GNSS Fault Detection and Exclusion (FDE) under sidewalk constraints for pedestrian localization in urban canyons

Abstract

The Global Navigation Satellite System (GNSS) has gained widespread use in smartphones, providing support for various pedestrian applications. In urban areas, multipath effects introduce large errors in different measurements, severely degrading the GNSS accuracy. To mitigate multipath effects, several Fault Detection and Exclusion (FDE) methods have been developed. However, in urban canyons, their effectiveness is significantly degraded due to the lack of fault-free measurements in the cross-street direction. In urban canyons, the pedestrian network provides an opportunity to improve the urban GNSS accuracy for pedestrians. The purpose of this study is to improve the GNSS FDE performance through the sidewalk constraints. A new scheme has been proposed to distinguish the correct side of the street effectively. The Hough Transform estimator was used to find the most consistent GNSS measurements under sidewalk constraints. To assess the proposed algorithm’s performance, extensive tests were conducted in urban canyons. The analysis of Carrier-to-Noise density ratio (C/N0) shows that 92% of sidewalks can be distinguished from the opposite sidewalk along the same street. The static test shows that the positioning accuracy can be improved from 22 m to 4.9 m, a 77% improvement over the residual based FDE. The dynamic test showed that the proposed method can achieve the sidewalk positioning, which is essential for many pedestrian applications such as last-mile delivery, emergency caller positioning and jaywalking monitoring.