A composite stochastic model considering the terrain topography for real-time GNSS monitoring in canyon environments

The site locations of real-time Global navigation satellite system (GNSS) monitoring are usually located in a canyon environment, where the signals are frequently affected by multipath, diffraction, and even non-line-of-sight (NLOS) reception, etc. How to establish an accurate mathematical model is crucial at this time. In this paper, a composite stochastic model based on elevation, azimuth, and carrier-to-noise-power-density ratio (C/N₀) is proposed, which can reflect the terrain topography of the monitoring station. Specifically, according to a mapping function of azimuth, a so-called geographic cut-off elevation is introduced to detect and exclude the NLOS reception and even outlier, then a constrained elevation is obtained. Besides, based on the template functions of C/N₀ and its precision, a procedure is implemented to determine the equivalent elevation, where the contamination of multipath and diffraction are considered properly. To validate the effectiveness of the proposed method, a designed experiment and real deformation monitoring in canyon environments are tested. The results show that the real terrain topography can be reflected to a great extent after using the proposed method. The positioning precision and reliability have been improved, and the performance of ambiguity resolution is also enhanced compared with the other traditional approaches. In real-time kinematic positioning, single-epoch centimeter-level and even millimeter-level accuracies can be obtained under these challenging conditions.