Bayesian approach for indoor wave propagation modeling

Abstract

This paper presents a parsimonious Bayesian indoor wave propagation model for predicting signal power in multi-wall multi-floor complex indoor environments. The received power is modeled as a Bayesian multiple regression model. The parameters of the model are assessed and validated using a two-tier validation strategy in which Bayes factor and posterior probability are used in the first tier and second tier, respectively. The performance of the two-tier strategy is then assessed using Bayesian information criterion. The proposed indoor propagation model is tested in a two-storey building with access points operating at 2.4GHz.