Quantitative comparison of indoor RFID channel models using bootstrap techniques

Channel characterization is the primary phase of wireless radio frequency identification (RFID) system based tracking processes. Traditional approaches of path-loss modeling of indoor RFID channel lack of the ability to assess the consistency and reliability of the fit. In this paper, we propose two approaches to ensure the reliability of the goodness-of-fit parameters of a fit and quantify its consistency. The first approach uses a standard bootstrap validation method to measure the confidence levels of the goodness-of-fit parameters of the fit. The second approach combines a bootstrap resampling technique with a statistical significance testing method to quantify the consistency of a fit across a number of independent experiments obtained in a particular indoor environment. We present experimental analysis to show the advantages of the proposed approaches over traditional fitting method. We suggest that for proper characterization of indoor channel, it is required to consider the systematic fluctuations along with the path-loss attenuation.