Task-space tracking of robot manipulators via internal model principle approach

Abstract

This paper presents an internal model-based adaptive control method for uncertain robot manipulators, addressing the task-space asymptotic tracking problem. In the proposed scheme, the reference trajectory is assumed to be a multi-tone sinusoidal signal with unknown amplitude and frequency parameters, and the robot kinematic and dynamic parameters are considered uncertain. Unlike existing approaches that assume the reference signals are directly measurable, we propose an error feedback controller that requires only measurements of the task-space tracking error, joint position, and joint velocity. Specifically, based on the internal model principle, an internal model-based dynamic compensator is developed to reproduce the reference signals. By using the parameter linearity properties of the robot kinematics and dynamics, adaptive laws are derived to handle the unknown parameters. The stability of the closed-loop system and the asymptotic convergence of the tracking error are analyzed using output stability concepts. The effectiveness of the proposed approach is validated through numerical simulations with a three-DOF manipulator.