Distributed Fusion Estimation for Nonlinear Systems under Denial-of-Service Attacks using the Try-Once-Discard Protocol

This paper tackles the challenge of distributed state fusion estimation in nonlinear, time-varying systems subjected to external denial-of-service (DoS) attacks. Given the susceptibility of network channels to such attacks, a transmission strategy employing the try-once-discard (TOD) protocol was introduced to reduce their impact. In this approach, each sensor in the network is linked to local estimators through multiple channels, with the TOD protocol determining the transmission priority. Data transmitted via the highest priority channel is retained, while data from other channels are discarded. A local estimator was constructed using the cubature Kalman filter (CKF), and a consensus-based approach was utilized to merge the local estimates, forming a distributed fusion estimation algorithm. The algorithm’s convergence was analyzed, and sufficient conditions for maintaining bounded filter error covariance were established, accounting for network disruptions and random nonlinearity. The proposed algorithm’s performance was demonstrated through numerical simulations.