关键词: A^*算法, 动态窗口法, 机器人, 避障, 运动学约束, 路径评价

Abstract: To solve the problems of the traditional A^* algorithm in practical application,such as insufficient dynamics,low obstacle avoidance efficiency and long running time,a method combining kinematically constrained A^* algorithm with Dynamic Window Approach (DWA) was proposed. Firstly,considering the robot kinematics,the node extension method and heuristic function model of A^* algorithm were improved to improve the search efficiency and reduce the number of path turns,so that the path generated by A^* algorithm can better meet the requirements of robot motion. Secondly,the path evaluation function of DWA algorithm was optimized to make the local path planned by DWA algorithm more smooth and coherent,which is more conducive to robot motion execution. Finally,key nodes were extracted from the global path planned by A^* algorithm to guide DWA algorithm to local planning and dynamic obstacle avoidance. Simulation results show that compared with other algorithms,improved A^*-DWA algorithm can save up to 45.09 % of planning time and reduce the driving distance by up to 13.49 %. Experiments show the effectiveness of improved A^*-DWA algorithm.

Key words: A^* algorithm, Dynamic Window Approach(DWA), robot, obstacle avoidance, kinematically constrain, path evaluation