Estimation of correlated and quantized spatial random fields in wireless sensor networks

We develop a novel approach for estimation and inference of spatial field defined over a two-dimensional space under quantized and imperfect channels communication constraints in wireless sensor networks. The objective is to derive an efficient solution for spatial intensity and exceedance estimation. The algorithm developed is based on a series expansion approach in which the first order is the familiar Laplace approximation utilising a Gaussian basis. We illustrate the efficiency of our approach through extensive simulations.