NLOS Multipath Detection by Using Machine Learning in Urban Environments

Abstract

In global navigation satellite system (GNSS) positioning, GNSS satellites are often obstructed by buildings, leading to reflected and diffracted signals, which are known as non-line-of-sight (NLOS) signals. These cause major positioning (also known as “NLOS multipath”) errors in GNSS positioning. This paper proposes a novel NLOS multipath detection technique that uses a machine-learning technique to improve positioning accuracy in urban environments. The key idea behind this technique is to construct a classifier that discriminates NLOS multipath signals from the output of the multiple GNSS signal correlators of a software GNSS receiver. In the code-tracking process within GNSS receivers, the code correlation peak is determined and tracked using the outputs of the signal correlators. In the case of an NLOS signal, there are no direct signals; the first reflected signal has low power compared to the direct signal. Hence, the correlation function is expected to be more distorted in the case of NLOS signal correlation. We use this phenomenon to detect NLOS signals. For realizing machine learning, we extract the features of the NLOS signal from the shape of the NLOS correlation function, using an actual dataset, to construct an NLOS classifier. To evaluate the proposed technique, we conduct NLOS classification experiments using signal correlation data acquired at different locations in the Shinjuku area of Japan. We propose to construct an NLOS classifier based on a support vector machine. From the experiments, 87% of the LOS signals and 99% of the NLOS signals are correctly discriminated.