Hybrid Lunar Walking Path Planning Based on Improved Bi-RRT and A*

This paper proposes a hybrid path‑planning algorithm that integrates an improved Bidirectional Rapidly‑exploring Random Tree (Bi‑RRT) with an enhanced A* search to support safe astronaut traversal on the lunar surface under complex constraints. First, an obstacle density adaptive goal-bias Bi-RRT algorithm is used to rapidly explore complex lunar terrain and generate an initial feasible path, effectively managing the computational challenges posed by large-scale search spaces. Then, the search space is constructed by morphological expansion operation on the initial feasible path, which provides effective guidance for fine path optimization. Finally, within this region a multi‑constraint weighted A* algorithm plans the final path, focusing on balancing multiple indicators such as energy consumption, risk, and illumination conditions. Experiments on real lunar terrain data show that the hybrid method achieves a superior balance across path length, success rate, metabolic energy consumption, risk avoidance, and illumination conditions when compared with existing planners. By uniting the fast exploration of Bi-RRT with the precise optimisation of the weighted A*, the algorithm improves both efficiency and quality of path planning in complex lunar environments and provides an efficient, reliable solution for astronaut lunar walking.