Distributed Clustering Algorithm for Spatial Field Reconstruction in Wireless Sensor Networks

In this paper, we consider the problem of distributed spatial estimation for field reconstruction in wireless sensor networks. In order to estimate the field, a geostatistical technique called kriging is used. Centralized spatial estimation algorithms with a large number of sensors lead to significant computational cost and energy wastage. We present a novel distributed clustering algorithm for estimating spatial interference maps, which are essential for operations and management in future wireless networks. In this algorithm, clusters are adaptively formed with a small subset of sensors by minimizing the kriging variance. The semivariogram computation and kriging prediction are locally performed in each cluster in a distributed fashion. The complexity of the clustering algorithm is analyzed and its performance is evaluated by comparing it with centralized and other distributed approaches.