Performance Analysis of Vehicular Localization Accuracy With Non-Line-of-Sight Identification

Recently, vehicular localization has drawn a lot of attention due to increasing location-based applications in vehicular ad-hoc networks (VANETs). However, similar to the conventional localization problem, vehicular localization suffers from the biases caused by non-line-of-sight (NLOS) propagation. To reduce its impact, NLOS identification and mitigation are commonly used, where the accuracy of NLOS identification is crucial. This paper aims to investigate the relationship between localization performance and misidentification of NLOS. The impacts to vehicle localization accuracy of two types of NLOS misidentification, namely false and miss respectively, are analyzed. It is demonstrated with Monte Carlo simulations that those two factors have different impacts on localization accuracy. Furthermore, the dominating factor of localization for different scenarios with two types of localization strategies is concluded, which could be further used to chose the strategy leading to higher localization accuracy in different scenarios.