Research On Pedestrian Collision Avoidance Algorithm Based On Data Fusion And Trajectory Prediction

Aiming at the problem that the vehicle active collision avoidance system is not effective and the collision avoidance mode is simple,this dissertation analyzes and studies the pedestrian detection,trajectory prediction,active collision avoidance strategy and vehicle collision avoidance control,and develops a pedestrian collision avoidance algorithm based on data fusion and trajectory prediction.First of all,after fully investigating the research status and key technologies of vehicle active collision avoidance system at home and abroad,this dissertation designs the overall framework of pedestrian collision avoidance algorithm with "perception,prediction,decision making and control" as the main idea.Secondly,in order to improve the pedestrian detection effect,a pedestrian detection method based on millimeter wave radar and Mobileye camera data fusion is designed in this dissertation.The detection data of the two sensors are fused by the weighted average method.Based on the analysis of common pedestrian trajectory prediction methods,a pedestrian trajectory prediction method based on least square curve fitting is designed.Thirdly,based on the TTC collision model and pedestrian trajectory prediction method,this dissertation divides the lateral and longitudinal collision areas between vehicle and pedestrian.For different collision scenarios,four collision avoidance modes,such as collision avoidance by decelerating,collision warning,emergency braking and emergency collision avoidance by steering,are developed.Accordingly,the braking,path planning and path tracking control methods are designed.Finally,the co-simulation platform based on Pre Scan/Car Sim/Simulink is built,and the algorithm is simulated and verified.On this basis,the real vehicle test platform is built,and the real vehicle test is carried out under different scenarios.The results of simulation and real vehicle tests show that the pedestrian collision avoidance algorithm based on data fusion and trajectory prediction can achieve good detection and trajectory prediction of pedestrians,and achieve pedestrian collision avoidance in various scenarios,which is practical and advanced.