Iterative learning control

Abstract

The chapter presents original research results in the area of nonlinear iterative-learning control. We propose a novel ILC scheme developed using neural networks. The following two cases are described: dynamic and static learning controllers and in both cases the controller is designed in such a way as to minimize the tracking error. This task is accomplished by an appropriate training of the neural controller after each repetition of the control system. Additionally, the chapter contains both the stability and convergence analysis of the proposed nonlinear ILC. The portrayed control strategies are tested on the examples of a pneumatic servomechanism and a magnetic suspension system.