, and
the camera in the smartphone. Red Eye utilizes the data from GPS to generate a
number of features such as

Distance to the


Traffic light






Fig 2: RLR Classifier

The RLR classifier predict
whether a RLR behavior is likely to happen.


5.Distance-based SVM (D-SVM)


D-SVM takes one set of sensor
data every 5 meters. The algorithm starts collecting the data when the scooter
is X meters from the intersection and stops when the scooter is Y meters from
the intersection to perform a prediction. The SVM prediction model returns a
confidence value and compare it with Pre-configured threshold.


6.Multi-Distance-based SVM


MD-SVM is designed to address
the problem of D-SVM.If confidence value of that prediction is higher than the
Pre-configured threshold,it reports RLR behavior .The system waits for the
scooter to travel for another 5 meters to perform the next prediction .The
threshold for the confidence value were initially set to a higher value, and
then gradually decreased by multiplying the original threshold with a R value
(0 oR1). MD-SVM allows the system to report a RLR event early, when there is
strong evidence in the data.



  4.Warning Message Transmitter/Receiver


3: Warning Message Transmitter/Receiver



Almost all modern smartphones
are equipped with an IEEE 802.11b/g/n WiFi radio.

Red Eye incorporated a
custom-made range extender. Antenna coupler to capture or inject the WiFi
signal from/to the built-in antenna.Signal amplifier to compensate for the loss
caused by the back cover of the phone.





The area of this paper is based
on the different collision avoidance system for various vehicles. Inship
collision avoidance, the method was based on model predictive control. It was
based on

simulation and optimisation .In
path planning and tracking for vehicle collision avoidance, the

method was also same with multi
constraints. But the paper was only possible in straight roads. Incyclist
collision avoidance system, sensors was used and it provided reliable
detections but with highspeeds, the sensors failed to provide consistent
detection .Then a new framework for collisionprediction was provided but it was
more complex. In the proposed paper, the scooter collisionavoidance was based
on smartphones which replaced sensors/radar. It not only prevents collisionbut
also warns the other scooters. It has high accuracy and good computational



In this paper, a collision
avoidance system proposed to prevent collision caused by RLR scooters.  The paper utilizes smart phones on board
scooters for predicting RLR behaviors. Once the RLR patterns are detected, the
rider would be advised to de-accelerate immediately and nearby vehicles would
receive repeating warning messages via built-in WiFi radios in smart
phones.  As smart phones are owned by a
large number of riders today, low-cost, onscooter after-market solution can be
easily adopted in a manner similar to any other smart phones apps


1 Vehicle collision
avoidance via control over a finite-time horizon. Authors: Marco Ariola;
Gianmaria De Tommasi; Francesco Amato 2017 IEEE 14th International Conference
on Networking, Sensing and Control (ICNSC) Year: 2017 


2 Ship Collision Avoidance
and COLREGS Compliance Using Simulation-Based Control Behavior Selection with
Predictive Hazard Assessment. Authors: Tor Arne Johansen; Tristan Perez; Andrea
Cristofaro IEEE Transactions on Intelligent Transportation Systems


 3Path Planning and Tracking for Vehicle
Collision Avoidance Based on Model Predictive Control With Multiconstraints.
Authors: Jie Ji; Amir Khajepour; Wael William Melek; Yanjun Huang IEEE
Transactions on Vehicular Technology Year: 2017 


 4 Field Testing of a Cyclist Collision
Avoidance System for Heavy Goods Vehicles. Authors: Yanbo Jia; David Cebon IEEE
Transactions on Vehicular Technology Year: 2016 


5 A New Framework of Vehicle
Collision Prediction by Combining SVM and HMM. Authors: Xiaoxia Xiong; Long
Chen; Jun Liang IEEE Transactions on Intelligent Transportation Systems Year:


6 The research on the
vehicle collision avoidance control based on vehicle motion estimation.
Authors: Ren Yue; Zheng Ling; Li Zhe; Yang Wei; Li Yinong; Wang Ke; Li Yusheng;
Xiong Zhoubing IET International Conference on Intelligent and Connected
Vehicles (ICV 2016) Year: 2016