Skip to Main Content
In this paper, we tackle the problem of spatial interpolation for distributed estimation in Wireless Sensor Networks by using a geostatistical technique called kriging. We present a novel Distributed Iterative Kriging Algorithm (DIKA) which is composed of two main phases. First, the spatial dependence of the field is exploited by calculating semivariograms in an iterative way. Second, the kriging system of equations is solved by an initial set of nodes in a distributed manner, providing some initial interpolation weights to each node. In our algorithm, the estimation accuracy can be improved by iteratively adding new nodes and updating appropriately the weights, which leads to a reduction in the kriging variance. As a consequence, each cluster is constructed adaptively by the set of nodes that achieves the best estimation over the sub-area covered by them. We analyze the most influential parameters to implement this algorithm. Finally, we evaluate the performance of our algorithm and we also analyze its complexity.