Lunar satellite orbit determination analysis and quality assessment from Lunar Prospector tracking data and SELENE simulations

Low lunar satellite orbit accuracy is assessed by means of the analysis of tracking data residuals, orbit overlap statistics and covariance propagation. A full-scale determination of a degree and order 75 spherical harmonics lunar gravity field model from 3 months of Lunar Prospector tracking data shows the influence of processing strategies for gravity field modelling. Despite large differences in gravity anomalies over the far side between this model and a comparable existing model, both models show similar results in terms of orbit performance. Analysis of residuals and overlap statistics using the LP150Q gravity field model shows the Doppler data fit to be at the level of a few mm/s. Three-dimensional orbit consistency is found to be at the level of 50–300 m, for polar orbits with altitudes ranging from 30 to 100 km. Covariance analysis shows that the expected radial orbit error as computed from the covariance of an anticipated gravity field model using SELENE data will be less than 1 m. Current gravity field models are shown to be tuned towards the polar orbit, while having expected orbit errors of several orders of magnitude larger for other inclinations, especially in the mid-range, from 40° to 80°.