Nonlinear Robust Robot Control in Cartesian Coordinate

Abstract

Many assembly tasks such as gear mating, fine polishing or deburring need to use both force and position control in the cartesian coordinate to avoid damaging parts and achieve the desired smoothness. It is ineffective to control force variable in the joint coordinate. In the past, several position control methods in the cartesian coordinate were proposed. However, none of them can maintain fast and accurate motion control in the face of modeling errors and external disturbance. In this paper, we propose a new nonlinear robust control scheme for robot motion control in the cartesian coordinate. The control input of this scheme consists of a nonlinear and a linear part. The nonlinear part decouples and stabilizes the robot dynamics in the cartesian coordinate. The linear part utilizes the robust servomechanism theory to suppress effects of modeling errors or unknown external disturbance. This scheme is used in the control of a two-joint, SCARA-type robot, and simulation results demonstrate that this nonlinear robust control scheme can achieve fast and yet precise tracking control in the cartesian coordinate even at the presence of severe modeling errors.