Frequency-domain design of iterative learning controllers for feedback systems

For a precise tracking control of linear systems, iterative learning controllers are combined with feedback controllers in three different configuration. A multi-step learning controller is proposed in frequency-domain and convergence conditions are derived for each feedback configuration, moreover, their characteristics are compared using Nyquist plots. It is illustrated by simulations that the iterative learning controller enhances the tracking performance of the feedback control system and the multi-step learning controller can be effectively used to attenuate the measurement noises and disturbances.