Intermediate Estimator-based Attack Reconstruction for Cyber-Physical Systems

This paper is concerned with the attack reconstruction problem for cyber-physical systems (CPSs) under false data injection (FDI) attacks, where both of sensor attacks and actuator attacks are taken into consideration. While existing attack estimation methods typically require satisfying observer matching conditions, such conditions are often challenging to meet in many systems. To overcome this limitation, a reconstruction method based on intermediate variables is developed to estimate the attack occurring at sensors and actuators. The developed method has low computational complexity and high scalability. It is proved that the states of the error system are uniformly and finally bounded. Simulation examples demonstrate the effectiveness of the proposed method.