A new framework of learning control for a class of nonlinear systems

This paper illustrates a new nonlinear learning control design based on Lyapunov's direct method. The design is applicable to the class of nonlinear systems consisting of finite cascaded subsystems in performing repeated tasks. A class of difference or difference-differential learning laws is proposed. It is shown that, under a difference learning control, the class of nonlinear systems is guaranteed to be asymptotically stable with respect to the number of trials. For better rejection of measurement noise, the difference-differential learning law can be applied to yield arbitrarily good accuracy. The proposed approach provides closed-form expressions of learning controls, and it gives the designer much flexibility in choosing various combinations of feedforward and learning control parts.