Angular Information-Based NLOS/LOS Identification for Vehicle to Vehicle MIMO System

Distinguishing between line-of-sight (LOS) and non-LOS (NLOS) based on measured channel impulse responses is a fundamental problem for localization systems, as well as channel modeling. In this work we show that the use of angular information can considerably enhance the accuracy of such a scenario discrimination, and we propose the use of support vector machines (SVM) for performing this task. We demonstrate, using real-world channel measurements, that the new method provides significantly improved detection, reducing in our measured channels the mis-identification rate from 10% to less than 4%.