GNSS CLOCK PERFORMANCE ANALYSIS USING ONE-WAY CARRIER PHASE AND NETWORK METHODS

摘要

Analysis of GNSS clocks performance is traditionally carried out by network analysis techniques where orbits and all participant clocks are computed in a batch or Kalman Filter least-squares adjustment to a reference time scale and reference frame. Several companies, organizations, and IGS analysis centers produce postprocessing orbits and clock estimations mainly at 15 or 5 minutes respectively, from which it is possible to extract the onboard clock performance on the medium or long term. Final IGS solutions are used as reference for validation of any new algorithm.Within the associated experimentation to the Galileo In Orbit Validation Element-A (Giove-A), GPS, and Giove-A satellite orbits and clocks are estimated by the Orbit Determination and Time Synchronization (ODTS) software from observations of a dual worldwide GPS/Galileo Network, being GIOVE-A results kindly augmented by Laser Ranging observations. Following the Galileo risk mitigation, the Giove-A satellite was launched on December 2005 carrying two RAFS clocks manufactured by Spectratime (formerly Temex Time).While the medium- and long-term performance is of the highest importance for navigation and integrity at the corresponding user level, since the prediction errors are directly mapped into the User Equivalent Range Error (UERE), the short-term performance of the timing signal has to be low enough to allow the user receiver to integrate the received signals over long intervals.The present paper analyzes ODTS estimations, IGS new 30-second clock solutions, and the new one-way carrier phase technique for the short term, in order to analyze the full clock performance from 1 second. Additionally, ground results for Giove-A rubidium clocks allow assessment of the difference between the physical and the signal (or "apparent") clocks. Detailed results will be presented for each GPS and Giove-A FM4 and FM5 clocks. In particular, special emphasis will be given to the clock estimations of the stations located at time laboratories, the analysis of the noise of the different methods, and limitations when characterizing the onboard clock performance.