Event-Based Distributed Estimation of Stochastic Systems over Sensor Networks with Deception Attacks

Abstract

This paper addresses H _∞ state estimation problem for a class of discrete-time stochastic delayed systems subject to deception attacks over sensor networks. We consider the scenario where the sensor output and estimator output are transmitted to its corresponding estimator and other estimators, respectively, over two different digital channels. To cope with the problem of constrained bandwidth, the event-triggering ideas are implemented on both sensor-to-estimator and estimator-to-estimator channels. Meanwhile, we consider there is an attacker that possibly corrupts the measurement output. Under these circumstances, we aim to design the state estimator such that the exponential mean-square stability with H _∞ performance is satisfied. By virtue of Lyapunov-Krasovskii functional, we can derive the explicit estimator parameters by solving a set of linear matrix inequalities (LMIs). Finally, a numerical example is established to verify our theoretical results.