A Hierarchical Framework for Dynamic Global Path Planning

This article presents a hierarchical global path planning algorithm designed for mobile robots. The global path planning problem is divided into three distinct components: waypoint generation, reference path generation, and path optimization. In the waypoint generation phase, we allocate the computation of environmental changes and map updates within each frame point cloud based on dynamically updated visibility graph. We then derive waypoints by identifying visibility edge vertices. Next, a smooth reference path is obtained through B-spline fitting optimization, delineating the convex driving region within the environment for further optimization within homotopy classes. To mitigate collision issues arising from B-spline fitting, we formulate the path optimization problem in the Frenet coordinate system as a quadratic programming task. This approach yields a globally collision-free path with high computational efficiency. Simulation experiments have been conducted to validate the effectiveness of the algorithm proposed in this paper.