Path planning for unmanned load–haul–dump machines based on a VHF_A* algorithm

Abstract

This study aimed to address the low search efficiency, long planning time, and unsmooth path problems in narrow and curved underground mining cave environments using traditional path planning algorithms. Thus, a new path planning method (VFH_A *) based on VFA and A* is proposed for unmanned load–haul–dump (LHD) operating in underground mines. First, the trajectory of the LHD articulation point is considered, the nodes are extended using this point, and collision detection is performed. Notably, the extended nodes neither collide nor conform to the trajectory of the LHD. Second, the vector field histogram (VFH) algorithm is introduced, and the steering and collision threat costs are combined with the comprehensive cost function of the traditional A* algorithm to form the VFH_A* algorithm. Thus, the node with the lowest comprehensive cost is selected for expansion, and redundant nodes are eliminated from the searched node paths. Third, the path is smoothed via Bezier interpolation. This ensures that the LHD operates smoothly and prevents excessive changes in the articulation angle. The proposed method was evaluated using above-ground and underground simulations. Compared with A* and GA algorithms, the VFH_A* algorithm significantly improved the search efficiency and can efficiently generate safe and smooth task paths in different scenarios. The planning time for each scenario was reduced by 85%. Finally, tracking experiments were conducted on the planned task paths of 1.5 and 2.5 m/s, indicating that the tracking error was less than 0.23 m. Overall, the planned path meets the requirements of unmanned LHD, indicating that the proposed method can adapt to practical applications.