Research And Simulation Of Vehicle Collision Warning System Based On VANET

Based on VANET, this paper proposed a vehicle collision warning system. Each vehicle takes advantage of the driving information of itself and the environmental information to establish the trajectory prediction model, which is used to predict the trajectory of the vehicle in the next N seconds. Vehicles exchange the predicted information via VANET broadcast protocol. After receiving all the neighbors’ data, the collision warning algorithm can estimate the possibility of collisions, and warn the driver to take measures to avoid danger. Therefore traffic accidents can be greatly reduced.The main contents and innovations of this paper are:1) The system used VANET to overcome the environmental limitations, and with the help of the GPS device and on-board screen, the driver can get comprehensive perception of the statuses and locations of all the vehicles around.2) The paper proposed the concept of vehicle safety boundary. A vehicle is regarded as a rectangle which can be positioned by four points’ coordinates. This method can improve the accuracy of the collision determination algorithm.3) An efficient adaptive broadcast protocol is proposed. The protocol uses the number of neighbor nodes’ driving direction and the distance from the neighbors to itself as the metric to select a relay node. The simulation results show that our algorithm has several advantages like: less rebroadcast packets, high transmission efficiency, and higher channel utilization.4) The paper proposed a new algorithm of collision detection. According to the predicted trajectories of the vehicle itself and all other neighbors, the vehicle can determine if there are collision dangers in the near future.