Transitive Iterative Learning Control for Switched Systems With Performance-Driven Switching Laws

There is a tricky problem in iterative learning control (ILC), where network delays during data transmission may lead to mismatches between the updated values of controller and required parameters, thereby disrupting the experience inheritance mechanism. Multimode switched systems can describe a wider range of complex processes, but most of the existing switching laws are fixed and lack flexibility. To address these issues, on the one hand, a transitive ILC (T-ILC) method is proposed to mitigate the effect of data network delay; on the other hand, a novel tracking performance-driven switching law (TPD-SL) is proposed. T-ILC can ensure the correct order of data through timestamps and memory identification. TPD-SL can ensure that the system switches to the subsystem with better tracking performance according to the change rate of the tracking error. Furthermore, we adopt an event-triggered communication method to save network computing resources. The stability of the system is demonstrated under the T-ILC and TPD-SL, and the relationship among network delay, event-triggering, and synchronous switching (controller switching aligned with system mode transition) versus asynchronous switching (controller switching with delayed mode transition) is analyzed. Numerical simulation results confirm that the method is effective.