Distributed State Estimation Under Multistep Random Transmission Delay and Packet Loss

This article investigates the problem of underwater distributed state estimation with multistep random communication delay and packet loss. Since underwater communication adopts acoustic communication, the information transmission speed is relatively slow, and multistep delay and packet loss situations are frequently encountered. Based on the hybrid consensus filtering algorithm and a proper buffer setting, a distributed Kalman filtering algorithm is proposed, which can deal with possible multistep random time delay and packet loss. A sufficient condition for the filtering stability is rigorously established. In addition, the state estimation problem for an uncertain target model in underwater environments is addressed by combining the interacting multiple model method. Finally, numerical simulations validate the effectiveness of the proposed algorithm and the accuracy of the conclusions, while also demonstrating that this method achieves a 13.9% reduction in the cross-scenario composite root-mean-square error compared to similar buffer-based approaches, further emphasizing its superiority.