Improved A* Algorithm-Based Optimal Path Planning of Rescue Robots Within Multi-Environment Maps
Jingrui Zhang, S.X. Wu, Houde Liu, Xiaojun Zhu, Bin Lan · Machines · 2025
The traditional A* algorithm performs well in single-map environments, but it is prone to path redundancy and obstacle handling delays in complex multi-map collaborative scenarios, making it unsuitable for the characteristics of multi-environment maps. To address these challenges of traditional A* algorithms, this paper proposes a multi-environment map rescue robot path planning method based on an improved A* algorithm. This method introduces an expected cost evaluation function to achieve weighted fusion of path costs and heuristic values from multiple maps, allowing the algorithm to integrate obstacle distributions and weight information across different environments. A random obstacle replacement mechanism is further designed to maintain path feasibility by locally substituting blocked nodes with adjacent accessible nodes, thereby ensuring continuity without global replanning. Through the combination of multi-map information fusion and local obstacle handling, the algorithm generates a globally optimized path that balances planning efficiency, robustness, and adaptability in uncertain rescue scenarios. Experiment results for a 50 × 50 map scenario show that the improved algorithm significantly outperforms single-map planning results in terms of path redundancy, total length, and turning characteristics. The expansion experiments demonstrate that the paths planned by the proposed algorithm are highly consistent with the optimal paths in terms of direction and local deviations, verifying its good feasibility and effectiveness.