Multi-GNSS Real-Time Precise Orbit Determination
Uloženo v:
| Vydáno v: | PQDT - Global (2025) |
|---|---|
| Hlavní autor: | |
| Vydáno: |
ProQuest Dissertations & Theses
|
| Témata: | |
| On-line přístup: | Citation/Abstract Full Text - PDF Full text outside of ProQuest |
| Tagy: |
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
|
| Abstrakt: | Providing real-time precise positioning services with Global Navigation Satellite Systems (GNSSs) has a profound impact on various fields such as autonomous driving, natural hazard monitoring and early warning, and surface loading. Stable, reliable, and high-precision real-time satellite orbits and clocks are the prerequisites for real-time Precise Point Positioning (PPP) services. Serving as the essential information for real-time precise positioning, satellite orbits are usually estimated with the latest available observations and predicted for real-time applications. However, the orbit accuracy drops progressively when the orbit update interval becomes longer. This thesis focuses on investigating the provision of high-precision satellite orbits in the real-time, including dynamic orbit modeling, ambiguity fixing and data processing strategy.Solar radiation pressure (SRP) is the most critical non-gravitational force acting on satellite orbits, especially in the eclipsing seasons. Taking GPS orbit as an example, it is demonstrated that the number of unknown parameters in the Empirical CODE Orbit Model (ECOM), which served as a parameterization model, can be reduced if there is a precise a priori box-wing model. For those eclipsing satellites, the shadow factor is recommended to apply in the D (pointing toward to the Sun) direction instead of all three directions. The active parameters in Y and B directions could absorb some unknown forces under eclipse seasons. The superiority of combining the a priori precise box-wing model with five-parameter ECOM (ECOM1) as well as adding shadow factor only in the D direction is proved by orbits and Earth rotation parameters. Compared with the solution with only the ECOM1 model as a parameterization model, the RMS values of orbit day boundary discontinuity (DBD) are improved by 17.8%, 22.7%, and 26.1% for the BLOCK IIR satellites in eclipsing seasons in the along, cross, and radial direction, respectively.Another key role for high-precision satellite orbits is carrier phase integer ambiguity resolution (IAR). Besides double difference (DD) IAR, undifferenced (UD) IAR has also been proven to be achievable in precise orbit determination (POD). The POD solution derived from UD IAR is demonstrated to be superior to that from DD IAR. For example, the orbit accuracy of BDS MEO satellites is improved by 21.7% and 10.4% in the along and cross component, respectively. Similar results can be observed from geodetic parameters including ERPs, station coordinates and geocenter coordinates. The orbits and geodetic parameters demonstrate that the differences between DD IAR and UD IAR solutions stem from the absence of independent DD ambiguities and incorrectly fixed DD ambiguity, in which the former takes the primary leading to orbit differences. Solving the above two problems is challenging, particularly when dealing with a massive network, so UD IAR is highly recommended for daily GNSS data processing.In the last part, a novel data processing strategy that parallels the epoch processing and significantly enhances the computation efficiency is proposed. In our proposed strategy, a 24- IV Abstract hour processing job is split into several sub-sessions that are processed in parallel, and then stacked to solve and recover parameters. Together with paralleling other procedures such as orbit integration and using open multi-processing (openMP), the multi-GNSS POD of 120 satellites using 90 stations can be fulfilled within 30 min. With historical information, including fixed UD ambiguities and cleaned observations, the network solution with 100 stations and 120 satellites can be finished in 10 min, in which one iteration of parameter estimation only costs 3 min. The predicted orbits derived from epoch-parallel-based solution equal to the legacy sequential batch solution. Compared with IGS products, the average 1D RMS values of user-available predicted orbits updated per 10 min are 3.1, 5.8, 3.4, 138.8, 18.1, and 6.8 cm for GPS, GLONASS, Galileo, BDS GEO, BDS IGSO, and BDS MEO satellites, respectively. The DBDs of corresponding orbits are around 1.0 cm for MEO satellites.This thesis demonstrates the feasibility of the proposed strategies in providing high-precision near real-time satellite orbits. The refined satellite force modeling and ambiguity fixing strategy improve the accuracy, while the proposed epoch-parallel strategy can shorten orbit update time effectively with the aid of multi-nodes and historical information. The stable and high-precision satellite orbits are beneficial for real-time clock estimation, precise point positioning, and atmospheric sounding. |
|---|---|
| ISBN: | 9798290615660 |
| Zdroj: | ProQuest Dissertations & Theses Global |