An efficient localization method using RFID tag floor localization and dead reckoning

Abstract

A method for using radio-frequency identification tag floor based localization (RFTL) for indoor mobile robot localization is proposed. This method does not suffer from illumination or line of sight impairment issue, and it is less affected by the environment than is conventional localization. However, it faces an inherent limitation, in terms of its low accuracy. To overcome this problem, an efficient localization method using RFTL and dead reckoning is presented. To utilize this method, it is important to find a model of map for the tag recognition probabilities. With this in mind, this paper proposes two model structures: a deterministic ellipse model, and a two variable Gaussian random variable (RV) model. The results of experiments to verify and compare the performances of these two model structure methods are presented.