Relevant CIR parameters selection for fingerprinting based location algorithm

Abstract

This paper investigates the relevance of Channel Impulse Response (CIR) parameters for fingerprint based localization algorithm. We consider in this study a WLAN environment where a Multi-Layer Neural Network (MLNN) learns off-line the location from the location-dependent parameters of the CIR. Then, it calculates on-line the accurate position of a mobile station from given fingerprints. More importantly, an advanced study is conducted to identify the relevant parameters of CIR to be used as fingerprints which offer lower complexity and high location accuracy.