Hierarchical Fusion Estimation for WSNs with Link Failures Based on Kalman-Consensus Filtering and Covariance Intersection

This paper studies hierarchical fusion estimation for wireless sensor networks (WSNs) with link failures based on Kalman-consensus filtering and covariance intersection (CI). The spatial distribution of multiple sensors are divided into several clusters, and each cluster has a certain number of sensors. First of all, the data of sensor nodes are processed by Kalman-consensus filtering algorithm. The sensors nodes can communicate with each other to maintain the consensus of local estimation. Sensors nodes in the same cluster are connected to a cluster head node as local estimators. Considering link failures, these failures are modeled as a set of independent Bernoulli processes. Secondly, CI fusion algorithm is used to generate fusion estimates from the collected local estimates. In addition, the proposed hierarchical fusion estimates are suitable for real applications. The simulation results verify the effectiveness of the proposed method.