Forecasting Approach in VANET based on vehicle collision alert

This paper deals with the forecasting of collision events in Vehicle Ad hoc Networks (VANET). Using significant parameters of each vehicle such that: position, speed, acceleration and direction, it is possible to avoid a possible collision. We present a collaborative forecasting module for VANET. The proposed module is focused on the estimation of these parameters using a kinematic model of each vehicle. As a result, each vehicle of the network could have a posterior global view of the dynamics of its network. The forecasting module is embedded in all vehicles of VANET and estimates the parameters of the subject vehicle and those of its neighbors. Estimated parameters is used beforehand to generate alerts in case potential risk. The communications between nearby vehicles (V2V communications) and the road-side infrastructure (V2I communications) is carried out by DSRC protocol (Dedicated Short Range Communication). Simulation results illustrate the estimation vehicles parameters.