Mag-Planner: A Hierarchical Path Planning Method with Exploring Priority for Narrow Space in Open-pit Mining Scenario
Haowei Lu, Yafei Wang, Yichen Zhang, Ruoyao Li, Mingyu Wu, Jiyuan Zhang · 2024
Autonomous driving technology for mining trucks is gradually being applied in mining scenarios to reduce accidents and driving hazards. However, due to the debris dropped during loading and transportation, the drivable space is constrained and becomes uncertain, which greatly increases the difficulty of path planning. The current mainstream planning algorithms often fall into blind exploration in such scenario due to the lack of complete ground free-space distribution information. To deal with this challenge, we propose a novel Mag-Planner to extract ground free-space distribution with exploring priority and plan path based on it. First, we determine the exploring priority value of each area in the map by magnetic conduction method and build a free-space configuration node graph. Then, we design a hierarchical planning method based on the exploring priority and search and backtrack nodes through the graph to find the optimal global path. We conducted field tests in an open-pit mine and got results demonstrating that Mag-Planner can reduce time cost by at least a quarter and increase success rates by at least 80% compared to K-PRM*and improved Hy-A*.