Definition and implementation of a new service for precise GNSS positioning

• Autor(a) principal: Oliveira Junior, Paulo Sergio de
• Data de Publicação: 2017
• Tipo de documento: Tese
• Idioma: eng
• Título da fonte: Repositório Institucional da UNESP
• Texto Completo: http://hdl.handle.net/11449/152111
• topic: GNSS; PPP-RTK; ZWD; Troposphere; Ionosphere; Ambiguity resolution; Modeling; Reference network; Troposphère; Ionosphère; Modèles; Réseau de référence; Résolution des ambiguïtés; Troposfera; Ionosfera; Modelos; Rede de referência; Resolução das ambiguidades

PPP (Precise Point Positioning) is a positioning method by GNSS (Global Navigation Satellite Systems), based on SSR (State Space Representation) concept that can provide centimeter accuracy solutions. Real-time PPP (RT-PPP) is possible thanks to the availability of precise products, for orbits and clocks, provided by the International GNSS Service (IGS), as well as by its analysis centers such as CNES (Center National d'Etudes Spatiales). One of the remaining challenges on RT-PPP is the mitigation of atmospheric effects (troposphere and ionosphere) on GNSS signals. Thanks to recent improvements in atmospheric models, RT-PPP can be enhanced, allowing accuracy and centimeter initialization time, comparable to the current NRTK (Network Real-Time Kinematic) method. Such performance depends on topology of permanent stations networks and atmospheric conditions. The main objective of this project is to study the RT-PPP and the optimized infrastructure in terms of costs and benefits to realize the method using atmospheric corrections. Therefore, different configurations of a dense and regular GNSS network existing in France, the Orpheon network, are used. This network has about 160 sites and is owned by Geodata-Diffusion (Hexagon Geosystems). The work was divided into two main stages. Initially, ‘float PPP-RTK’ was evaluated, it corresponds to RT-PPP with improvements resulting from network corrections, although with ambiguities kept float. Further on, network corrections are applied to improve “PPP-RTK” where ambiguities are fixed to their integer values. For the float PPP-RTK, a modified version of the RTKLib 2.4.3 (beta) package is used to take into account for the network corrections. First-order ionospheric effects were eliminated by the iono-free combination and zenith tropospheric delay estimated. The corrections were applied by introducing a priori constrained tropospheric parameters. Periods with different tropospheric conditions were chosen to carry out the study. Adaptive modeling based on OFCs (Optimal Fitting Coefficients) has been developed to describe the behavior of the troposphere, using estimates of tropospheric delays for Orpheon stations. This solution allows one-way communication between the server and the user. The quality of tropospheric corrections is evaluated by comparison to external tropospheric products. The gains achieved in convergence time to 10 centimeters accuracy were statistically quantified. Network topology was assessed by reducing the number of reference stations (up to 75%) using a sparse Orpheon network configuration to perform tropospheric modeling. This did not degrade the tropospheric corrections and similar performances were obtained on the user side. In the second step, PPP-RTK is realized using the PPP-Wizard 1.3 software and CNES real-time products for orbits, clocks and phase biases of satellites. RT-IPPP (Real-Time Integer PPP) is performed with estimation of tropospheric and ionospheric delays. Ionospheric and tropospheric corrections are introduced as a priori parameters constrained to the PPP-RTK of the user. To generate ionospheric corrections, it was implemented a solution aligned with RTCM (Real-Time Maritime Services) conventions, regarding the transmission of ionospheric parameters SSR, which is a standard Inverse Distance Weighting (IDW) algorithm. The choice of the periods for this experiment was made mainly with respect to the ionospheric activity. The comparison of the atmospheric corrections with the external products and the evaluation of different network topologies (dense and sparse) were also carried out in this stage. Statistically, the standard RT-IPPP takes ~ 25 min to achieve a 10 cm horizontal accuracy, which is significantly improved by our method: 46% (convergence in 14 min) with dense network corrections and 24% (convergence in 19 min) with the sparse network. Nevertheless, vertical positioning sees its convergence time slightly increased, especially when corrections are used from a sparse network solution. However, improvements in horizontal positioning due to external SSR corrections from a (dense or sparse) network are promising and may be useful for applications that depend primarily on horizontal positioning.