Decision-making of active collision avoidance system based on comprehensive evaluation method of dangerous scenarios

Abstract

Dangerous scenario modeling is of great significance for the development of automatic driving functions, and the active collision avoidance system directly affects the safety of vehicles and the intelligence level of vehicles. However, the current existing collision risk evaluation methods are difficult to comprehensively evaluate different dangerous scenarios, the evaluation methods are complicated and there is no normalized description form. In addition, the traditional decision-making method of active collision avoidance system based on collision time distance will lead to the problems of inconsistent risk evaluation scale and single way of collision avoidance. To solve the above problems, the mechanism analysis of the longitudinal, lateral, and parallel collision risks is carried out respectively, and then the model based on kinematics and dynamics is established. After that, the corresponding expression form of collision risks is derived. Finally, the comprehensive evaluation method of dangerous collision scenarios is proposed. Secondly, based on the comprehensive evaluation method of collision risk, considering longitudinal constraints, lateral constraints, parallel constraints, and road regulations, an active collision avoidance decision-making method that used the cooperative collision avoidance mode of breaking and steering is designed. Thirdly, different simulation experiments are designed based on PanoSim/Simulink. The results show the comprehensive evaluation method can effectively model the collision risk of vehicles in the motion process. At the same time, the decision-making method of the active collision avoidance system achieves the expected working effect, thus improving driving stability and safety.