An Efficient Algorithm for Global Optimization in Redundant Manipulations

Abstract

A method is developed to resolve the redundancy of serial robotic manipulators. The main characteristics of this method are as follows: (a) The solution is conservative (unique); (b) It globally (over the task period) optimizes the objective function; (c) It is computationally very fast. In fact, it is about three orders of magnitude faster than solving for the exact boundary value problem.