Joint Self-Tuning With Cartesian Setpoints

Abstract

A joint coordinate self-tuning manipulator control method is presented which uses Cartesian setpoints. The method is capable of both position and hybrid control. Position and force errors are transformed from Cartesian coordinates to position and force errors at the joints. The position and force errors at each joint are combined into one hybrid error that is eliminated using pole-placement self-tuning. Real time position and hybrid control results are given. No prior knowledge of manipulator or load dynamics is required and real time control results show that the goal of consistent control with changing load dynamics is achieved. The major cause of error in position and hybrid control is the large friction effects in the joints.