Proposing, QP-unification and verification of DLSM based MKE-IIWT scheme for redundant robot manipulators

To remedy the problem of joint-variable oscillation and instability as well as the non-zero final joint-velocity phenomenon arising in the conventional inertia-inverse weighted torque (IIWT) minimization scheme, in this paper, a novel scheme is proposed and investigated for redundant manipulators with joint-physical limits (i.e., limits of joint angle, joint velocity, joint acceleration and joint torque) considered on the basis of different-level simultaneous minimization (DLSM) technique. Such a scheme combines the velocity-level minimum kinetic energy (MKE) scheme and the torque-level IIWT minimization scheme via two weighting factors, which guarantees the stability and final joint-velocity of motion to be near zero. Since the IIWT scheme is an optimization scheme with global characteristics and the MKE scheme is a local counterpart of the global kinetic energy minimization scheme, the proposed DLSM based MKE-IIWT scheme (or termed, the DLSM-MKE-IIWT scheme) possesses the advantages of both local and global optimization schemes. Based on Zhang-equivalency lemma and the relation between joint torque and joint acceleration, such a DLSM-MKE-IIWT scheme is finally reformulated as a unified quadratic program (QP). Computer simulation performed on a 3 degrees-of-freedom (DOF) redundant robot manipulator comparatively verifies the effectiveness and superior performance of the proposed DLSM-MKE-IIWT scheme.