Abstract
Troposphere slant wet delays (SWD) are usually assumed to be of isotropic nature, which allows simple modelling in precise point positioning (PPP) applications. However, this assumption can lead to considerable errors, up to ten centimeters under extreme weather conditions. The traditional approach to model an inhomogeneous troposphere is based on a two-axis gradient model, of which the accuracy and flexibility are both limited. In order to better consider isotropy in the troposphere, we propose a new model based on B-splines. This B-spline model considers the variation of SWD with azimuth, but tries to incorporate troposphere information at different elevation angles. In this study, we evaluate traditional models, e.g., a two-axis gradient model and a 2nd-order harmonic gradient model, and compare them to our B-spline model. We compare the ability of these approaches to model non-isotropic atmospheric conditions using data from a commercial GNSS simulator. We show that the two-axis gradient model has non-negligible residual errors when modelling the inhomogeneous troposphere, while the second-order gradient model and the B-spline model can better handle complex atmospheric conditions. We also carry out tests based on real data from IGS geodetic stations in Europe. We compute multi-GNSS undifferenced-uncombined (UDUC) PPP solutions, which are based on float integer ambiguity values. The results show that the B-spline model works well within PPP analysis and delivers meaningful troposphere estimates that better reflect complex atmospheric scenarios.
Users
Please
log in to take part in the discussion (add own reviews or comments).
