GPS-Based Real-Time LEO Satellite Navigation Using Kinematic and Reduced Dynamic Filtering Approaches

Precise onboard real-time satellite navigation based on Global Navigation Satellite Systems (GNSS) is becoming increasingly important for Low Earth Orbit (LEO) missions. Besides obtaining a point solution, statistical filtering techniques allow not only an increase in precision, but also to improve the robustness in the absence of measurements. The system model on which these techniques are based can vary from a kinematic approach, which use only the instantaneous state of the satellite, to a dynamic approach, where the orbital dynamics are exhaustively modeled. A reduced dynamic approach, where a force model is based on dynamic parameters included in the estimation, presents a balance between both. This work presents the implementation and testing of two orbital navigation systems for LEO satellites. Each system is based on a kinematic and reduced dynamic approach respectively, for the statistical filtering of dual-frequency GPS pseudorange and delta-range measurements. The system model used in each case is described, as well as the details of the implementation in a in-house designed and manufactured GPS receiver. As an evaluation, the results obtained through testing with a GNSS signal simulator in an orbital scenario are presented, comparing the navigation solutions to the simulated orbit. The results show correct performance of both systems, obtaining a potential position and velocity precision of 1 m and 10 mm/s for the reduced dynamic approach. The computational load of each approach is also analyzed, showing that although the reduced dynamic approach has a higher computational cost, it performs well in the hardware employed and provides greater precision and robustness.