LOS and NLOS identification based on UWB stable distribution

Abstract

Ultra-wideband (UWB) technology for the localization of wireless sensor networks has received considerable attention last few years. This technology is dedicated for indoor localization using a fine delay of resolution and obstacle-penetration capabilities. A lot of challenges remain before implementation of UWB can be deployed on a large scale like non-line-of-sight (NLOS which is especially critical for most location-based applications because the NLOS propagation introduces positive bias in the estimation of distance, which can seriously affect the performance of localization.In this paper, we present a technique for identifying between both line-of-sight(LOS) and non-line-of-sight (NLOS) contexts based on stable distribution parameters using SVM (Support Vector Machine) methods for classification. This characterization was applied to UWB measurements collected from whyless.com project by the IMST group, over bandwidth of 10 GHz.