EXPERIMENTAL IDENTIFICATION OF FRICTION AND ITS COMPENSATION IN PRECISE, POSITION CONTROLLED MECHANISMS

There are two primary objectives of this paper. The first is to present methodologies developed for experimentally determining accurate models for the nonlinear friction inherent in most mechanisms. The second objective is to present alternative closed-loop controller strategies for decoupling the effect of friction in order to improve positioning accuracy. The identification methodology is novel in the manner in which it extracts the nonlinear friction properties from the closed-loop errors via an iterative signal processing technique. The paper is based on both theoretical modeling and on a practical position control problem that was substantially resolved in developing the methodologies. The application was a robotic gripper with a highly preloaded rack-and-pinion mechanism. The paper provides both measurement and controls design methodologies to help systematically circumvent the problems of nonlinear friction in precise, position-controlled mechanisms.