A Novel NLOS Identification Algorithm for Indoor Localization Based on Ranging Residuals

Abstract

Wireless signals’ Non-line-of-sight (NLOS) propagation will seriously affect the localization accuracy of indoor localization systems, so mitigating NLOS effects is essential for improving indoor localization accuracy. Many algorithms improve localization accuracy by identifying the distance with NLOS noise and discard it. Most of existing NLOS identification algorithms requires a lot of priori information about the NLOS signal. However, this information is difficult to obtain in practical engineering problems. In this paper, the main goal is to identify NLOS noise without any priori information. To obtain this goal, we proposed a novel NLOS identification algorithm for indoor localization based on ranging residuals. The algorithm adds two auxiliary anchor nodes for the tag to locate anchor nodes in reverse. Since the anchor node location is known, we use the anchor node error to identify NLOS noise. The algorithm only identifies NLOS noise based on the measurement distance and anchor nodes’ location. Then we choose the anchor nodes without NLOS noise to locate the tags. The results of simulation demonstrate that the identification success rate of the NLOS identification algorithm under different NLOS distributions can reach 95.85%, and the localization accuracy is less than 1m.