Path planning for unmanned mining vehicles based on improved A^* algorithm under multi-gradient risk constraints

To address the problems of poor path smoothness and potential collision risks in existing path planning algorithms for unmanned mining vehicles in unstructured underground environments, this paper proposed an improved A^* algorithm for path planning of unmanned mining vehicles, namely the RA-A^* algorithm, which integrated multi-gradient risk constraints with path search. Based on a risk potential field, the environment was divided into obstacle core zones, risk buffer zones, and absolute safety zones, and a multi-gradient risk map was constructed to provide a fine-grained representation of roadway safety margins. The comprehensive cost function of the traditional A^* algorithm was reconstructed. Risk values and turning penalties were incorporated into the actual cumulative cost function, enabling paths to actively stay away from high-risk areas and reducing unnecessary turns. A risk prediction term was introduced into the heuristic function to avoid potential hazards ahead. A dynamic update and re-planning mechanism for the multi-gradient risk map was introduced. When dynamic obstacles were detected, the multi-gradient risk map was updated in real time, and the path was replanned to avoid the dynamic obstacles. Cubic B-spline curves combined with an iterative collision detection mechanism were used to smooth the planned paths. Multi-scenario simulations and underground experiments showed that the paths planned by the RA-A^* algorithm could actively shift toward low-risk areas. The algorithm maintained a safe obstacle clearance distance when encountering static obstacles, replanned paths and generated safe detour trajectories when encountering dynamic obstacles, avoided approaching roadway walls and obstacles, and produced relatively smooth paths at corners. The RA-A^* algorithm reduced path risk value, the sum of turning angles, and number of turns, while improving average obstacle clearance distance. It ensured that vehicles maintained a relatively large minimum obstacle clearance distance even under extremely narrow working conditions.