RSS-Based Localization Using Bayesian Hierarchical Model with Spatially Correlated Shadow Fading

A new localization technique using received signal strength (RSS) is proposed in this letter, which considers the case of spatially correlated shadow fading. Particularly, a three-stage Bayesian hierarchical model (BHM) is introduced to simulate the signal propagation process of spatially correlated shadow fading. Using the data generated by the wireless sensor networks (WSN) to estimate the parameter of BHM by applying Markov chain Monte Carlo (MCMC) algorithm. When the transmission power of the source is unknown, the parameters' distributions are updated according to the collected RSS data. Finally, the location of the signal source could be estimated from the statistical characteristics of the posterior predictive distribution. Performance results are presented, which demonstrate that the proposed MCMC-based method provides a significant improvement over the spatially correlated shadow fading on RSS measurement and improves positioning accuracy.