Adaptive LQ-Based Computed-Torque Controller for Robotic Manipulator

Abstract

In this paper, the motion control of robotic rigid two-link manipulator using an adaptive LQ-based computed-torque controller is studied in the presence of unknown parameters in the dynamic model. This controller consists of a linear quadratic computed-torque one, and an adaptive update rule for estimating the parameters. The linear quadratic regulator (LQR) presents an optimal solution to minimize the tracking error. The advantage of the adaptive mechanism in this controller is that the accuracy of a manipulator carrying unknown loads will improve with time because the adaptation rule continues getting feedback signal from the tracking error.