Cluster Based Routing Scheme for Distributed Regression in Wireless Sensor Networks: Gaussian Eliminations

This paper presents a cluster based routing scheme to perform the distributed Gaussian elimination algorithm for solving the regression algebraic equation towards to data regression. Under this routing framework, the junction tree structure is incorporated to coordinate the clusters so as to decouple the intersections among the neighboring clusters via a new message passing protocol. Simulation-based experimental results are presented. It is shown that the use of synchronous protocol of message passing scalably implemented among clusters is more efficient than the purely asynchronous protocol over entire sensor networks.