Perfect Tracking of Repetitive Signals for a Class of Nonlinear Systems

Abstract Perfect tracking of the output of a class of nonlinear systems that has a unique response for a given input and is subject to repetitive reference input is considered in this paper. A conditional learning scheme guaranteeing sufficient knowledge can be learned iteratively to improve the input to achieve perfect tracking is proposed. The sufficient condition for monotonic convergence of the input sequence and the choice of the learning gains are given. The tracking performance of the proposed scheme is illustrated by a simulated example.