|
Recently, indoor positioning and location-based service have attracted considerable attentions. In harsh indoor environment, non-line-of-sight (NLOS) propagation might cause serious localization bias. Therefore, an efficient method for screening line-of-sight (LOS) measurements is necessary for reliable positioning. In this paper, a novel LOS identification scheme, in which the geometric relationship between position error and angle of arrival (AOA) and the infomation of separated multipath are utilized, is proposed. Specifically, the identification procedure can be divided into three steps. Firstly, range measurements that may contain NLOS bias are used to obtain coarse location. Secondly, we derive a LOS confidence region about the two-dimensional AOA base on the initial coarse position, with the region threshold related to the initial positioning error and estimated AOA deviation. After filtering out the NLOS measurements, an improved two step weighted least squares (TSWLS) position estimator base on time difference of arrival (TDOA) and AOA is developed and used for accurately positioning. Simulation results reveal that the proposed algorithms can yield a high LOS/NLOS identification accuracy and localization precision. |
|
Keywords:indoor positioning, LOS Identification, geometry, separated multipath. |
|