Feedback-aided minimum joint movement (FAMJM) scheme for redundant manipulators with simulations and experiments

To keep the minimum joint movement of redundant manipulators over the task duration, a feedback-aided minimum joint movement (FAMJM) scheme is proposed for the redundant robots. The optimal index of the FAMJM scheme, in essence, is to minimize the drift between the instantaneous joint state and the initial joint state, which is subject to the robot forward-kinematics equation, the joint-angle limits and the varying joint-velocity limits. Besides, to improve the accuracy of planning and control, a position-error feedback is incorporated readily into the scheme via a neural-dynamic method. Based on Lyapunov theory, the convergence performance of the feedback-aided scheme is proved. The FAMJM scheme is then reformulated and unified as a quadratic program (QP) subject to an equality constraint and a bound constraint, which is solved by a discrete-time QP solver. Furthermore, this FAMJM scheme is simulated and implemented on an actual planar six degrees-of-freedom (6-DOF) push-rod joint (PRJ) manipulator to write English letters, i.e., to track “M” and “P” shaped paths. Simulation and experiment results verify the efficacy and accuracy of the FAMJM scheme and its corresponding QP solver. Moreover, this work shows an interesting trend that is becoming more and more important in the field of automatic control and robotics; i.e., to combine planning and reactive control methodologically and systematically, in view of the real-time requirement and the possibility that unexpected events arise from time to time.