Distributed Filtering for Markovian Jump Systems with Packet Loss Compensation

In this article, the distributed-based filtering problem is studied for a type of Markov jump system subject to mode-dependent random packet losses. Packet dropout happens during the communication channel, which can be described via a random variable. In order to describe the nonhomogeneous Markov process, transition probabilities (TPs) on polytopic-structure-based are introduced. In view of packet loss, added by the single exponential smoothing (SES) technique, the packet loss compensation is provided to deal with the prediction value of the missing output. Besides, as for the physical limitations, it is quite difficult to ensure that all mode signals of the plant are available to filter. In the sequel, we establish a partially mode-dependent filter (PMDF). By using the Lyapunov theory, some sufficient conditions are obtained to guarantee the stochastically stable (SS) of the filtering error system (FES), and the filter parameters can be derived. Lastly, the validity of the presented results is tested through a numerical example.

Automated summary

In this article, the distributed-based filtering problem is studied for a type of Markov jump system subject to mode-dependent random packet losses. Transition probabilities on polytopic-structure-based are introduced to describe the nonhomogeneous Markov process. Packet loss compensation is provided using the single exponential smoothing technique, and a partially mode-dependent filter is established to deal with the prediction value of the missing output. By utilizing the Lyapunov theory, sufficient conditions for the stochastic stability of the filtering error system are obtained, and the filter parameters can be derived. The presented results are tested through a numerical example.