Empirical Analysis of Multipath and Non-Line-of-Sight Errors on Receiver Measurements in an Urban Environment

Signals from global navigation satellite systems (GNSS) are prone to multipath interference and non-line-of-sight (NLOS) reception in urban environments. Buildings surrounding a receiver are liable to block the line-of-sight (LOS) to GNSS satellites and reflect copies of GNSS signals, causing much greater errors in the pseudorange measurements compared to open-sky environments. Such pseudorange errors can lead to navigation solution inaccuracies that can introduce considerable mission risk, particularly if the errors are comparable to or greater than the distance between buildings (canyon width). This study analyzes statistical trends in pseudorange measurements from two commercial GNSS receivers onboard a vehicle traveling through a dense urban canyon. Data collection included multiple laps around the same city blocks to average out the influence of instantaneous satellite geometry. A 3D digital elevation map and ray-casting methods were used to classify whether measurements were LOS or NLOS. To analyze the statistical behavior of pseudorange errors attributed to multipath interference and NLOS reception for each receiver, separate histograms were generated from the errors on the LOS measurements and the errors on the NLOS measurements. The primary findings include a quantitative comparison of the difference in the weight of the tails in the statistical distribution of the errors on the NLOS measurements and the study of one particular case where the statistical distribution of the pseudorange errors from individual satellites have asymmetric tails which appear to mirror each other. This empirical analysis can inform what statistical performance may or may not be consistent across receivers in urban environments.