Protocol-Based Distributed Security Fusion Estimation for Time-Varying Uncertain Systems Over Sensor Networks: Tackling DoS Attacks

Abstract

This article studies the distributed fusion estimation (DFE) issue for networked multi-sensor systems (NMSSs) with stochastic uncertainties, bandwidth-constrained network and energy-constrained denial-of-service (DoS) attacks. The stochastic uncertainties reflected in both the state and measurement models are characterized by multiplicative noises. For reducing the communication burden, local estimation signals are subject to dimensionality reduction processing. And the improved Round-Robin (RR) protocol is used on the channels from local estimators to the fusion estimator. To reflect the actual situation, the dimensionality reduction strategy is designed from the defender's point of view in the sense of minimum fusion error covariance (FEC). And the attack strategy is designed from the attacker's point of view in the sense of maximum FEC. Then, based on a compensation model, a recursive distributed Kalman fusion estimation algorithm (DKFEA) is proposed. The stability conditions making the mean square error (MSE) for DFE bounded are derived. In the end, the validity of the presented DKFEA is verified by an illustrative example.