An optimum Markov random field-based localization algorithm wireless sensor networks

The received signal strength (RSS) based localization algorithm is proposed in this paper. Here, the RSSs from neighboring sensors are assumed to be statistically dependent. The Markov random field model is employed to explain this dependency. From the model, the optimum sensor locations are obtained from the maximum likelihood estimate. Our experiment has shown that our proposed algorithm can improve the localization accuracy by 9.59% over the traditional localization algorithm without neighboring nodes' information.