Distributed Filter Design Over Sensor Networks Under Try-Once-Discard Protocol: Dealing With Sensor-Bias-Corrupted Measurement Censoring

A new distributed filtering problem is fully studied in this article for time-varying systems over sensor networks under measurement errors and measurement censoring, where the measurement error is modeled as stochastic sensor bias driven by a dynamical equation and the measurement censoring is described by the Tobit measurement model. To reduce data congestion and transmission burden, a weighted try-once-discard protocol (WTODP) is applied to transmission channels to efficiently orchestrate data communication. The transmission priority is determined in a dynamical way depending on the importance of missions. The aim of this article is to construct an optimal distributed Tobit Kalman filter (TKF) such that filter parameters are rigorously determined in the minimum mean squared error sense under the consideration of the bias, censoring and WTODP effects. Specifically, sparsity of the network topology is comprehensively considered by using the novel matrix simplification technique. Furthermore, the resultant filtering error is ensured to be exponentially bounded in the mean squared sense. Finally, an illustrative example is used to show the applicability of the proposed filter.