Data-Based Iterative Learning Control: A Nonconservative Approach via LMI Techniques

This paper aims to propose a data-based iterative learning control (ILC) framework that addresses the tracking issues without imposing additional assumptions on the sufficiency of sampled data. By introducing the concepts of k-state system and robust k-stability, we establish a connection between time-domain tracking issues and iteration-domain k-stabilization. Moreover, with the application of some helpful linear matrix inequality (LMI) techniques, we convert the data-based ILC synthesis into solving equivalent LMI conditions. As a result, the tracking error is recursively corrected and satisfied tracking performances are achieved by leveraging as few sampled data as possible. To demonstrate the effectiveness of the proposed ILC framework, illustrative simulations on an injection molding process are also provided.