Fully Distributed Secure State Estimation for Nonlinear Multi-Agent Systems Against DoS Attacks: An Edge-Pinning-Based Method

This paper investigates the fully distributed secure state estimation problem of nonlinear multi-agent systems (MASs) against denial-of-service (DoS) attacks. The design of state observer depends on the relative measurement outputs between neighboring agents rather than the absolute ones of agents themselves. To cooperatively estimate system states with the influence of DoS attacks in a fully distributed scenario, an edge-pinning-based synchronous update strategy for coupling weights is proposed. The purpose of adopting edge pinning method is to adjust partial coupling weights between neighboring agents instead of all ones, which can effectively reduce estimation complexity and save computing resources, especially for large-scale MASs. Further, an edge-pinning-based asynchronous update strategy for coupling weights is designed to meet different actual demands. Then, the security state estimation schemes under synchronous and asynchronous update strategies are proposed without utilizing any global information, such as the Laplacian matrix of communication graph or its smallest nonzero eigenvalue, and so on. Finally, a simulation example is provided to verify the theoretical results.
Note to Practitioners
—This paper addresses the distributed secure state estimation problem for nonlinear MASs, which plays an important role in unmanned aerial vehicle (UAV) formation, intelligent transport systems, smart grids, and so on. The absolute measurement outputs of agents themselves are usually unavailable in practical situations, so the relative measurement outputs between neighboring agents are used to design observers. Moreover, to overcome the difficulty that global information is unavailable in industrial applications with DoS attacks, the edge-pining-based synchronous and asynchronous update strategies for coupling weights are proposed in a fully distributed scenario.