Vulnerability Analysis of Distributed State Estimator Under False Data Injection Attacks

This paper focuses on the vulnerability and strict vulnerability of distributed state estimators under false data injection (FDI) attacks, where adversaries aim to exert unbounded effects on the estimation error dynamics by injecting malicious data into sensor nodes, communication links, or both. In particular, a distributed system is characterized as vulnerable (or strictly vulnerable) if there exists an unbounded FDI attack that leads to bounded changes (or no changes) in residuals. By utilizing invertibility theory and carefully designed attack sequences, we establish the conditions for systems to be (strictly) vulnerable, based on different attack scenarios. Additionally, we provide a comparative analysis to evaluate the varying impact of different attacks on system security. Finally, a three-area grid system model is presented to illustrate the validity of the theoretical results.