Rough Terrain Path Planning for Autonomous Ground Robot
Xuan-Phat Truong, Seong Hyeon Hong · 2024
This paper proposes a new methodology to solve the path planning problem for rough terrains such as planetary surfaces, particularly for autonomous mobile robots. For the purpose of the algorithm development, the map is assumed to be known. The decomposition step is performed to approximate rough surfaces into smaller cells, preserving the height information. Any unpreserved information due to decomposition is considered for the path planning step since the height that represents the cell may not be accurate. The A Star (A*) algorithm is adopted as a basic skeleton for our approach and necessary modifications are made to perform the path planning on 3D maps. This is accomplished by considering the height information of the map and incorporating this information into the cost function in variety of ways. Additionally, not only the known obstacles but the algorithm is able to detect and avoid the unknown obstacles by replanning when significant map information is updated. The simulation experiments were conducted to validate the effectiveness, speed, and robustness of the proposed method.