Contact Sequence Planning for Hexapod Robots in Sparse Foothold Environment Based on Monte-Carlo Tree

Abstract

Legged robots can pass through complex field environments by selecting gaits and discrete footholds carefully. Conventional methods plan gaits and footholds separately and treat them as a single-step optimal process. However, such approaches cause poor passability in sparse foothold environments. This letter proposes a novel coordinative planning method for hexapod robots. It treats gait and foothold planning as a sequence optimization problem while considering the harshness of the environment as the leg’s fault. The Monte Carlo tree search (MCTS) algorithm is used to optimize the entire traversing motion sequence. A slidingMCTS method is proposed to effectively strike a balance between optimization and search operations by introducing a moving root node and controlling the sampling time. The proposed planning algorithm takes advantage of the fault-tolerant mechanism, lifting legs without valid footholds and planning the contact sequence of the remained legs, to improve the passability of the hexapod robot in harsh terrains. The method has been compared with the RRT-based search method for terrains with different densities of foothold, and experiments on challenging terrains are carried out to verify the efficiency. The results have shown that the proposed method dramatically improves the hexapod robot’s ability to pass through sparse-foothold environments.