Data-driven H∞ controller tuning for sensitivity minimisation

By achieving low sensitivity, it is possible to suppress perturbations of a plant and disturbances. In model-based controller syntheses, the solution of sensitivity minimisation is analytically obtained through a mathematical model of a plant. However, a plant model which is enough accurate to design an appropriate controller is not always available. Moreover, if there exists a large uncertainly between a plant model and an actual plant, the designed controller might not show desired performance and at worst the system would be destabilised. This paper focuses on data-driven controller tuning methods because they require no plant model and skip system identification. The controller is designed or tuned by using only input/output data. This paper proposes a design method to achieve sensitivity minimisation, which is a typical controller synthesis problem in model-based approaches, by using only input/output data.