Consensus-based distributed estimator of cyber–physical systems against false data injection attacks: A two-layer estimation approach

Abstract

The problem of consensus-based distributed estimation for cyber–physical systems (CPSs) against false data injection (FDI) attacks is taken into account. Different from most research perspectives, the FDI attack is modeled with high randomness in value, target, and period. The CPSs architecture with a group of sensors and estimators is considered in the distributed estimation. A reconstructed measurement, depending on the combination of an attack detector and estimation results, is established to obtain accurate measurement data when the transmitted measurements are falsified by FDI attacks. The two-layer estimator, which includes measurement fusion estimation and interactive consensus estimation, is advanced to achieve secure consensus-based distributed estimation. In contrast to current approaches, due to accumulating measurement data from different sensors, and adopting the appropriate consensus gain matrix obtained by the Lyapunov function of estimation error, the proposed approach can realize consensus distributed estimation while requiring fewer communication resources. In addition, numerical simulations of the 2-dimension tracking system are provided to illustrate the effectiveness of the proposed method.