Indoor NLOS Localization Based on Collaboration of Multiple Base Stations

In indoor environments, Non-Line-of-Sight (NLOS) propagation is universal and causes the deterioration of localization accuracy of the existing solutions, which mainly rely on the information of Line of Sight (LOS) propagation paths. In fact, multipath signals reflect indoor geometry information and target location in NLOS environment. Based on this fact, an indoor NLOS localization method based on collaboration of multiple base stations is proposed, which utilizes the NLOS information to perform localization. Firstly, the LOS propagation regions of multiple base stations and the prior information of the room structure are used to determine the fuzzy regions of scatterer locations. Secondly, the search ranges of scatterer locations are further constrained according to the angle of arrival (AOA). Then a localization model is established based on estimations of time of flight (TOF) differences and locations of scatterers and target. Finally, the target location is acquired with the Leverberg Maquardt (LM) algorithm by employing coarse locations of the target obtained by Genetic Algorithm (GA). Simulation results show that the proposed algorithm is able to use multiple base stations to localize the target with only NLOS paths.