Research on Path Planning for Robots with Improved A* Algorithm under Bidirectional JPS Strategy

Abstract

Aiming to address the A* algorithm’s issues of traversing a large number of nodes, long search times, and large turning angles in path planning, a strategy for multiple improvements to the A* algorithm is proposed. Firstly, the calculation of the heuristic function is refined by utilizing the Octile distance instead of traditional distance, which more accurately predicts the optimal path length. Additionally, environmental constraints are introduced to adaptively adjust the weight of the heuristic function, balancing the trade-off between search speed and path length. Secondly, the bidirectional jump point search method is integrated, allowing simultaneous path searches from both directions. This significantly reduces path search times and the number of nodes traversed. Finally, the path undergoes two rounds of smoothing using a path smoothing strategy until the final path is generated. To validate the effectiveness of the improved A* algorithm, simulations are conducted on ten types of grid maps. Results demonstrate that the improved A* algorithm markedly decreases path search times while maintaining path length, with greater speed improvements observed as the map size increases. Furthermore, the improved algorithm is applied in experiments with mobile robots, achieving significant reductions in average path search times of 79.04% and 37.41% compared to the traditional A* algorithm and the JPS algorithm, respectively. This enhancement effectively meets the requirements for rapid path planning in mobile robotics applications.