Clocks, Geodesy, and Differential Corrections

If you are a scientist and need to squeeze the last millimeter of accuracy and picosecond of time from GNSS measurements, then any difference between system elements is significant. Examples include location of the phase center of the transmitting antenna relative to the satellite’s center of mass, location of the tracking stations used to determine the orbit and clock of each satellite, the geodetic models which predict satellite orbits, the effect of continental drift and polar wander on orbit determination, the short-term and long-term stability of satellite clocks, the wander of each satellite clock relative to a national Universal Time Coordinated (UTC) standard, and how close the national time standard is to the global UTC.

In addition, scientists must be concerned with minute aberrations in signal structure, ionospheric and tropospheric refraction effects, variability in receiver antenna characteristics, including phase variations with azimuth and elevation angle, the local multipath environment, the physical stability of structures holding the antennas, and so on. Each one of these and more have been the subject of intense and ongoing study.

None of these is of much concern to consumers. Most GNSS receivers are in mobile phones (three billion or more), followed by automobiles. Accuracy is challenged by multipath, signal blockage, poor antenna structures, very limited battery power (which limits sophisticated processing), severely constrained physical space, an environment filled with electronic noise, and so on. A few meters of error apparently is tolerable and has not limited the growth of these markets.

In between these quite different applications are professional and commercial uses. The top accuracy requirement is about half a centimeter for survey, machine control, structure monitoring, volcano monitoring, earthquake monitoring, and so on. Perhaps surprisingly, many precision agriculture applications require accuracy between 1 and 10 cm. Another set of applications requires between half and one meter of accuracy.

The most prominent example is data collection for geographic information systems (GISs), such as locating and mapping physical structures. Less accuracy is needed for commercial navigation applications for aircraft, ships, boats, trucks, etc. More accuracy is needed when landing aircraft or navigating large ships in shallow waters.

The above paragraphs are meant to show that there are a vast number of GNSS applications, each with its own accuracy and other requirements. Therefore, the need for interoperability of different GNSS signals differs from one application to another. For general navigation, it appears interoperability has nearly been achieved. This is because every GNSS provider is striving to improve its accuracy and adhere to international standards. The international standard for time is Coordinated Universal Time, which, in French, is Temps Universel Coordonne. UTC was chosen as a compromise abbreviation. UTC is based on International Atomic Time (TAI) which is computed by the International Bureau of Weights and Measures (BIPM) located near Paris.

Interoperability at the level of about 10 ns (approximately 3 m at the speed of light) seems achievable in the near future. Even so, it is recommended that the measured time offset between systems be obtained. Sources include external measurements provided as part of satellite navigation messages or via other communication channels. The most accurate method is for the receiver to calculate time offsets when there are enough satellites visible from each system to include the system time offset as an unknown parameter in the navigation solution. Because the time offset between systems changes so slowly, it can be heavily filtered after the solution has settled.

The other standard which all GNSSs are attempting to replicate is known as the International Terrestrial Reference Frame (ITRF). The difference between the latest GPS reference frame and the latest ITRF is only a few centimeters. All other GNSSs are also approaching this level of agreement. Therefore, for applications which require no more than a few meters of accuracy, it is now - or soon will be - acceptable to combine signals from several GNSSs without concern for time or geodetic interoperability.

For applications requiring more accuracy, the answer for many decades has been the use of differential corrections. The basic concept is that one reference station at a defined location, or a network of reference stations at defined locations, tracks the available satellite signals and compute corrections which are supplied to users. There are differential systems with one reference station which operate over a limited distance of between 15 and 50 km from the reference station, and there are other systems with reference stations distributed over broad areas which can serve an entire continent.

Over short distances, the corrections may be only pseudorange adjustments or, for survey and machine control applications, they also include carrier phase readings. For larger systems, the corrections will include adjustments to each satellite’s orbit parameters as well as pseudorange and/or clock corrections. Some applications permit simple corrections; others may demand a more complex set of corrections. Systems with large coverage areas include the WAAS operated by the US Federal Aviation Administration (FAA) and similar SBASs around the world.

There are private systems operated by commercial companies such as OmniSTAR by Trimble, Ltd., StarFire by John Deere, and TerraStar by TerraStar GNSS Ltd. Each of these services employs communication satellites to distribute correction messages. Smaller systems use local radio transmitters and receivers to send correction messages.

Importantly, all differential systems eliminate basic interoperability issues of time or geodesy offsets. After differential corrections are applied, the time reference and the orbit coordinates for each satellite are adjusted to agree with common references, as defined by the differential system itself. The only remaining interoperability issues will occur in the user equipment, such as different time delays due to different signal center frequencies. Differential systems are widespread and widely used, even in cell phone networks, and they all but eliminate most concerns about interoperability.

References:
1 http://www.unoosa.org/oosa/en/ourwork/icg/icg.html
2 http://www.unoosa.org/documents/pdf/icg/activities/2007/WG-A-2007.pdf
3 http://www.unoosa.org/oosa/en/ourwork/icg/working-groups.html

4 Pullen, S. and Joerger, M., Chapter 23 GNSS Integrity and Receiver Autonomous Integrity Monitoring (RAIM), in Position, Navigation, and Timing Technologies in the 21^st Century, (eds. Y.J. Morton et al.), Wiley-IEEE Press, 2020.

5 Walter, T., Chapter 13 SBAS, in Position, Navigation, and Timing Technologies in the 21st Century (eds. Y.J. Morton et al.), Wiley-IEEE Press, 2020.

6 http://www.gps.gov/policy/cooperation/europe/2007/MBOC-agreement/
7 Lu, M. and Zheng, Y., Chapter 6 Beidou Navigation Satellite System, in Position, Navigation, and Timing Technologies in the 21st Century (eds. Y.J. Morton et al.), Wiley-IEEE Press, 2020.

8 Karutin, S., Testoedov, N., Tyulin, A., and Bolkunov, A., Chapter 4 GLONASS, in Position, Navigation, and Timing Technologies in the 21st Century (eds. Y. J. Morton et al.), Wiley-IEEE Press, 2020.

9 Karutin, S., Private communication.
10 Parkinson, B.W., Morton, Y.J., van Diggelen, F., and Spilker, J.J., Chapter 1 Introduction, Early History, and Assuring PNT (PTA), in Position, Navigation, and Timing Technologies in the 21st Century (eds. Y. J. Morton et al.), Wiley-IEEE Press, 2020.