A new nonlinear learning control for robot manipulators

Abstract

A new iterative learning control scheme is applied to the trajectory tracking of robot manipulators. The proposed learning control is based on a hybrid, continuous and discrete, Lyapunov argument so that global asymptotic stability can be achieved with respect to the number of trials. This scheme also provides the designer flexibility to design and to implement a learning control for robotic systems by choosing various combinations of robust and learning control parts. The proposed control does not require acceleration measurement, resetting of initial tracking errors and Lipschitz condition. It is also robust in the sense that the exact knowledge of either the nonlinear dynamics or uncertainties of the system is not required except for bounding functions on the magnitude.