Real Time, Autonomous, Precise Orbit Determination Using the Global Positioning System

摘要

In an effort to estimate precise satellite ephemeris in real-time on board a satellite, the Goddard Space Flight Center (GSFC) created the GPS enhanced Orbit Determination Experiment (GEODE) flight navigation software. This dissertation offers alternative methods and improvements to GEODE to increase on board autonomy and real-time total position accuracy and precision without increasing computational burden. First, GEODE is modified to include a Gravity Acceleration Approximation Function (GAAF) to replace the traditional spherical harmonic representation of the gravity field. Next, an ionospheric correction method called Differenced Range Versus Integrated Doppler (DRVID) is applied to correct for ionospheric errors in the GPS measurements used in GEODE. Then, Dynamic Model Compensation (DMC) is added to estimate unmodeled and/or mismodeled forces in the dynamic model and to provide an alternative process noise variance-covariance formulation. Finally, a Genetic Algorithm (GA) is implemented in the form of Genetic Model Compensation (GMC) to optimize DMC forcing noise parameters. Application of GAAF, DRVID an i DMC improved GEODE's position estimates by 28.3% when applied to GPS/MET data collected in the presence of Selective Availability (SA), 17.5% when SA is removed from the GPS/ MET data and 10.8% on SA free TOPEX data. Position estimates with RSS errors below 1 meter are now achieved using SA free TOPEX data. DRVID causes an increase in computational burden while GAAF and DMC reduce computational burden. The net effect of applying GAAF, DRVID and DMC is an improvement in GEODE's accuracy/precision without an increase in computational burden.