Data-driven flutter suppression of noised airfoil models: model identification and robust sliding mode control designed with multi-level control surfaces

The robust control of complex systems poses a persistent challenge in modern control theory. While several model-free robust control strategies have been developed, they often falter in the face of large perturbations. To address this issue, this paper introduces an integrated approach that combines data-driven model identification with robust control law design paradigm. The resulting control law demonstrates robustness against both large noise perturbations and model identification errors. Furthermore, it offers advantages such as strong interference resistance, smaller control energy requirement, and enhanced stability of controlled states when compared to traditional PID control strategies. Notably, our method yields valuable insights for tackling complex control issues in general multi-degree-of-freedom Lagrangian mechanical systems, particularly those characterized by uncertain governing equations and large discontinuous perturbations.