Combining Prior Knowledge and Data for Robust Controller Design

We present a framework for systematically combining data of an unknown linear time-invariant system with prior knowledge on the matrices or uncertainty robust controller design. Our approach leads to matrix inequality (LMI) based feasibility criteria which guarantee stability and performance robustly all closed-loop systems consistent available data. The design procedures rely combination multipliers inferred via learnt from measured data, where latter novel unifying disturbance description is employed. While large parts paper focus input-state measurements, we also provide extensions output-feedback noisy input-output against nonlinear uncertainties. illustrate through numerical examples that our provides flexible simultaneously leveraging thereby reducing conservatism improving significantly if compared black-box approaches data-driven control.