Convergence Rates for Learning Control Algorithms-A Repetitive Process Theory Approach

Abstract

This paper presents significant new results on the stability and convergence properties of a general class of iterative learning control schemes derived using two-dimensional(2D) systems theory. These results apply for a general learning law which (explicitly) uses information from previous iterations or trials. A key feature of these results is that they are expressed in terms of standard linear systems theoretic properties, such as relative degree and the location of the zeros.