Advanced GNSS Algorithms and Services Based on Highly-stable On-board Clocks

The on-board GNSS clock technology has evolved greatly in the last years, moving from the initial Caesium atomic clocks, to the latest Rubidium (Rb) and Passive Hydrogen-Maser (PHM) clocks. Both Rb and PHM technologies have proven to be highly predictable clocks, which opens the door to a series of improvements or modifications with respect to the “classical” Orbit Determination and Clock Synchronisation (ODTS) algorithms that would not only lead to the consequent and quite evident performance improvement, but it may also pave the way for the provision of advance services such as extended long-term predictions or an autonomous navigation service implemented on-board. The most common approach for any ODTS process is to estimate the satellite orbits and clock parameters by means of a Weighted Least-Square or a Kalman Filter processing, but whereas the satellite orbits are integrated by using a dynamical model with at most 15 parameters, the estimation of the satellite clock parameters is based on an epoch-by-epoch estimation of all clock parameters without taking into account any physical behavior of the satellite clocks. This means that the values of a clock bias at different epochs are considered independent of each other, regardless of their stability. This classical epochby-epoch clock estimation approach has the advantage of being somehow “insensitive” to the typical stochastic behavior of the atomic clocks and to clock jumps, whereas the clear disadvantage is that no a-priori information regarding the clock stability is used within the estimation process. The explanation for this clock estimation approach is that traditionally the stability of the on-board clocks was such that no deterministic model could match the clock behavior to the necessary accuracy level. In this regard, the Galileo satellites are equipped with Passive Hydrogen Masers clocks, which have shown excellent short-term performance only comparable to the Rubidium clocks (Rb) clocks on-board of the GPS Block IIF satellites. This aforementioned high clock stability makes it feasible to accurately parameterize those satellite clocks with a model within the ODTS process, reducing the number of snapshot parameters to be estimated, which will allow to reduce the computational burden, increase the robustness of the estimation process of the other parameters, increase the prediction performances validity and reduce the required ground tracking network, since a continuous satellite tracking by several stations would no longer be required. Furthermore, the clock stability does not only enable the improvement in the performance of the ODTS algorithms, but it would also increase the prediction performance validity of the predicted clocks in the navigation messages. The scope of this paper is to perform a preliminary assessment of the potential station network reduction and clock prediction performance improvement based on the usage of physical clock modelling within the ODTS processing.