In the first part of this article a parametric framework for Iterative Learning Control (ILC) was developed. Its validity for different classes of systems and for various parametrisations was shown. In the second part ILC in the frequency domain is discussed in more detail. An algorithm based on pole placement is introduced. It is easy to implement and shows good robustness against model uncertainties. Learning law and conventional control can be combined by two-stage pole placement. A suitable application for this cyclic control problem are test benches for materials and machine parts testing. The proposed algorithm is implemented on pneumatic and hydraulic test stands with good experimental results.
Print ISSN: 0178-2312
Volume: 55, 04/2007
Pages: 180 - 189