Iterative learning control optimization strategy for feedback control systems with varying tasks

Abstract

Iterative learning control (ILC) combined with feedback control is a common approach to repetitive systems with external disturbances, as it enables high tracking performance and guarantees time-domain stability. However, the variation of the reference trajectory in practical repetitive operations often degrades the control performance. To this end, this paper develops a feedback-based ILC to transfer the experience of repetitively operating a certain task to a brand new task without restriction on its time duration. This two-dimensional (2-D) design employs a parallel structure, where the ILC and the feedback controller are designed separately to achieve performance optimization. Then, the feedback plus feedforward controller is integrated into a new feedback controller with learning-based parameters. The convergence and robustness analysis of the design is given. Finally, numerical simulation experiments of a DC motor position control system verify the proposed scheme's effectiveness and robustness.