Memory-driven event-triggered bipartite consensus for multi-agent systems under dynamic leader and scaling attacks

This study explores the memory-event-driven bipartite tracking consensus problem in linear multi-agent systems under the influence of a scaling attack. In the model, the leader's control input is a non-zero and unknown variable to all followers. We examine scaling attacks within the communication networks characterised by cooperative and competitive interactions, where such attacks can manipulate the original signal by downsizing, upsizing, or zeroing, thereby subsuming both deception and denial-of-service attacks. To address this, a dynamic memory-event-based control mechanism is proposed, which balances network resource usage and system performance by dynamically adjusting the threshold function based on historic triggered signals. Additionally, the control protocol accounts for antagonistic interactions among agents, ensuring the secure bipartite tracking consensus through carefully designed gain matrices and control parameters. Finally, theoretical results are supported by two numerical examples.