Coordinated planning and control for multifingered robot hand

A coordinated planning method and control strategy for multifingered robot hand is proposed. We divide a typical grasp and manipulation task into two stages: an off-line planning stage and an online control stage. In the planning stage, a finite motion planning method and an optimization problem is formulated at first to find the fingertip locations, finger joint trajectory and distribution of the fingertip forces, then a position and force coordinated planning method is developed. In the control stage, a generalized mater-slave coordinated control method is designed and discussed in details, which includes a force/position hybrid controller for the masters and new revised coordinated force and position hybrid controller for the slaves. It is very useful to compensate errors of the forces and positions in real-time and keep the grasping and manipulation stable all the way. Simulation of the planning result and experiment of grasping and moving a ball with the GFKD three-finger robot hand show the applicability and validity of the proposed scheme.