Consensus of Heterogeneous Multi-agent Networks Under Sparse Attacks by Integral Type Observer

This paper is concerned with output consensus of heterogeneous multi-agent systems under sparse attacks. The leader agent is only available to a small fraction nodes. To estimate the leader's signal, a distributed observer is proposed, which is of the integral type and possesses good robustness and high convergence rate. Unlike the existing literature on distributed observer, this paper introduces a novel event-triggered strategy that utilizes the agents' measurement error state instead of the norm of the state errors. It is theoretically and numerically demonstrated that the triggering times and communication loads between agents can be significantly reduced by using this strategy, and a positive lower bound on the triggering time interval is also computed. Furthermore, an adaptive scheme with pinning strategy is considered to ensure the output consensus with uniformly ultimately bounded in the presence of false data attacks. Finally, a numerical example is given to illustrate the effectiveness of the results.