A new optimality based adaptive ILC-algorithm

In this paper a new optimality based adaptive Iterative Learning Control (ILC) algorithm is proposed. It can be seen as an extension of the feedforward algorithm u_k+1(t)=u_k(t)+γe_k(t+1), which is known to suffer from poor transient behaviour. It is in fact shown that this extended algorithm gives guaranteed monotonic convergence, which is a considerable improvement when compared to the algorithm. Furthermore. the extended algorithm contains a simple tuning knob that can be used to select a suitable convergence rate. The theoretical findings are illustrated with simulations, which support the theory presented in this paper.