Abstract: A particle model for vehicle collision avoidance is established to conduct a comparative analysis on the effectiveness of three different collision avoidance methods. The results show that during vehicle high-speed emergency collision avoidance, steering collision avoidance requires smaller longitudinal distance than braking collision avoidance does and is closer to that with combination of steering and braking collision avoidance. Based on this, an automatic steering collision avoidance controller combining feedforward and feedback is designed. By selecting seven-order polynomial reference path with smooth transition and combined with 2-DOF linear steering dynamics model, the feed-forward wheel turning angle can be obtained. In addition, PID feedback controller is designed with the lateral displacement deviation between actual path and reference path as feedbacks. Finally, a simulation verification is performed on the controller designed. The results show that the feedforward and feedback tracking controllers designed have good tracking effect in actual practice, high robustness and strong ability to resist side wind and other external disturbances, and hence can achieve effective collision avoidance.

Key words: vehicle, automatic emergency steering, collision avoidance, feedforward control, feedback control