The collision warning system, based on multi-sensor, can perceive more information than a unilateral system, just like vehicle motion status, driver intention, and road traffic environment; thus, it can provide more accurate judgments of driving safety. In this paper, physical and logical framework of a collision warning system based on multi-sensor is designed in detail. Vehicle to vehicle communication (V2V), vehicle to infrastructure communication (V2I), and vehicular sensors are used to acquire the motion status of the subject vehicle, neighboring vehicles, and pedestrians. Then, the short-term future motion status is predicted based on the Kalman algorithm, and the vehicle collision risk can be detected in advance by a collision detection algorithm. Furthermore, the safe distance threshold as a collision risk indicator is calculated and compared with the predicted distance. Finally, the vehicle collision risk can be detected according to the comparison. Simulation results show that the method is effective for judging the collision risk and providing an accurate warning prompt.
展开▼
