Adaptive LQ control of robot manipulators

Abstract

LQ control is a well known optimal control for linear systems without uncertainties. In this paper, however, an adaptive control scheme is proposed for robot manipulators with LQ performance. The robot dynamics is inherently high nonlinear and furthermore we assume that the robot parameters are all unavailable. We propose to use the function approximation technique to achieve effective feedback linearization. This ensures that the closed loop system exhibits a linear dynamics where the LQ controller can thus be designed. To verify the closed loop stability and boundedness of internal signals, the Lyapunov-like technique is applied. A 2-D robot manipulator is constructed to justify the effectiveness of the proposed scheme experimentally.