Detection of False Data Injection Attack for Interconnected Systems

Decentralized detection of network attacks for interconnected systems is addressed in this paper. Most of the existing researches on decentralized attack detection are focused on multi-agent systems and no related research has been done on systems interconnected over graph with complex spatial terms. This paper will discuss from the following three aspects. Firstly, to isolate the effect of neighbor subsystems, the problem of dealing with spatial terms is analyzed. The unknown input terms are introduced to decouple the interconnected structure of the interconnected system, which will be eliminated from measurement output equation based on the generalized inverse theory. Secondly, the sufficient conditions for the existence of the decentralized attack detector and attack detection principle is given. Detection threshold value is calculated by the estimation error under no attack. Finally, detectability analysis is given. Two illustrated examples are provided to demonstrate the effectiveness of the proposed strategy.