A Realistic Stochastic Model Considering the Multipath Effects Based on C/N0 for GNSS Positioning in Harsh Environments

The presence of unmodeled errors such as multipath and non-line-of-sight (NLOS) significantly degrades the positioning performance of global navigation satellite systems (GNSS), especially in harsh environments. We propose a novel stochastic model that considers multipath effects based on carrier-to-noise-density ratio (C/N0). This model utilizes C/N0 observations and their nominal function to construct a five-segment variance shrinking factor to enable a more accurate characterization of observation precision and yields reliable posterior residuals. Subsequently, the Institute of Geodesy and Geophysics 3 (IGG3) scheme based on posterior residuals is applied for quality control to enhance the performance of the results. Three sets of real-time kinematic (RTK) positioning experiments were conducted in harsh environments such as urban canyons and shaded areas to validate the effectiveness of the proposed method. For single-epoch static RTK positioning, the proposed method improves the horizontal accuracy of the Trimble Alloy from decimeter to centimeter level and the u-blox F9P from approximately 1.5 m to 1.0 m, with improvements of 2-13% in the ambiguity success rate compared to traditional equal-weight, C/N0, and elevation models. In kinematic positioning, the proposed method significantly improves positioning accuracy by more than 20% and increases the ambiguity success rate. The Trimble Alloy achieves decimeter-level horizontal positioning accuracy in deep urban canyons, while the u-blox F9P achieves an accuracy of around 1 m.