Distributed Secure Fusion Estimation With Lossless Data Compression Against Stochastic Event-Based Attacks in Power Systems

This paper investigates the issue of distributed secure fusion estimation in power systems under stochastic event-based (SEB) attacks. A lossless data compression method utilizing full rank decomposition transform is proposed to reduce redundant sensor data, thereby alleviating network transmission load. To further save communication resources and prolong network lifetimes, sensors use a stochastic event-triggered mechanism (SETM) to transmit compressed data to local estimators. However, SETM can be exploited by SEB attackers. We first derive local estimator iterations using Bayesian inference in both normal and attack scenarios. Moreover, the accuracy of fusion estimates can be guaranteed through the codesign of information compensation and an indexed transmission strategy that counteracts SEB attacks. Finally, this paper provides an upper bound for fusing error covariance under a diagonal matrix weighted fusion scheme, effectively balancing fusion estimation performance and resource demands. All theoretical results have been illustrated by the 5-Generator and 8-Line power system.