The applications for which the data from Earth observation satellites can be used are dependent on the precision and accuracy of the position of the satellite. Hence, precise orbit determination of satellites is crucial for the success of Earth observation satellites like the Sen
...
The applications for which the data from Earth observation satellites can be used are dependent on the precision and accuracy of the position of the satellite. Hence, precise orbit determination of satellites is crucial for the success of Earth observation satellites like the Sentinel family of satellites. TU Delft is an analysis center of the Copernicus precise orbit determination Quality Working Group and provides orbit solutions of Sentinel satellites for the external validation of the Sentinel orbits computed by the Copernicus Precise Orbit Determination Service. The GIPSYX/RTGx software is used for orbit determination using a reduced dynamic approach.
Existing research has focused on the overall accuracy of the orbit solutions to improve orbit accuracy. However, knowledge of the overall accuracy cannot provide much information about modeling errors. In this research, a data analysis approach was taken to aid the identification of the sources of errors in TU Delft's precise orbit determination strategy of the Sentinel satellites. From a thorough study of the precise orbit determination strategy, a set of parameters were selected for analysis, and a database was created for each satellite. Data analysis methods were customized to create a data analysis framework for the orbit determination data analysis. This framework was applied to the orbit determination data of Sentinel 1A, 1B, 2A, 2B, 3A, and 3B satellites, and the results were interpreted.
The two most common sources of error speculated were errors in the antenna phase center offset or phase center variations, and errors in the location of the center of mass of the satellite. Recommendations were issued regarding ways to confirm the error sources. A significant (> 95 %) correlation was observed between drag coefficient and geomagnetic activity index (Ap index) which shows the improper modeling of geomagnetic activity in the DTM2000 atmospheric density model.