.png){kind=icon}
.png){kind=icon}
Return
LTSFA-based attack detection and isolation algorithm for remote state estimation
.png){kind=icon}
PRE
AI
DOI：10.1016/j.ejcon.2026.101513.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
This paper investigates a security issue in Cyber-Physical Systems (CPSs), focusing on the performance of multi-sensor remote state estimation under the optimal innovation-based deception attack. We consider a scenario where a system equipped with N sensors monitors its state, with these sensors transmitting their measurements to a remote estimator via a wireless communication channel. The attacker is assumed to be capable of launching a linear attack to alter some sensor measurements, aiming to maximize the degradation of estimation performance while evading detection by a chi 2 detector. To mitigate the attack, a remote state estimator is employed, incorporating an attack detection and isolation algorithm based on Long-Term Dependency Slow Feature Analysis (LTSFA). This algorithm enhances the estimation performance by identifying attacks and recovering the compromised measurements. We derive recursive expressions for the error covariances of the remote state estimator both with and without the LTSFA-based attack detection and isolation algorithm under the optimal innovation-based deception attack. The proposed algorithm's effectiveness under the given conditions is theoretically demonstrated and further validated through simulation examples.
Keywords:
CPS security
LTSFA
Remote state estimation
Strictly stealthy attacks
Journal
IF:
2.6
Papers:
323
Citations:
2.5K