Privacy‐Preserving Asymmetric Bipartite Consensus for Multi‐Agent Systems Under Cyber Attacks

ABSTRACT This paper addresses the privacy‐preserving and security control issues of multi‐agent systems (MASs) in cooperative‐competitive networks subject to multiple cyber attacks. Firstly, a unified attack model is developed to represent the simultaneous occurrence of Denial‐of‐Service (DoS) and False Data Injection (FDI) attacks in MASs. Secondly, considering the existence of privacy‐theft attacks, a sub‐state exchange (SSE) approach is introduced to protect the privacy of the agents' initial state from being stolen. Then, based on the Kalman filter, a state estimation‐based security protection mechanism is proposed under the joint attacks. The mechanism ensures that the agents use secure state values to achieve asymmetric bipartite consensus for MASs. Finally, simulation experiments are conducted to validate the effectiveness of the proposed methods.