Suppression Method of Disturbance Torque Due to System Drift via an Observer in the Robot Control System

Abstract

The degradation of performances in the robot manipulator control system results from disturbance torques due to the system drift, frictions in the gears, Coriolis, gravity loading, and centrifugal effects of the link. The torque due to the system drift is closely related to the difference between ground levels and the variation of characteristics of the control system, while the others depend on the manipulator's physical parameters, instantaneous joint configuration, joint velocity and acceleration, and the load it is carrying. In this paper, an estimation algorithm of the system drift in the robot manipulator control system is proposed, which is based on the gradient search method. This algorithm is executed in the steady state for the elimination of the system drift because the variation of the characteristic of the control system due to the change in temperature, supply voltage, and etc. are slow compared with the motion of the robot manipulator. The proposed algorithm is implemented on the control system for a robot manipulator with six degrees of freedom, which has an 80386/80387 as a main processor and a TMS 320C30 as a processor for motion control. Experimental results show that the proposed suppression method of the system drift via an observer can be effectively used in controlling robot manipulators with high dynamic performance at both the transient and steady states.