Automated GPS-based operational orbit determination

Satellite operations depend on being able to generate accurate predictions of a spacecraft's orbit in a very short period of time, typically a few hours, after observations are made. The satellite ephemeris generated in this process is used by mission controllers for planning operations such as vehicle pointing and orbit adjust generation. The research described in this dissertation, investigates the methods and parameterizations necessary to achieve a fast and accurate ephemeris. To accomplish these investigations, an automated system is used. Two distinct spacecraft missions are discussed. They each have specific goals that must be met by their operational orbit determination systems. The first is ICESat, a scientific satellite that is part of NASA's Earth Observation System (EOS), and is operated by the Laboratory for Atmospheric and Space Physics (LASP). The primary OD requirement for ICESat is to provide predictions accurate to 10 meters cross-track for 48 hours to accomplish instrument pointing planning. The second mission is Quick-Bird, a commercial imaging satellite that is owned and operated by Digital Globe, Inc. QuickBird requires post-processed orbits with 3 meters (1sigma) accuracy in total position and 30 day orbit predictions to accomplish imagery planning. A variety of measurement processing schemes and error corrections are explored for each of these spacecraft. It is shown that it is possible to achieve approximately one meter (1sigma) orbits for both spacecraft in a orbit determination system that is designed for use in spacecraft operations. In the ICESat case, it was found that using single-differenced measurements met the requirements while reducing both the processing time and the logistical load for importing external data. QuickBird benefitted from the addition of the DRVID method of ionospheric removal and from using double-differenced measurements.