A Two-Stage Fingerprint Filtering Approach for Wi-Fi RSS-Based Location Matching

Abstract

This paper presents a two-stage fingerprint filtering (FF) approach for the received signal strength (RSS) based location matching in Wi-Fi area . W ith the popularity of ubiquitous computing and location - based services in a recent decade , the Wi-Fi RSS -based location matching has been one of the most widely recognized method to locate users’ positions due to its low cost of deployment and maintenance. However, the indoor or urban wireless channel is characterized by time varying, non-line-of-sight, and multi-path interference. The efficiency of fingerprints is significantly deteriorated. Therefore, construct ing a reliable radio map for the location matching becomes a key and unavoidable challenge. T he experiments conducted in the real Wi-Fi environment indicate that the FF approach can yield better location precision (or the probability of locating the test position at matching position) and accuracy (or the cumulative distribution functions of errors) compared with the conventional neighbor matching .