Resumen de Mejora de la estabilidad a medio y largo plazo de la transferencia de tiempo por microondas
The realization and maintenance of an international reference time scale is essential for a great number of state-of-the-art technological applications, such as telecommunications or navigation. Since 1975, the reference time scale for world-wide time coordination has been the Coordinated Universal Time (UTC), which is created from the International Atomic Time (TAI), adding to it the so-called "leap seconds", in order to keep UTC close to the time derived from the Earth rotation (mean solar time). TAI is a uniform time scale that provides a precise reference for scientific applications. It is realized and maintained by the International Office of Weights and Measures (BIPM), from a weighted mean of about 400 atomic clocks around the world. The realization of an atomic time scale, such as TAI, is based on the readings of the participating clocks, so it requires methods for comparing these distant clocks (time and frequency transfer, TFT). The realization of an atomic time scale then, not only relies on the quality of the atomic clocks that participate to it but also on the quality of the methods for comparing them. With the constant improvement of the stability and accuracy of the atomic clocks, TFT methods should be increasingly accurate and precise, if we want that they contribute to improving the quality of time scales. UTC relies today on a network of international time links for clock comparisons based on GNSS (Global Navigation Satellite Systems), concretely GPS and recently GLONASS, but also on the TWSTFT (Two-Way satellite Time and Frequency Transfer) technique, used in about 20% of the time laboratories. The objective of this research was to develop new time transfer techniques with better stability in the medium- and long-term using GNSS systems and TWSTFT.
Since the UTC time transfer network is highly redundant, i.e., several time links are maintained through more than one technique, the first work was focused on combining both methods (with in particular the GPS navigation system data) with the principle of keeping the best characteristics of each one, i.e., a good time resolution and a frequency stability at 10^{-15} at one day from GPS, and the 1-ns accuracy of TWSTFT. We developed an original method for combining GPS with TWSTFT, that is able to compare remote atomic clocks with independence of the baseline length. The main purpose of the combination of GPS and TWSTFT is to calibrate the ambiguous GPS phase measurements with the TWSTFT data, which have a better accuracy than the GPS codes. We have demonstrated that for typical TWSTFT links with a precision of 0.5 ns, the combined solution with GPS and TWSTFT keeps the 1-ns accuracy of the TWSTFT data, reaches the 10^{-15} frequency stability at integration times of one day, solves for the large day-boundary discontinuities existing in some geodetic clock solutions, and has day-boundary discontinuities with a standard deviation similar to the discontinuities of the geodetic solutions for the GPS stations having the best performances in that terms. Moreover, we have shown that the combination approach is also able to remove the large day-boundary discontinuities associated with the GPS-only time transfer solutions, either due to GPS hardware delay changes, or due to the coloured signature of the noise of GPS codes. The long-term stability of the combined solution depends on the long-term stability of the TWSTFT link, as the combined solution is fitted on the TWSTFT data.
The second research done in this project was focused on the use of the new code signal E5AltBOC, provided by the future European Galileo, for time transfer applications. The current main limitations of GNSS time transfer are the noise level and impact of multipath of the GPS and GLONASS code measurements. The Galileo E5AltBOC code, exhibiting a very low range noise and multipath error down to the values of 20 cm, has revealed itself as a very promising signal for improving the medium-term stability of time transfer. For achieving this goal, this very precise code must not be combined with any other code, as it is currently done for GNSS time transfer at the most accurate level, because, in such a case, we would lose the advantages of its precision. The main limitation in such a case is to determine the contribution to the signal delay due to the ionosphere. Our investigations have led to the two following approaches for using the E5AltBOC code:
1.- Using the E5AltBOC measurements after correcting them for the ionospheric delay. To benefit from the precision of E5AltBOC, a sub-nanosecond precision in the ionospheric delay must be achieved. The current most precise way of determining the ionospheric delay that has been found is by means of a dual-frequency combination of code and carrier-phase measurements.
2.- Using a new observable, the code-plus-carrier (CPC) combination, applied to E5AltBOC. CPC E5 is obtained from the addition of the code and carrier-phase measurements on the same frequency. The advantage of this observable is that it is at first-order ionosphere-free, as the GPS P3 observable, but without increasing the noise and multipath of the code measurement. In fact, by means of current available GPS and GIOVE data, we have demonstrated that the CPC E5 combination is more than 15 times less noisy than GPS P3. Besides, this combination has a reduction of factor 2 in the noise and multipath with respect to the E5AltBOC code measurement. CPC is although an ambiguous observable. Thereafter, it cannot be used for time transfer unless it is first corrected for its ambiguity term. In order to solve for it, we have used another combination called code-minus-carrier, that results from doing the difference between the code and the carrier-phase measurements on a same frequency, because it has the same ambiguity that appears in the CPC observation equation of E5. However, the code-minus-carrier observable needs to be corrected for the ionospheric delay first. To that purpose, the same strategy used in the first approach has been applied.
We have demonstrated both analytically and using simulated or true data, that both strategies lead to a time transfer solution that is equally affected by the long-term multipath as the solution obtained using the ionosphere-free dual-frequency code combination. This is due to the fact that a second code has been used for determining the ionospheric delay, and throughout the process, the long-term multipath error of this second code is transferred to the time transfer solution. Moreover, the contribution of the multipath of the E5AltBOC code is increased as well. Therefore, the properties of the very precise Galileo E5AltBOC code are lost. Only the short-term stability of time transfer could be improved if we used the second strategy proposed, i.e., the E5CPC approach. Nevertheless, we have also seen that the CPC E5 is still 3 times noisier than a dual-frequency ionosphere-free combination of carrier phases. Hence, this latter is still better for improving the short-term stability of time transfer. According to that, we have proposed a third approach that consists of using the ionosphere-free carrier-phase combination with L1 and L5, and determine its ambiguity with respect to the CPC E5 observable. This last strategy should lead to the highest improvement in the medium-term stability of time transfer with the E5AltBOC code, if we are able to find another strategy for solving for the ambiguity of CPC E5.
