A Consensus Nonlinear Filter With Measurement Uncertainty in Distributed Sensor Networks
Kai Shen, Zhongliang Jing, Peng Dong · IEEE Signal Processing Letters · 2017
This letter addresses the consensus-based nonlinear state estimation in distributed sensor networks with unknown measurement noise statistics. The existence of naive nodes and the communication constraint in sensor networks requires a hybrid consensus filtering method. In the frame of consensus filtering, a novel consensus nonlinear filtering approach named variational Bayesian consensus cubature Kalman filter (VB-CCKF) is proposed, in which the CKF is employed to handle the nonlinear state estimation and the VB approximation is adopted to iteratively estimate the sufficient statistics of the measurement noise covariance on each step. Simulations are performed in order to demonstrate the effectiveness of the proposed approach.