System Identification and Low-Order Optimal Control of Intersample Behavior in ILC

Although iterative learning control (ILC) algorithms enable performance improvement for batch repetitive systems using limited system knowledge, at least an approximate model is essential. The aim of the present technical note is to develop an ILC framework for sampled-data systems, i.e., by incorporating the intersample response. Hereto, a novel parametric system identification procedure and a low-order optimal ILC controller synthesis procedure are presented that both incorporate the intersample behavior in a multirate framework. The results include i) improved computational properties compared to prior optimization-based ILC algorithms, and ii) improved performance of sampled-data systems compared to common discrete time ILC. These results are confirmed in a simulation example.