A Novel Event-Triggered Iterative Learning Model Predictive Control for Linear Systems

In this work, a novel event-triggered iterative learning model predictive control (ILMPC) framework is developed for a class of linear systems subject to input saturation and initial shift. First, a nominal ILMPC strategy is proposed to clearly demonstrate the design philosophy, in which an initial state learning scheme is incorporated to remove the identical initialization condition. Furthermore, in order to reduce the communication and computation loads of the proposed ILMPC approach, two efficient ILMPC strategies, namely, a time-direction event-triggered ILMPC and a hybrid time-iteration-direction event-triggered ILMPC schemes, are developed by considering the event-triggered strategies along the time and iteration axes. It is shown that the proposed event-triggered ILMPC strategies are able to save the communication and computation resources significantly while maintaining the tracking control performance. The convergence of the proposed ILMPC schemes are analyzed rigorously via the contraction mapping methodology and the Lyapunov-like theory, and the effectiveness of the proposed ILMPC method is verified through a numerical example.