A Novel Location Algorithm Based on Dynamic Compensation Using Linear Location Prediction in NLOS Situations

Due to the time-varying and multi-path environment in mobile communications, Non-Line-Of-Sight (NLOS) has become the most challenging problem when reaching higher accuracy in wireless location estimation. In this paper, we propose a novel location algorithm named Linear Location Prediction (LLP) to gain favorable TOA location accuracy without bringing much complexity. By fully exploiting the common characteristics of moving objects, we provide linear location prediction estimation to MS in advance as a reference value to the newly estimated TOA measurement at the next sampling moment in NLOS situations. Therefore, the effect of NLOS errors is greatly reduced as the NLOS-contaminated estimated value can be compensated dynamically using the predicted one, yielding higher location accuracy. The simulation shows that the location error of our algorithm is restricted to an encouraging level of only 0.074 kilometers, which is one third of that of the best algorithm, BFA [2]. In addition to that, the slowly rising curve of location error versus number of NLOS-contaminated BSs proves its greater robustness to NLOS contamination as well.