Clock Modelling in GNSS Processing

Between 2010 and 2013, the Astronomical Institute of the University of Bern (AIUB) took part in the "e;Satellite and station clock modelling for GNSS"e; ESA project as the prime-contractor in a consortium consisting of AIUB, Technical University of Munich (TUM) and the Swiss Federal Institute of Technology of Zurich (ETH). The main objective of the project was to explore and use the stability of the best satellite and ground receiver clocks in Global Navigation Satellite System (GNSS) data processing, which is currently completely ignored in nowadays processing. This is due to the very high requirements put on the stability of the clocks to be useful in that direction: considering the ionosphere-free linear combination has a noise of 3 mm (i.e. 10 ps), the clock has to be stable to roughly 15 ps over the sampling rate of the processed data. This sampling rate can vary from several Hz to 300 s, as typically used for post-processing activities. Such stability is reached by the current best satellite clocks, such as the rubidium clocks onboard the GPS block IIF satellites or the H-Maser clocks of the Galileo satellites.

Clock modelling is a very promising approach in GNSS data processing. It can potentially significantly reduce the number of parameters to be estimated since the epoch-wise satellite clock parameters could be replaced with a very basic model (e.g. 1st order polynomial) over the processed interval. This highly reduced parameterization would help stabilizing the solution in the short-term range (whereas the ambiguity resolution stabilizes the solution primary in the long-term range) if one thinks in terms of the other parameter types such as station height and troposphere which highly correlate with the clock parameters.

On the other hand, independent epoch-wise clock parameters are known for absorbing other model deficiencies, in particular solar radiation pressure modeling defects in orbit models. This leads to another very promising aspect of better clocks in space: by monitoring their performances, one can directly judge about the orbit modeling for example. This was already demonstrated with the H-Maser clocks onboard the Galileo satellites.

Results from this project have been presented for instance at:

Orliac, E., R. Dach, D. Voithenleitner, U. Hugentobler, K. Wang, M. Rothacher, D. Svehla ; 2011: Clock Modeling for GNSS Applications. AGU Fall Meeting 2011, San Francisco, USA, December 5-9, 2011.

Orliac, E., R. Dach, K. Wang, M. Rothacher, D. Voithenleitner, U. Hugentobler, M. Heinze, D. Svehla; 2012: Deterministic and Stochastic Receiver Clock Modeling in Precise Point Positioning. EGU General Assembly 2012, Vienna, Austria, April 22-27, 2012.

Orliac, E., A. Jäggi, R. Dach, U. Weinbach, S. Schön; 2012: Receiver Clock Modelling for GPS-only Gravity Field Recovery from GRACE. EGU General Assembly 2012, Vienna, Austria, April 22-27, 2012.

Orliac, E., R. Dach, K. Wang, M. Rothacher, U. Hugentobler, P. Steigenberger, W. Enderle: Satellite Clock Modelling and Multi-GNSS Solutions. The 2013 Joint UffC, EFTF and PFM Symposium, Prague, Czech Republic, July 21-25, 2013.