Collision-Free Path Planning for Uavs in Narrow Mine Tunnel Environments Based on Voronoi Diagram

In narrow mine tunnel environments, unmanned aerial vehicles (UAVs) are prone to collisions due to proximity to tunnel walls or objects. Nevertheless, few existing path planning algorithms adequately address dynamic obstacle environments. This paper proposes a dynamic obstacle avoidance path planning algorithm for UAVs based on Voronoi diagram initialization. First, an optimal path search method based on Voronoi diagrams is introduced. Delaunay triangulation is performed on static obstacles within the environmental grid map to generate a Voronoi skeleton and extract feasible waypoints. Next, real-time dynamic obstacle information is integrated into the trajectory planning phase, considering uncertainties in obstacle velocity and direction, and triggering local replanning when collision risks are detected. Finally, the A* algorithm is employed to search for the shortest collision-free path, and a smooth final path is generated by fitting cubic B-spline curves. Numerical simulations and experiments in the ROS-Gazebo physical simulation environment demonstrate that the proposed method enhances the UAV's dynamic obstacle-avoidance capability. By increasing the distance between the UAV and obstacles during trajectory planning, it effectively reduces collision risks when operating in narrow tunnel environments.