Adaptive integrated security control of ICPS based on data-driven and mechanism analysis fusion method

For a class of industrial cyber-physical systems (ICPS) where stealthy false data injection (FDI) and actuator faults coexist, the data-driven and mechanism analysis methods are combined to study the integrated security control and communication co-design problem. Firstly, a discrete event-triggered communication scheme (ADETCS) obeying exponential adaptive law is designed, and an adaptive ICPS framework is constructed to withstand both network FDI attacks and physical component faults. Secondly, based on data-driven technology, the prediction model PSO-CatBoost of FDI attack is established through optimization to accurately reconstruct and compensate the attack. Then, by means of less conservative techniques such as the augmented Lyapunov-Krasovskii functional and improved affine Bessel-Legendre inequality, the solution methods of the robust observer and integrated safety controller are deduced. Finally, the effectiveness of the proposed method is verified through the example of a quadruple tank. The results show that the deep integration of data-driven stealthy FDI reconstruction compensation and mechanism analysis compensation error suppression, the effective tolerance of active-passive cooperation to network atacks, the combination with active fault tolerance, and adaptive adjustment of trigger parameter with the change of system behavior under the ADETCS can significantly increase ICPS dual security defense capability, save more network resources. © 2024 Northeast University. All rights reserved.