Efficient Locomotion Planning for a Humanoid Robot with Whole-Body Collision Avoidance Guided by Footsteps and Centroidal Sway Motion

摘要: In this paper, we propose a locomotion planning framework for humanoid robot with an efficient footstep and whole-body collision avoidance planning, which enables the to traverse unknown narrow space while utilizing its body structure like human. The key idea of proposed method is reduce large computational cost by executing global first, has much smaller search space, then performing sequential posture resulting footsteps centroidal trajectory as guide. phase, modify bounding box based on sway motion. This planner obtain appropriate next motion planning. Then, execute prioritized inverse kinematics trajectory, plan each step within less than 100ms at worst. major contribution our paper solving problem increasing enabling adaptive spot in space.