GNSS-to-GNSS time offsets: study on the broadcast of a common reference time

Abstract

We present two different approaches to broadcasting information to retrieve the GNSS-to-GNSS time offsets needed by users of multi-GNSS signals. Both approaches rely on the broadcast of a single time offset of each GNSS time versus one common time scale instead of broadcasting the time offsets between each of the constellation pairs. The first common time scale is the average of the GNSS time scales, and the second time scale is the prediction of UTC already broadcast by the different systems. We show that the average GNSS time scale allows the estimation of the GNSS-to-GNSS time offset at the user level with the very low uncertainty of a few nanoseconds when the receivers at both the provider and user levels are fully calibrated. The use of broadcast UTC prediction as a common time scale has a slightly larger uncertainty, which depends on the broadcast UTC prediction quality, which could be improved in the future. This study focuses on the evaluation of two different common time scales, not considering the impact of receiver calibration, at the user and provider levels, which can nevertheless have an important impact on GNSS-to-GNSS time offset estimation.

Impact of Galileo-to-GPS-Time-Offset accuracy on multi-GNSS positioning and timing

“…On the other side, using the UTC as a pivot to get GGTO can induce an error in GGTO so-determined, as the UTC predictions provided by the different GNSS are not equal. In Sesia et al (2020), an error of up to 20 ns was shown for an inter-system bias assessed from this method. We consider this as an extreme case, keeping in mind that the GNSS systems are continuously improving as well as their realizations of UTC.…”

“…Signorile et al (2018) propose two options for this reference: either an average of the 4 GNSS time scales (GPS, Galileo, GLONASS and BeiDou) that can be computed easily by the 4 GNSS providers, or the prediction of UTC provided in the navigation messages of each GNSS. Sesia et al (2020) show that the first approach (average of GNSS times) provides very good performances in terms of XYTO accuracy but requires adaptations at the system level. For what concerns the second approach, using the broadcast UTC information as reference does not require any change at the system level, but generates an error that can be as large as 20 ns on the XYTO, depending on the quality of the broadcast UTC prediction.…”

Precise Onboard Time Synchronization for LEO Satellites

“…The officially stated accuracy of Galileo's GGTO values ranges from 2.2 ns to 4.6 ns during the first quarter of 2021 (GSA, 2021). Sesia et al (2021) showed that the broadcast GGTO varied approximately ± 5 ns over the course of one month in 2018. The GGTO provided in the INAV and FNAV messages is represented by a first-order polynomial, allowing the user to obtain an epoch-wise value for time correction relative to GPS.…”

“…As described in Section 3, the ISB lumps together system time differences and GNSS/signal-specific receiver biases. The actual time difference between GPST and GST is known to be at the few-nanosecond level with slow variations (Sesia et al, 2021). Furthermore, post-processing showed that the physical biases of the PODRIX receiver between Galileo and GPS observations were almost constant over one day (Montenbruck et al, 2021).…”

“…Predicted GNSS-UTC offset parameters are broadcast via the system's respective navigation messages to allow a user to derive UTC based on the obtained GNSS time. Sesia et al (2021) examined the quality of broadcast GNSS-UTC time offset predictions and the feasibility of using UTC as a common timescale for multi-GNSS processing. They specifically highlighted the fact that broadcast GNSS-UTC offsets of various constellations refer to different UTC realizations.…”