Refinement of Reduced-Dynamic Orbit Determination for Low Earth Satellites

Abstract

Precise orbit information for Earth observation satellites gains ever more importance as the request for high quality remote sensing products increases. Global PositioningSystem (GPS)-based reduced-dynamic orbit determination has evolved as the state-of-the-art for high-precision orbit estimation of low altitude spacecraft. It combines a priori models of the spacecraft dynamics with varying levels of empirical parameters to best exploit the high precision of the available GPS observations. The optimum trade-off between the quality of dynamical models, the required level of stochastic parameters, and the geometric strength of the GPS observations is a matter of ongoing research and focus of this work. Given the high quality of orbit determination solutions that has been demonstrated even without any non-gravitational force models in missions such as CHAllenging Minisatellite Payload (CHAMP), Gravity Recovery And Climate Experiment (GRACE), Gravity field and steady-state Ocean Circulation Explorer (GOCE), and Swarm, the present thesis aims to answer the question whether and to what extent refined dynamical models and refined GPS processing techniques can contribute to further improve the achievable orbit determination accuracy. The study is based on a comprehensive set of current Earth observation missions in low Earth orbit that are equipped with geodetic-grade GPS receivers. This comprises the Sentinel-1A, Swarm-C, and TerraSAR-X/TanDEM-X missions with altitude of 450 km to 693 km, where notable perturbations due to atmospheric forces affect the satellite motion. For each of these missions, dedicated macro models have been established and used for the description of atmospheric drag and lift forces, solar radiation pressure, and Earth radiation pressure with targeted modeling accuracies at the one nm/s2 level. The benefit of using such models is assessed through different performance metrics including self-consistency tests, satellite laser ranging (SLR), and radar ranging as an external validation technique. Use of advanced atmospheric density models and spacecraft macro models for atmospheric forces is found to slightly reduce the associated empirical accelerations but does not allow to entirely waive the estimation of such parameters. With respect to radiation pressure, the macro model appear to be essential for a realistic description of Earth radiation pressure (ERP). In particular, it benefits the modeling of the radial ERP acceleration, which directly impacts the height leveling of the resulting orbit, and is such of key relevance for altimetry missions. With respect to GPS observation modeling, the use of ambiguity fixing is shown as a key technique to achieve improved orbit determination accuracy for all orbit geometries. Its benefit largely outweigh that of non-gravitational force models and contribute to major improvements in the horizontal (along-track/cross-track) position knowledge. On the other hand, the reduced vertical dilution of precision still makes the resulting orbit solutions sensitive to geometric orbit modeling errors in radial direction and justifies the use of refined radial acceleration models. Overall, ambiguity fixing can offer 33 % improvement in orbit determination accuracy and allows to reach a one-cm level (1-D) performance, as evidenced by the analysis of SLR residuals for the aforementioned missions. Radar-ranging is shown to enable independent validation of precise orbit determination solutions in Synthetic Aperture Radar (SAR) missions. However, it is not yet fully competitive with SLR in terms of both accuracy and coverage. In particular, it is confined to high resolution SAR imaging, and a global network of corner cube reflectors would be required for a wider use.