Co-Design of Data-Driven Attack Observer and Resilient Controller for Unknown NCSs Under Hybrid Attacks

This paper proposes a data-driven estimation and control co-design method for a class of unknown linear network control systems (NCSs) under denial-of-service (DoS) and false data injection (FDI) attacks simultaneously. Firstly, a new piecewise observer is designed to provide the real-time estimates of the unavailable system state and the unknown FDI attack signal in the presence of DoS attacks. Under the assumption that the system matrices of the considered NCSs are known in advance, model-based exponential stability criteria are established by utilizing attack-dependent time-varying Lyapunov function (TVLF) method. Secondly, to extend the obtained model-based stability criteria under hybrid attacks to data-driven paradigm, a new technical lemma, called data-driven system representation under DoS attacks is proposed. Based on this technical lemma, data-driven stability conditions of unknown NCSs under hybrid attacks are derived in terms of data-based linear matrix inequalities. Thirdly, a security controller design method is proposed within the framework of data-driven methodology. Finally, the effectiveness of the proposed data-based security control strategy for unknown NCSs under hybrid attacks is verified through numerical simulation.