Robust Iterative Learning Control for Vibration Suppression of Industrial Robot Manipulators

Abstract

Vibration suppression is of fundamental importance to the performance of industrial robot manipulators. Cost constraints, however, limit the design options of servo and sensing systems. The resulting low drive-train stiffness and lack of direct load-side measurement make it difficult to reduce the vibration of the robot's end-effector and hinder the application of robot manipulators to many demanding industrial applications. This paper proposes a few ideas of iterative learning control (ILC) for vibration suppression of industrial robot manipulators. Compared to the state-of-the-art techniques such as the dual-stage ILC method and the two-part Gaussian process regression (GPR) method, the proposed method adopts a two degrees-of-freedom (2DOF) structure and gives a very lean formulation as well as improved effects. Moreover, in regards to the system variations brought by the nonlinear dynamics of robot manipulators, two robust formulations are developed and analyzed. The proposed methods are explained using simulation studies and validated using an actual industrial robot manipulator.