Vehicle infrastructure integration system using vision sensors to prevent accidents in traffic flow

This study describes the development of a vision sensor for detecting shock waves which is one of the main factors of accidents in highway traffic flow. The major contributions of this research are development of vehicle tracking and detection of shock wave in saturated traffic. Moreover, realisation of a vehicle infrastructure integration (VII) system for providing arrival information of such propagation to drivers is proposed. The experiment on the analysis of the propagation with the developed image sensor has shown that an error might occur in the arrival time information of the propagation provided to the driver. Therefore a prediction technique at the arrival time of the propagation is integrated in the authors´ system. By using this prediction technique and taking the error tolerance of drivers into account, the experimental results show that prediction success rates are improved by about 5%.