Iterative learning control (ILC) is a control scheme that takes advantage of the cyclic nature of a process, in order to improve the plant output from cycle to cycle. As the ILC uses the measured system output trajectory for calculation of an improved input trajectory, disturbances in the measurements highly affect the performance of this control scheme. The integration of time variant filters into the ILC-loop is a suitable method for removing these disturbances and thus improves the quality of control. In this paper, we present a method to design time variant filters based on wavelet decompositions of the control signals. The article describes how the filtering methods are integrated into the context of discrete time ILC. The theoretical results are complemented by results obtained from a real world processing plant.
展开▼
