Multirate State Tracking for Improving Intersample Behavior in Iterative Learning Control

Abstract

Iterative learning control (ILC) enables highperformance output tracking at sampling instances for systems that perform repetitive tasks. The aim of this paper is to present a state tracking ILC framework that reduces oscillatory intersample behavior often encountered in output tracking ILC. A multirate inversion is performed to achieve state tracking in ILC, which achieves perfect state tracking at every $n$ samples, where $n$ denotes system order. Consequently, this improves the intersample tracking performance. Moreover, convergence criteria based on frequency response data are derived and exploited in a design approach. The approach is successfully applied to a motion system confirming improved intersample tracking accuracy compared to standard frequency domain ILC.