Reputation-Aware Distributed Filtering for Nonlinear Bias-Corrupted Systems Over Sensor Networks Under Event-Triggered Mechanism

Abstract

This article is concerned with the reputation-aware Kalman-type distributed filtering (RAKTDF) problem for a class of nonlinear bias-corrupted systems (NBCSs) with a dynamical event-triggered mechanism (DETM). First, a representative unknown input, namely the dynamical bias, is introduced, which is evolved by a dynamical equation with Gaussian white noise. A DETM is employed to regulate the frequency of data transmission so that data conflict and network congestion are avoided. In order to identify and eliminate abnormal data from neighbors, a trust-based scoring strategy, referred to as the reputation-aware mechanism, is modeled and utilized to improve the accuracy of the filtering algorithm. The RAKTDF algorithm is recursively developed such that the covariance upper bound of the filtering error dynamic (CUBFED) is derived, after which the filter gain is determined by minimizing the trace of the CUBFED. Furthermore, a sufficient condition is established to guarantee the boundedness of the filtering error dynamics. Finally, an illustrative example is provided to verify the effectiveness of the proposed RAKTDF algorithm.