UWB Geo-Regioning Using Multivariate Channel Statistics

Ultra-wideband (UWB) geo-regioning is an approach to localize UWB transmitters by means of their channel impulse responses. A received channel impulse response is assigned to a geographic region based on a-priori known characteristics of this region. The first approach relying on knowledge of the average power delay profile of the channel was presented and analyzed in (F. Athaus et al., 2005), (C. Steiner et al., 2006). In this paper we extend this method by including the knowledge of the correlation between channel taps. We discuss this new modeling assumption, derive the respective multivariate maximum likelihood decision algorithm, provide an analytic expression for the probabilities of misclassification, and present simulation results based on measured channel impulse responses. A significant performance gain is visible compared to the previous average power delay profile approach, where the taps are assumed to be independent.