Hybrid Consensus-Based Cubature Kalman Filtering for Distributed State Estimation in Sensor Networks

In this paper, the high-dimensional distributed state estimation problem is investigated for a class of sensor networks within the cubature Kalman filtering (CKF) framework. The network consists of two types of nodes, i.e., communication ones and sensor ones. First, a hybrid consensus-based cubature Kalman filtering (HCCKF) is developed by blending the two existing approaches, namely, consensus on measurements (CM) and consensus on information (CI). As a result, the proposed filtering algorithm has complementary features of CM and CI, which turns out to be a better solution to the distributed state estimation problem. Secondly, estimation errors in HCCKF are proved to be exponentially bounded in mean square. Finally, a target tracking case-study in an example sensor network is given to demonstrate the effectiveness of the proposed HCCKF.