Modeling tropospheric wet delays with dense and sparse network configurations for PPP-RTK

Abstract

Precise Point Positioning (PPP) is a well-known technique of positioning by Global Navigation Satellite Systems (GNSS) that provides accurate solutions. With the availability of real-time precise orbit and clock products provided by the International GNSS Service (IGS) and by individual analysis centers such as Centre National d'Etudes Spatiales through the IGS Real-Time Project, PPP in real time is achievable. With such orbit and clock products and using dual-frequency receivers, first-order ionospheric effects can be eliminated by the ionospheric-free combination. Concerning the tropospheric delays, the Zenith Hydrostatic Delays can be quite well modeled, although the Zenith Wet Delays (ZWDs) have to be estimated because they cannot be mitigated by, for instance, observable combinations. However, adding ZWD estimates in PPP processing increases the time to achieve accurate positions. In order to reduce this convergence time, we (1) model the behavior of troposphere over France using ZWD estimates at Orpheon GNSS reference network stations and (2) send the modeling parameters to the GNSS users to be introduced as a priori ZWDs, with an appropriate uncertainty. At the user level, float PPP-RTK is achieved; that is, GNSS data are performed in kinematic mode and ambiguities are kept float. The quality of the modeling is assessed by comparison with tropospheric products published by Institut National de l'Information Geographique et Forestiere. Finally, the improvements in terms of required time to achieve 10-cm accuracy for the rover position (simulated float PPP-RTK) are quantified and discussed. Results for 68 % quantiles of absolute errors convergence show that gains for GPS-only positioning with ZWDs derived from the assessed tropospheric modeling are about: 1 % (East), 20 % (North), and 5 % (Up). Since ZWD estimation is correlated with satellite geometry, we also investigated the positioning when processing GPS + GLONASS data, which increases significantly the number of available satellites. The improvements achieved by adding tropospheric corrections in this case are about: 2 % (East), 5 % (North), and 13 % (Up). Finally, a reduction in the number of reference stations by using a sparser network configuration to perform the tropospheric modeling does not degrade the generated tropospheric corrections, and similar performances are achieved.