Motion Planning for Ricochetal Brachiation Locomotion of Bio-primitive Robot

摘要

Brachiation is a special form of locomotion used by primates swinging from branches to branches with their upper limbs. Primates are capable of adapting their behavior according to different situations. Among them, Ricochetal Brachiation is quite a sophisticated one which can fly primates across a very long distance. Ricochetal Brachiation requires precise coordination of two brachiating and one free-flying phases. The tight coupling, nonlinear and underactuated characteristics bring great challenges on motion flexibility and controller robustness. This paper intends to solve those two problems by planning feasible and flexible motion trajectories on-the-fly, which deals with the planning problem in ricochetal brachiation process with different distance, attitude and energy variations. The key of ricochetal motion planning is to determine the free fly trajectory. However, tight coupling of three phases introduces too much uncertainty. To reduce the search space, geometric, translational, rotational and brachiation pattern constraints are explored. The rotational constraints brought by momentum conservation is analyzed to discover the internal bounding conditions and landing patterns. Natural constraint planner(NCP) is proposed to organize the conditions and find feasible ricochetal motion. Simulation results show effectiveness and flexibility of the proposed methodology.