.png){kind=icon}
.png){kind=icon}
Return
Attack Detection and Reconstruction for CPSs via an Improved T-N-L Observer and Zonotopic Analysis
.png){kind=icon}
PRE
AI
B
C
Z
X
赵
DOI：10.1109/tgcn.2026.3708874.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Safety-critical Cyber-Physical Systems (CPSs) are inherently vulnerable to malicious attacks. While detecting these attacks is crucial, handling practical unknown but bounded (UBB) disturbances without triggering conservative thresholds remains a significant challenge. To fill this gap, this paper investigates attack detection, reconstruction, and isolation for safety-critical CPSs under UBB disturbances. In order to represent complex nonlinear systems, the Takagi-Sugeno (T-S) fuzzy model is utilized. First, an improved T-N-L observer integrated with the <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$H_{\infty }$ </tex-math></inline-formula> technique is designed to estimate the system states. To improve estimation accuracy, we develop a decoupling approach that decouples the partial disturbance matrix, effectively attenuating the effect of disturbances on the error system. Second, compared to conventional observers, a strictly tighter interval estimation of the system state is derived through the reachability analysis of the error system. Based on this, an improved attack detection method using residual reachability analysis is proposed to bypass conservative thresholds and reduce computational complexity. Furthermore, the impact of disturbances and the error system on attack reconstruction and isolation is investigated. Finally, simulation results confirm the superiority of the proposed framework. In particular, the proposed method reduces the missed detection rate to 0%, compared with 4.84% for traditional methods, and significantly improves the reconstruction accuracy for both actuator and sensor attacks. For sensor attacks, it achieves an approximate 90.1% reduction in root mean square error (RMSE), while enabling zero-false-alarm isolation for T-S fuzzy systems.
Keywords:
Attack detection and reconstruction
cyber-physical systems
$H_{\infty }$ technique
T-N-L observer
reachability analysis
Journal
I
IF:
6.7
Papers:
1.3K
Citations:
4.3K
Organization
Cited Papers