Distributed probability density function estimation of environmental function from sensor network data

The problem of distributed estimation of the probability density function (PDF) of any environmental function from sensor network measurement is addressed. The proposed algorithm estimate the local spatial parameter of some environmental function as well as the global parameters in distributed manner by fusing the local parameters among the neighboring nodes. The sensor data is modeled using Gaussian mixture PDFs and an algorithm is proposed to estimate the parameters by maximizing the log likelihood function of the sensor data. This algorithm for local and global parameter estimation of the environmental function has been validated using some simulated data. Also real world data of a sensor has been used to estimate the local parameters of an environmental function.