A Review of GLONASS

The aim of this paper is to review the implementation of the GLONASS system since the launch of the first satellite and report on the developments in equipment available for its use commercially. Findings of some of the trials that have taken place are provided to give an indication of the advantages offered in comparison with GPS under the various operating modes. The difficulties involved in integrating these two systems and ongoing research activities are identified.

Introduction

Since the launch of the first GLONASS satellite in 1982, four years after the equivalent event for GPS, various operational and political issues have provided a spasmodic development programme. A second satellite system could provide enhanced coverage and increased redundancy to operational users. In addition it provides healthy competition between the system operators. One wonders whether users of GPS would still be at the mercy of the United States Department of Defense if theirs were the only system available. As it happens this is currently the case, but a white paper declaring that GPS would be maintained and made fully available, without interference by the operators, was released at about the time that declarations were made by the operators of GLONASS reviving the plan for implementation of their system.

While the politicians on opposite sides of the globe have been making promises, the manufacturers of commercial equipment have been developing receivers to cope with the alternative systems. They have been spending their development budgets on the basis of their interpretation of declarations made. Many commercial survey companies have been investigating the potential of the different systems, or a combined approach.

Technical Specifications

In comparison with GPS, the principal specifications for the GLONASS constellation are given in Table 1. The greater inclination of the orbital planes for GLONASS will give slightly higher elevation satellites in polar regions, however this advantage is offset to some extent by the lower orbits. It is thought that the variation in the orbital characteristics is due to the ground path requirements. In order to predict ephemerides and upload these for retransmission to the user, the satellite path around the earth must be monitored. In the case of the United States Department of Defense, ground tracking stations can be established internationally. However the former Soviet Union had less opportunity to locate monitoring stations, so it was necessary to ensure that satellites passed over the Soviet Union regularly, and the given constellation achieves this. The difference in the format for transmission of ephemerides is of no significance. It is important that new parameters are computed from recent observations and uploaded regularly.

The satellites in both systems carry atomic clocks and therefore maintain atomic time (IAT), the fundamental unit of which is the SI second. Universal Coordinated Time (UTC) runs at IAT, but integer seconds are inserted as necessary to compensate for variations in earth rotation. UTC(Russia)=UTC+3h and is incremented by leap seconds, GPS is not but the offset between GPS time and UTC is provided.

Table 1. Comparison of GPS and GLONASS specifications, Chalmers (2000)

Receivers are capable of making time observations to within some part of the bit length being measured, current levels are at about 1%. As the bit length within the GLONASS system is double that of GPS, the receiver's capability and hence precision available will be reduced. This is not particularly significant in terms of the error budget for pseudo range observations.

Political Developments

In its early days, GLONASS implementation suffered under the political situation in the former Soviet Union as the republic was dissolved into the various states. Funding a United system became an immediate problem that was finally resolved in 1995, when on March 7th a decree was issued by Chernomyrdin (1995) that promised:

• operation with a full complement of satellites in 1995;
• the formation of a coordination council on use of the system by national and foreign civil users;
• the manufacture of equipment and provision of differential services within the states of the former Soviet Union;
• the provision through the public media of information on system status;
• liaison with international bodies.

Two and a half years later, few of these objectives had been fully realised and a further decree, again issued by Chernomyrdin (1997) on 15th November 1997, identified:

• a number of ministries within Russia that had a requirement for GLONASS;
• that funding would be made available from the federal budget;
• further geodetic applications for the system.

On 18th February 1999 a further decree was issued by Eltsin (1999) which placed the system operation and maintenance outside of the Ministry of Defence of Russia and the Russian Space Agency. These bodies are now customers of the GLONASS system that will be maintained by the Russian federation for dual use. It further identified the need for foreign investment in order to meet the programme and suggested the use of this system as a base on which to create an international navigation satellite system.

Operational Features

The specifications for commercial use are identified as a standard precision navigation signal to provide:

• positioning accuracy within 57-70m (99.7%);
• vertical positioning accuracy within 70m (99.7%);
• velocity factor components within 15cm/sec (99.7%);
• timing within 1mks (99.7%).

These characteristics may be improved by the use of differential or carrier phase techniques.

The Russian federation hosts a network of ground tracking stations across Russia. At various locations across the continent, observations of telemetry data and ranges (to laser reflectors on GLONASS satellites) are made and passed to CGS Moscow. Ephemerides and control information are computed and uploaded to satellites. Under the full constellation at least 5 satellites will be available at any location, at any time worldwide.

Current Status

There have been 28 satellite launches since 1982, and a total of 76 satellites have been placed into orbit. The schedule is given in Table 2. Launch intervals have been irregular from the outset, with lengthy breaks at times that coincide with political events. During perestroika in 1989/90 there was a 12 month period between launches, then 14 months, a year after the dissolution of the Soviet Union which happened early in 1992. Nine satellites were then launched each year for the next two years and it was in this period that the first decree was issued identifying GLONASS as a commercial system and promising a full constellation in the near future. There was then a break for 3 years and it is clear from the decree of 1997 that this was due mainly to financial problems. The irregularity of launches would not be a problem if the satellites were lasting. The average life of about 18 months in the first five years has improved to about 3 years recently. By comparison however it is not unusual for a GPS satellite to remain active for three times this period.

Table 2. GLONASS satellite launches to date, GLONASS history (2000).

According to the information that was used to compile Table 2, as of May 2000 there were 11 satellites active. This information is out of date and a further satellite has been unusable since some time this year. The remaining ten usable satellites occupy slots in the orbital planes as given indicated in Table 3. Channel refers to the frequency of transmission that is different for each satellite.

Table 3. GLONASS satellite constellation at May 2000, Status Information Group (2000).

Six of the ten useable satellites sit in the same plane. Independent monitoring by Status Information (2000), Germany, has classified the observables to these satellites into quality information as given in Table 4. Note that this includes only nine of the satellites given in Table 3 as this source identifies satellite 10 as being in bad health.

• 9 - All functioning satisfactorily within acceptable tolerances
• 8 - significant number of outliers (>10%)
• 7 - Large number of outliers (>25%)
• 6 - Average RMS not good (C1>2m, P1,P2>7m)
• 5 - Average RMS not good and significant number of outliers
• 4 - Average RMS not good and large number of outliers

Table 4. Quality of observables to GLONASS satellites,

Status Information (2000).

Satellites are taken out of service for maintenance for short periods of time, notices are issued to users by the Summary Information Group (2000), and the status is flagged in the satellite message. Figure 1 shows periods for which satellites were unhealthy during the year 2000 to the end of May and Figure 2 indicates typical global coverage at a particular point in time. A mission planning service is available on-line through the GLONASS Updated Information Service (GUIS, 2000a).

Fig. 1: Unhealthy satellites for 144 days in May 2000, GUIS (2000b)

Fig. 2:Typical GLONASS global coverage, Information Analytical Centre (2000)

As it stands there are insufficient operational satellites within the constellation to be of any practical use in positioning a moving platform. There is however the potential to enhance the GPS system with additional GLONASS satellites.

Integrated GPS/GLONASS

There are two problems that exist when combining data from these two systems:

1. Time

When using either GPS or GLONASS we require a minimum of 4 satellites to solve for X Y Z coordinates and time. The solution is available under the assumption that all satellites are synchronised in time and this is achieved by broadcasting a difference between GPS or GLONASS time and the satellite clock. GPS time is maintained by an atomic source that is located at the control centre and the measured difference between this and the satellite clock is used as the correction. GLONASS operates on the same principle, but a clock in Moscow is used as the reference time for the system. While the synchronisation clocks are operating in atomic time, and the offset is known, they are not synchronised together and there is a possibility that some discrepancy exists. Ideally a combined receiver should solve for five parameters X,Y,Z,T GPS and T GLONASS . This being the case, then incorporating a single GLONASS satellite will not contribute to the overall solution, it will merely solve for the offset between GPS time and GLONASS time.

2. Datum

GPS satellites broadcast their navigation information in the form of Keplerian parameters which enable satellite location to be computed in geocentric format with reference to WGS84. The GLONASS format is position, velocity and acceleration in geocentric coordinates that relate to PZ90. The difference between the two spheroids is less than 15m with an average difference of 5m. Misra and Abbot (1994) published results that showed two significant disrepancies between the datums:

1. Z translation of 4m.
2. Rotation about z of 3x10 -6 radians.

Research in this field has proved difficult as little is known about PZ90. Researchers have used the broadcast ephemerides from the satellites themselves to determine the datum. These early results were therefore unreliable and adjustment provided residuals in the order of 30-40 metres. Rossbach, Habrich and Zarraoa (1996) report that in May 1996, three European organisations (Institute of Geodesy and Navigation, Institute of Applied Geodesy, and the German Aerospace Centre) undertook a 5 day observation campaign at 6 ground stations across Europe and into West Africa. Data was shared and used by each organisation independently to coordinate stations in PZ90 and evaluate transformation parameters. Due to large discrepancies between results, and relatively small values for 6 out of the 7 parameters, the conclusion focussed on the z rotation alone as the only significant value at -1.622x10 -6 radians. The two different results have come from observations made in the USA and Europe independently, they both identify a significant rotation about the z axis, but evaluate different magnitudes. At an altitude of 23000km (from the centre of the earth) the satellite position is shifted by 37m using the European result and 69m using the USA result, which would also shift z by 4m.

Integrated Receivers Commercially

Both 3S Navigation in the USA and Skipper in Russia have been building integrated receivers for about the last 10 years. The revived interest and decrees issued by the Russian federation led to further interest and Ashtech released products more recently. Integrated circuits from Javad Positioning Systems (2000) will be available at sometime this year, which may introduce more receivers and reduce cost.

Stand Alone Operation

Figure 3 shows some results obtained by Beser and Balendra (1993) from 3S navigation using a static receiver for a one hour observation period. During this time the constellation was as follows:

The effects of selective availability on GPS can clearly be seen, however GLONASS has an offset due to the unknown transformation parameters at that time. The significant advantage that can be readily identified from the numerical information is the significant improvement in geometry due to the additional satellites. Raby (1994) addressed the problems involved in combining these two data sources, one with a deliberate time variant error and the other with much less noise, but a fixed bias.

Fig. 3: Results from GPS/GLONASS Static Receiver, Beser and Balendra (1993).

Fig. 4: GLONASS Positions, Borjesson, Johansson and Darin (1999)

Fig. 5: GPS Positions, Borjesson, Johansson and Darin (1999)

Fig. 6: Integrated GPS+GLONASS Positions, Borjesson, Johansson and Darin (1999)

More recently, using improved transformation parameters, Borjesson, Johansson and Darin (1999) again showed the advantages of a combined solution. Figure 4 shows GLONASS results for a static receiver. The cluster of solutions in the centre show that GLONASS results are normally within 20m, the outliers are due to poor geometry due to the deficiency in the number of satellites available. Figure 5 shows the same information for GPS, while the geometry is now improved, the scatter plot is distributed over 50m. When the data is combined however, as shown in Figure 6, the scatter is highly concentrated within 20 m.

Differential Systems

Johnston, Shardlow and Toor (1997) undertook field trials to compare single station differential GPS with differential GLONASS, and the combined solution using their Multifix system to provide a bench mark. Results of the comparison of the two systems operating independently are shown in Figures 7 and 8, with the DGLONASS (Ashtech) solution offset from the DGPS (Trimble) by 3m to separate the lines. Large errors in the DGLONASS solution tend to correlate with high HDOP values, but not always. Higher HDOP values exist in the Ashtech solution due to the reduced number of satellites available within the GLONASS constellation in comparison with GPS, hence there are periods of reduced coverage.

Fig. 7: Differential GPS, Johnston, Shardlow and Toor (1997).

Fig. 8: Differential GLONASS, Johnston, Shardlow and Toor (1997).

When the DGPS and DGLONASS are combined, the number of satellites overall increase. Figure 9 shows solutions that are comparable to those of the DGPS while Figure 10 illustrates the significant improvement in HDOP which now rarely exceeds 1. From the graphical information alone it is difficult to identify a clear improvement in performance of the combined approach, but the better geometry must provide an improvement in the stability of the solution. The increased number of satellites will also provide coverage in areas where some part of the sky is masked.

Fig. 9: Integrated Differential GPS and GLONASS, Johnston, Shardlow and Toor (1997).

Fig. 10: HDOP for GPS and Integrated GPS/GLONASS, Johnston, Shardlow and Toor (1997).

Carrier Phase Systems

Fig. 11: Short Base Line Rapid Static Solutions, Martin and Ladd (1997)

For static operations, results from Martin and Ladd (1997) provided in Figure 11, show that a single frequency integrated GPS/GLONASS receiver provides a comparable performance to that of dual frequency GPS. Again this is largely attributed to the improvement in geometry due to the increase in the number of satellites available, although Leick, Beser, Rosenboom and Wiley (1998) identify the need for variation in double difference processing methods from those used in GPS systems.

Integrated RTK (Real-time Kinematic) systems are also on the market and Diggelen (1997) provides results from a commercial system which show that for short baselines initialisation of the combined solution is obtained within 1.5 seconds in 50% of tests undertaken. Over longer baselines however, the GPS alone provides more rapid convergence on the integer ambiguity. While manufacturers would not give away their trade secrets, one must assume that the integrated system provides such performance because it is not truly single frequency. The fact that each GLONASS satellite operates on a different frequency is perhaps used to some advantage in the initialisation of RTK systems over short distances.

Further Research

In 1998 the International GLONASS Experiment (IGEX98 (2000)) was conceived by a number of international organisations to study GLONASS and assess precise satellite orbits, coordinate stations and PZ90/WGS84 relationships amongst other objectives. Between October 1998 and April 1999, observations of code and carrier phase were made at 61 stations (primarily in Europe) in 26 different countries by various organisations using a variety of receivers. Data was pooled and to summarise the primary outcomes identified by Slater, Willis, Beutler, Gurtner, Lewandowski, Noll, Weber, Neilan and Hein (1999):

• Precise ephemerides - are now available from some IGS web sites, for example IFAG (2000). Results are validates between 6 analysis centres to the 20-30cm level.
• Reference frames - transformation parameters between PZ90 and WGS84 are found to vary periodically. This may be due to earth orientation parameters used to generate GLONASS ephemerides. The dominant parameters are confirmed as a rotation about the z axis of 0.3 to 0.4 seconds and a 1.1m translation along the z axis.
• The International GLONASS Service - Pilot Project (IGLOS-PP (2000)) has been instigated from May 2000 to establish a global GLONASS tracking network and provide GLONASS data and precise orbits internationally. Data will be further used to assess GLONASS performance, assess earth orientation parameters and improve other IGS products such as atmospheric information.

Conclusions

In all modes of operation, the additional satellites provided by GLONASS when combined with GPS provide improved coverage, improving the geometry and hence dilution of precision available. This feature is of particular benefit when operating in areas where some part of the horizon is masked and insufficient GPS satellites may be visible to compute a solution at a single epoch. An improvement in the ability to resolve the integer ambiguity under RTK operation for short baselines is also apparent. This is likely to be due to the variation in frequency broadcast by the different GLONASS satellites.

While the variation in reference frames is not completely resolved, the nature of discrepancies is understood and results presented from recent observations show conformality between the GPS and GLONASS systems.

The future of GPS is confirmed by White House (2000). In this recent press release the intention to permanently remove selective availability and make further improvements for commercial users over the next decade is stated. It is also clearly identified that the system will be maintained and made freely available. When considering some of the current applications such as civil aviation, the stage has now been reached where removal of this system for even a short period could jeopardise life. The future of GLONASS is however less certain. With no satellites being launched in the last 18 months, no indication of further launches, and comparatively short life expectancy of GLONASS satellites, there is insufficient trust in the system to warrant capital outlay be many prospective commercial users. Yet some manufacturers have spent significant sums in the development of equipment that will provide this capability. As a consequence the cost of acquiring access to integration is still high and while this may reduce over time, justification of this outlay on the basis of the reliability and future GLONASS may be difficult.

Further Reading

1. Full technical specifications are available from the official GLONASS Interface Control Document (revised 1998), published in English by the Russian Space Forces which may be downloaded from ftp://ftp.nz.dlr.de/nav/glonass/ICD/GLONASS_ICD.pdf
2. A comprehensive document detailing operational features and computations is provided by Heinz Habrich in Geodetic Applications of the Global Navigation Satellite System (GLONASS) and of GLONASS/GPS Applications which can be downloaded from: ftp://igs.ifag.de/dist/habrich_glonass.pdf
   

References

Beser J. and Balendra A. (1993). Integrated GPS/GLONASS Navigation Results, 3S Navigation.

Borjesson J. Johansson J and Darin F. (1999). GLONASS: Experiences From The First Global Campaign. Radio Vetenskap pch Kommunikation 99 (RVK99), Karlskrona, Sweden, June 14-17.

Chalmers (2000). http://www.oso.chalmers.se/~geo/gg_comp.html

Chernomyrdin (1995). http://www.rssi.ru/sfcsic/decree.html

Chernomyrdin (1997). http://www.rssi.ru/sfcsic/decree2.html

Van Diggelen F. GPS and GPS+GLONASS RTK. Proceedings of the Satellite Division of the Institute of Navigation 10th International Technical Meeting, ION GPS-97, Kansas City, September 16-19.

Eltsin (1999). http://www.rssi.ru/sfcsic/38rp-e.html

GLONASS history (2000). http://www.rssi.ru/sfcsic/hist_e.html

GUIS (2000a). http://www.nz.dlr.de/gps/glonass2.html

GUIS (2000b). ftp://ftp.kn.nz.dlr.de/glonass/status_i nf/nagu.gif

IFAG (2000). ftp://igs.ifag.de/IGEX

IGEX-98 (2000). http://164.214.2.59/GandG/ion/index.htm

IGLOS-PP (2000). http://igscb.jpl.nasa.gov/projects/i glos/glonass.html

Information Analytical Centre (2000). http://www.mcc.rsa.ru/IACKVO/dnoiseen1.html

Javad Positioning Systems (2000). http://www.javad.com

Johnston, G. Shardlow, P. and Toor P. (1997). GLONASS Receiver Evalution Report, CSO 196, Internal Racal Report, February 1997.

Leick A. Beser J. Rosenboom P. and Wiley B. (1998). Assessing GLONASS Observations. Proceedings of the Satellite Division of the Institute of Navigation 11th International Technical Meeting, ION GPS-98, Nashville, September 15-19.

Martin W. and Ladd J. (1997). GPS+GLONASS Surveying: Post Processed and Real-Time Results. Proceedings of the Satellite Division of the Institute of Navigation 10th International Technical Meeting. ION GPS-97, Kansas City, September 16-19.

Misra P.N. and Abbot R.I. (1994). SGS85-WGS84 Transformation. Manuscripta Geodaetica, No 19.

MIT Lincoln Laboratory (2000). http://satnav.atc.ll.mit.edu/

Raby P. (1994). Integrating Signal information From GPS and GLONASS. University of Leeds. Department of Electronic and Electrical Engineering. August 1994.

Rossbach U. Habrich H. and Zarraoa N. (1996). Transformation Parameters Between PZ90 and WGS84. Proceedings of the 9 th International Technical Meeting of the Satellite Division of the Institute of Navigation, ION GPS-96, Kansas City, September 17-20.

Slater J.A. Willis P. Beutler G. Gurtner W. Lewandowski W. Noll C Weber R. Neilan R.E. and Hein G (1999). The International GLONASS Experiment (IGEX-98): Organisation, Preliminary Results and Future Plans. Proceedings of the Satellite Division of the Institute of Navigation 12th International Technical Meeting, ION GPS-99, Nashville, September 14-17.

Status Information (2000). ftp://ftp.kn.nz.dlr.de/glonass/status/upd ate.inf

Status Information Group (2000). http://www.rssi.ru/sfcsic/nagu.txt

Summary Information Group (2000). http://www.rssi.ru/sfcsic/nagu.txt

White House (2000). http://www.navcen.uscg.mil/gps/policynot es/SA.htm