Vehicle active security based on driver modeling

The vehicle passive safety technology can only solve the problems caused by traffic accidents. The active safety technology, which can prevent and reduce accidents, would suffice for more far-reaching applications. In this paper Elman neural network is adopted to predict driver's behavior ahead of time. The "people oriented" driver-vehicle-road closed loop model is set up. The system would record the habits of the driver and warn in time when the behaviors of the driver deviate from the forecasted trajectory to a certain extent. Real time simulation is carried out, which is based on 3D urban road that acquired by GPS equipment. The results indicate that Elman algorithm can be used to establish the warning system of driver's improper operation and provide the reliable and valuable information for safe driving.