BAYESIAN APPROACH FOR INDOOR WAVE PROPAGATION MODELING
Progress In Electromagnetics Research M
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
... sed using Bayesian information criterion. The proposed indoor propagation model is tested in a two-storey building with access points operating at 2.4 GHz. INTRODUCTION With the expanding of wireless communication technologies in the recent years, it is becoming increasingly difficult to ignore the fact that these technologies have become more complex to design and more expensive to install and maintain. In the new global economy, cost has become a central issue for developing such technologies. A major area of interest within the field of minimizing the cost of wireless communication systems is by reducing the human intervention. A considerable amount of literature has been published on this matter. Some of the studies emphasize the propagation of Radio Frequency (RF) and multipath phenomenon, where RF signals arrive at the receiver from multiple directions due to reflection, diffraction, and scattering. By incorporating the knowledge of RF propagation, researchers were able to reduce the human interference factor by using theoretical propagation models instead of empirical methods. However, a major problem with this kind of propagation models is their accuracy compared to empirical methods. More recently, literature has emerged, which offers extremely meritorious findings about the use of Bayesian models in various applications. In location determination systems, Bayesian models are used to significantly reduce the size of the data set required for model training    , in predicting the popularity of a tweet in social media  , in record linkage and duplication detection  , in the field of blind image deconvolution , and many more. Three essential operations are needed in Bayesian data analysis  . The first step is defining a probability model composed of observed variables and hidden causes, then conditioning on these observed variables in the second step, and lastly evaluating the fit of the model. The objective of this study is to design a Bayesian model that is capable of predicting the signal power level in a complex multi-floors and multi-walls indoor environments. This study aims to contribute to this growing area of research by integrating Bayesian models with the field of signal propagation. This study is divided into four sections. Section 2 provides a literature review on the area of wave propagation modeling. Section 3 introduces the proposed model and describes the methods used to validate the model. Section 4 shows the results and performance analysis of the model, and Section 5 describes summary and conclusion.