A Novel GNSS Quality Control Method With Satellite Geometric Network for Dynamic Urban Positioning

Quality control of raw measurements in global navigation satellite systems is important for precise positioning, navigation, and timing (PNT), particularly in mission-critical applications like urban air mobility (UAM) and autonomous driving. Positioning accuracy as a key performance of satellite navigation systems is affected by both the precision of raw measurements and the geometric configuration of all tracking satellites. Significant research has been investigated to detect satellites with non-line-of-sight (NLOS) or multipath (MP) errors and subsequently exclude or de-weight those satellites. However, excluding all detected satellites in urban areas can profoundly deteriorate the geometric configuration of the tracking satellites, and de-weighting all detected satellites suffers from inaccurate weighting models leading to poor transferability. To address this issue, this article presents a novel quality control algorithm with satellite geometric networks. The proposed method employs two designed detectors and a nested sliding window scheme to detect satellites with NLOS or MP error. Subsequently, complex network statistics derived from the satellite geometric network are used to assess the contribution of each satellite to the geometric configuration. Based on the determined contribution of each satellite, an exclusion or de-weighting strategy is applied to the corresponding satellites. The positioning results from two vehicular tests in urban areas show that 3-D root mean square errors of 5.92 and 8.39 m can be achieved in moderate and deep urban areas, respectively. These results exhibit superiority in positioning accuracy over the state-of-the-art of over 50% in moderate urban areas and over 80% in deep urban areas, respectively.