A Reliable Propagation Channel Model for a Better RFID-Based Indoor Localization System

Abstract

Fingerprinting has been extensively researched for localization solutions in complex environments. The fingerprint is a dataset of the Received Signal Strength Indicators (RSSIs) coming from the emitting antennas associated with different wireless technologies. Indoor, RSSIs are affected by multipath. To overcome these variations impact, we propose an advanced calibration approach applied on a set of RSSs collected at several positions distributed in a classroom. The first step of this method consists in determining the propagation channel attenuation coefficients. Then, the accuracy of the environment modelling is enhanced by the Weighted Average Attenuation Factors (WAAF) procedure. It is validated within the offline stage of an RFID-based indoor localization system. In fact, it improves both the reliability of the built radio map by covering the whole environment, as well as the location accuracy obtained by the positioning process. Experimental results confirm that the proposed system based on the WAAF channel model features better positioning errors compared to the conventional systems; they are reduced by 35 percent, while deploying 0.062 RFID tag per square meter only.