Proper closed-loop specifications for data-driven model-reference control

In control applications where finding a model of the plant is costly and time consuming, direct data-driven approaches represent a valid alternative for the design of model reference controllers. However, the selection of a proper reference model within a model-free setting is known to be a critical task, as such a model typically plays the role of a hyperparameter. In this work, we extend the existing theory so as to compute both a reference model and the corresponding optimal controller parameters from data to satisfy given behavioral bounds on the desired closed-loop performance. The effectiveness of the proposed approach is illustrated 011 a benchmark simulation example.