Data-driven parameter tuning of IMC for unstable plants

This paper proposes a data-driven controller parameter tuning approach to the modified internal model control (IMC), which was proposed by Yamada in [9], for unstable plants. Here we clarify how to use fictitious reference iterative tuning (FRIT) to tune parameters of the modified IMC controller with only one-shot experimental data. The approach in this paper simultaneously yields a desired controller and a mathematical model of the controlled plant. In addition, we also analyze and qualify the robustness of the system in the presence of modeling error.