Robust Learning Model-Predictive Control for Linear Systems Performing Iterative Tasks

In this article, a robust learning model-predictive controller (LMPC) for uncertain systems performing iterative tasks is presented. At each iteration of the control task, closed-loop state, input, and cost are stored used in design. This article first illustrates how to construct invariant sets safe policies exploiting historical data. Then, we propose an LMPC design procedure, where data generated by at $j$ next notation="LaTeX">$j+1$ . We show that procedure allows us iteratively enlarge domain policy, it guarantees recursive constraints satisfaction, input-to-state stability, performance bounds certainty equivalent system. The use different feedback along horizon key element proposed effectiveness scheme illustrated on linear system subject bounded additive disturbances.