Improved prediction of GPS satellite clock sub-daily variations based on daily repeat

Abstract

High-precision estimates of GPS satellite clock errors reveal systematic sub-daily clock bias variations on the order of 1 ns. The low noise levels in the Rubidium Atomic Frequency Standards onboard GPS IIR, IIR-M, and IIF satellites provide visibility of these small, but systematic behaviors. Prior studies have reported on this phenomenon and sought to characterize the specific frequency components present and to identify potential causes of the observed periodic variations. Our research focuses on the repeatability of the clock variations and the potential for using observed variations directly to predict future clock behavior. Results are presented and compared to IGS Ultra-rapid and broadcast message predictions for all operating GPS satellite clocks for a 1-month period in July 2017. During this time, the accuracy of the proposed sub-daily variation prediction is better than 0.15 ns (RMS) for 8 out of 9 GPS Block IIF Rb clocks and under 0.3 ns (RMS) for most GPS IIR and IIR-M Rb clocks. This approach is complementary to existing techniques for estimating longer-term clock rates and drift and can be combined with them to improve the fidelity of predictive satellite clock models for real-time GPS position, navigation, and timing applications.