COLLISION AVOIDANCE ON A CURVED ROAD IN VANET

Abstract

The world of transport system has transformational changes, from horse driven carriages and paved roads to a more complex road transport system consisting of a variety of vehicles and other advanced infrastructure, all put in place in order to provide safe and efficient mobility of vehicles. With this trans formational change in transport system, number of road accidents have also increased exponentially. To minimize the probability of road accident, it must be ensured that present day vehicles must be equipped with latest technology which has been evolved in transportation world. For collision avoidance, a large number of algorithms have been proposed. Advanced technology along with preventive algorithms, provides necessary information regarding probability of collision occurrence to drivers. This necessary information is provided either by wireless communication or by physical means like sign boards and warning signals. In wireless communication based systems such as Vehicular Ad-Hoc Networks (VANETs), there are two common approaches: vehicle to vehicle (V2V) and vehicle to infrastructure (V21). Each of them has some pros and cons. Generally, the two main challenges for any approach are - how to accurately estimate velocity and changes in trajectory (course). Any warning message which is transmitted to drivers must contain these two conditions. There are some algorithms which work on assumption of constant velocity, which is not fruitful. There is another category of algorithms that works efficiently on cross section of road but does not provide efficient result at the curved section of road. Hence for overall efficiency, there must be a system which handles all such constraints. We propose an algorithm for overcoming the problems of curved section of road, based on relative velocity vector method.