Unification and comparison on bi-criteria velocity, acceleration and torque minimization illustrated via three-link planar robot arm

Abstract

In this paper, the latest result that the bi-criteria redundancy-resolution schemes at different levels (i.e., the joint-velocity, - acceleration and - torque levels) can be unified into a quadratic-programming (QP) problem formulation, is presented and summarized for redundant robot arms with joint physical limits. In addition, comparison on the three schemes is presented for further investigation. The presented QP-based formulation, subject to equality, inequality and bound constraints simultaneously, is then solved online via a primal-dual neural network based on linear variation inequalities (LVI). Computer-simulation results based on PUMA560, PA10 and multi-link planar robot arms (specially, a three-link planar robot arm in this paper) have demonstrated the efficacy of the bi-criteria neural-weighting schemes, as well as the resultant QP-based unification of robots' redundancy resolution.