Further results on state estimation of multi-agent systems under homologous attacks

This note presents our recent results on state estimation of multi-agent systems under homologous attacks. We first present a sufficient and necessary condition on agents’ dynamics such that agents states can be reconstructed under the considered attack with finite time measurements. Violating the condition prohibits the estimation of agents’ states, regardless of the length of time measurements collected. Then, assuming that agents states are reconstructable, we propose a fully distributed discrete-time estimation algorithm. This algorithm allows each agent to estimate its state and the attack signal through local information sharing. Unlike existing discrete-time estimation methods, agents need not to access global information of networks and agents dynamics by the proposed algorithm.