Design of feedback gain in feedback-based iterative learning control

Abstract

Feedback loops have been introduced in iterative learning control (ILC) to improve the performance in the time domain. This in turn affects the convergence performance in the iteration domain. Such an impact on ILC performance has not been rigorously analysed in the literature. In this paper, the concept of dynamic influence in the iteration domain is introduced to characterise the impact of feedback in the convergence of ILC algorithms. An optimal feedback gain is thus designed in order to minimise such an impact with the help of nominal model, leading to an improvement of the transient response in the iteration domain. The result also reveals a clear performance trade-off between the robus..