Improvement and Fusion of A* Algorithm and Dynamic Window Approach Considering Complex Environmental Information

Abstract

Path planning is a key technology for autonomous robot navigation; in order to allow the robot to achieve the optimal navigation path and real-time obstacle avoidance under the condition of complex and bumpy roads, an optimization algorithm based on the fusion of optimized A* algorithm and Dynamic Window Approach is proposed. The traditional A* algorithm generates the optimal path by minimizing the path cost. But when the road in the area where the robot is located is uneven, the path planned by the traditional A* algorithm may be the shortest but not the optimal. Because the robot can pass the ravines and bumps on the road, the cost of passing is higher at this time. At this point, for the robot, path planning must consider the path length and the number of undulations in the path. For the heuristic function of the traditional A* algorithm, the weight information of the road surface is added to it. The optimization algorithm can obtain an optimized path avoiding a lot of bumpy roads. Since the path has a large number of redundant turning points, the turning point extraction strategy is adopted to delete the redundant points of the path, and finally an optimal path with a little bumpy road, short length, and few turning points is obtained. Secondly, in order for the robot to obtain local obstacle avoidance capabilities based on the global optimal path, the optimized A* algorithm is combined with the Dynamic Window Approach to obtain a fusion algorithm that combines global path planning and local path planning. Experimental simulation results show that this algorithm can effectively avoid unnecessary bumpy roads, remove redundant turning points, improve path flatness, increase path smoothness, and achieve a compromise between path length and road surface undulations. At the same time, the local real-time obstacle avoidance capability based on the optimal path is also increased.