Abstract
This work investigates the problems of attack detection and reconstruction for discrete-time linear cyber-physical systems (CPSs) with unknown-but-bounded disturbances and measurement noise. First, an improved unknown input observer (UIO) with fewer synthesis conditions is proposed to estimate the system state. To enhance the estimation accuracy, a decoupling method is proposed to decouple the disturbance matrix and exclude the effects of disturbances on the error system. Then, to reduce the computational complexity, an improved attack detection method based on the reachability analysis of the residual is proposed, which requires less computation than the existing attack detection methods. Moreover, the impact of disturbances and measurement noise on the process of attack reconstruction and isolation is analyzed. The accuracy of attack reconstruction is improved through the decoupling approach and $H_{\infty } $ technology. Finally, the simulation results verify the effectiveness and feasibility of the proposed approach.
Funding Information
  • National Natural Science Foundation of China (62376170, 62303287)
  • Natural Science Foundation of Shandong Province (ZR2024QF017)

This publication has 50 references indexed in Scilit: