Orbit determination and prediction accuracy analysis for a regional tracking network

China's COMPASS satellite navigation system relies on a regional tracking network to provide navigation services. Limited by its geographic border, the regional network is able to cover only 30% of the medium-earth-orbits (MEO). Accuracy of determined and predicted orbits is not able to satisfy system requirements if the tracking data processing strategy for global tracking network processing is used for the regional network. Two major error sources for orbital prediction are accuracy of initial orbital elements and dynamical modeling. To achieve better prediction accuracy, we propose a two-step orbit determination and prediction strategy. For step 1, only solar radiation pressure (SRP) parameters are estimated along with the orbital elements and other parameters; for step 2, all parameters are estimated but the SRP parameters are tightly constrained to their step 1 estimates. Experimenting with data from a regional GPS network, we conclude for orbital prediction using the proposed two-step strategy, the average user range error (URE) for 24-h prediction arcs is better than 0.6 m.