Non-Line-of-Sight Identification for Indoor Positioning Using Ultra-WideBand Radio Signals

Abstract

In this paper, Non-Line-of-Sight (NLoS) identification/mitigation is proposed, based on systematic statistical hypothesis testing. In the proposed method, the combined statistics of timing-based measurements and Received-Signal-Strength (RSS) are used, and the fact that all collected measurements are related to the same unknown position is exploited. Four implementations are proposed, ranging from the rigorous optimal approach to a much simplified and approximate approach, offering a trade-off between performance and complexity. Validation results, using actual Ultra-Wideband radio signals, show that the correct detection rate can go up to 99% and the corresponding positioning accuracy with NLoS mitigation is substantially improved: the Root Mean Squared Error (RMSE) of the estimated position is reduced by up to a factor of seven. The use of the simplest approach comes with only a minor decrease in performance, compared to the full optimal implementation. Copyright © 2013 Institute of Navigation