Event-Triggered Security Consensus of Continuous-Time Multi-Agent Systems Against Complex Cooperative Attacks

This article considers event-triggered security consensus of continuous-time multi-agent systems (MASs) subject to complex cooperative attacks. First, false data injection attacks (FDIAs) and denial-of-service (DoS) attacks are modeled under a unified framework through the Bernoulli process, which provides a comprehensive understanding of the challenges posed by these attacks in MASs. Second, an improved event-triggered mechanism (ETM) based on aperiodic sampling is constructed, which avoids the Zeno phenomenon and has the potential to improve the efficiency and resilience of MASs. In addition, it can be combined with detection signals to address the challenge of determining trigger times under DoS attacks, which is an important contribution. Third, to overcome the impact of complex cooperative attacks, this article analyses the security consensus of MASs in three steps by combining the characteristics of trigger time and attack time. Based on this, security consensus is guaranteed. Finally, a practical example is used to demonstrate the feasibility of the proposed ETM.