Path planning based on improved bidirectional A* algorithm for AGV

Abstract Considering the issues like numerous nodes, poor search efficiency, and sluggish operation of traditional A* in relatively complex factory setups, an optimized A* algorithm relying on 16-neighborhood bidirectional node search is proposed. By improving the search strategy of the 8-neighborhood of the traditional algorithm, 16 search neighborhoods are selected around the parent node. Different nodes are eliminated according to different strategies in the process of bidirectional node expansion. At the same time, the evaluation program undergoes an improvement process to make adjustments to the heuristic function according to the obstacle rate to avoid falling into the area with too high an obstacle rate. The ordered array combined with the binary search tree index is used for structural optimization to improve efficiency while guaranteeing the practicability of the route. Finally, the path is optimized. By comparing various algorithms through sim-ulations and experiments, the algorithm refined in this paper demonstrates remarkable effect-iveness. In terms of the count of expanded nodes, it has been reduced by 50%. At the same time, the search time has also been shortened by approximately 40%.