Secure and Privacy-Preserving Consensus for Multi-Agent Networks under Deception Attacks

This paper studies the distributed consensus problem for discrete-time multi-agent networks under deception attacks. For the privacy of agent initial state value, an additive secret sharing method is adopted at the initial time of the system. To overcome the adversarial effects of deception attacks on the network, a distributed secure method is used to identify and remove the tampered values from the attacks, and then a control protocol is designed so that all normal nodes in the network are able to converge to the average value of the initial values. Theoretical analysis shows that the proposed method can effectively resist the impact of deception attacks and achieve the accurate average consensus of the system. Numerical examples are also given to demonstrate the effectiveness of the results.