Selecting a suitable grasp motion for humanoid robots with a multi-fingered hand

This paper discusses grasp planning of a multi-fingered hand attached at the tip of a humanoid robotpsilas arm. Our planner can select a different grasping style even for the same object if a position/orientation of the object is changed. Also, if the planner cannot find a feasible grasping posture with arm/hand kinematics, a humanoid robot tries to use the whole body motion. These functions are necessary for realizing the robust grasp planning. Our planner defines convex models on both the object and each grasping style. By considering geometrical relationship among these convex models, we determine several parameters needed to define the final grasping configuration. The effectiveness of the proposed method is confirmed by several numerical examples and experimental results.