Coarse NLOS detection algorithm based on discrete power levels

Abstract

Indoor localisation has become increasingly important for many services. When the environment is not crowded, i.e. most links between the target and anchors do not suffer from non-line-of sight (NLOS) offsets, the centralised time-difference-of-arrival localisation can efficiently estimate the target position. However, such a method does not work in crowded environments unless a large number of anchors are installed, which is unrealistic for many applications. In this case, the objects that block the LOS links can serve as dynamic anchors (e.g. with round trip time-of-arrival localisation) to improve positioning accuracy. However, such a method tends to include too many unreliable dynamic anchors in estimation, which will significantly increase the computational complexity (e.g. to detect and discard such unreliable anchors in the NLOS mitigation stage). This study proposes a coarse NLOS detection algorithm based on discrete power levels to efficiently achieve the coarse NLOS mitigation, which automatically discards most unreliable dynamic anchors. The simulation results show that the proposed method can achieve very similar estimation performance as the optimal estimation method in the crowded environment, while analysis shows that it can significantly decrease the computational complexity.