Event-based model-free adaptive consensus control for multi-agent systems under intermittent attacks

This paper investigates the distributed event-triggered consensus tracking control problem for nonlinear multi-agent systems in the presence of intermittent attacks on both intra- and inter-agent communication channels. Intermittent attacks are characterised by their ability to manipulate the channel transmission factor to disrupt the reliability of communication. First, to address the problem that the agent model knowledge is unknown, a dynamic linearisation method is introduced to transform the nonlinear agent models into equivalent linear models that depend only on the agents' input and output data. Then, a dynamic event-triggered mechanism is developed to reduce communication transmission, which is based on individual agent information rather than consensus errors to avoid the impact of attacks occurring in the communication between agents. Building upon the attack model and the equivalent linear data model, a distributed model-free adaptive control scheme with dynamic event-triggering is devised to ensure that the consensus errors of all agents are ultimately bounded, even when intermittent attacks occur. The proposed scheme's effectiveness is demonstrated through numerical simulation.