A data-based approach for designing adaptive controllers for unknown systems

This paper presents an approach for designing gain-scheduled controllers for unknown systems using a set of measured data. In most control approaches, controllers are designed using a mathematical model, which is often obtained on the basis of some simplifying assumptions. Thus, controllers designed through model-based methods may result in degradation of the desired closed-loop performance due to complex dynamics. Hence, the proposed approach is motivated by the fact that: 1) errors associated with the modeling process are avoided since no mathematical model is required for the controller design, 2) the designed adaptive controllers are able to ensure desired performance specifications for the plant operated not only at a given operating point but over a range of operating conditions, and 3) the controller structure can be selected a priori. A simulation example to control a water heating system is presented to validate the proposed method.