Minimum-Energy Redundancy Resolution of Robot Manipulators Unified by Quadratic Programming and its Online Solution

This paper presents the latest result regarding the unification of minimum-energy redundancy resolution of robot manipulators via a quadratic program. The presented quadratic programming (QP) formulation is general in the sense that it incorporates equality, inequality and bound constraints, simultaneously. This QP formulation covers the online avoidance of joint physical limits and environmental obstacles, as well as the optimization of various performance indices. Every term is endowed with clear physical meaning and utility. Motivated by the real-time solution to such robotic problems, four QP online solvers are briefly reviewed. That is, standard QP optimization routines, compact QP method, dual neural network as a QP solver, and state-of-the-art LVI - based primal-dual neural network as a QP solver. The QP- based unification of robots' redundancy resolution is substantiated by a large number of computer simulation results based on PUMA560, PA10, and planar robot arms.