Design of a humanoid ping-pong player robot with redundant joints

This study investigates the design of a humanoid ping-pong robot system with a 7-DOFs redundant arm. The design of the arm mechanism, real-time binocular vision system, and a distributed control structure was presented. The methods for ball trajectory prediction and motion planning for the ping-pong racket were proposed. A Jacobian pseudoinverse method considering joint limits and manipulability optimization terms is employed to solve inverse kinematics in order to decrease the counterforce exerting on the humanoid robot. The proposed system is validated by a series of rally experiments between the robot and a human.