High precision GPS network with precise ephemerides and earth body tide model

Rede GPS de alta precisão usando efemérides precisas e modelo de maré terrestre

NOTAS BREVES DE PESQUISA & DESENVOLVIMENTO TÉCNICO SHORT RESEARCH & TECHNICAL DEVELOPMENT NOTES

High precision GPS network with precise ephemerides and earth body tide model

J. F. G. Monico¹, V. Ashkenazi² & T. Moore²

¹Unesp- Campus de Presidente Prudente

Rua Roberto Simonsen 305, CP 957, Fone [phone], Fax: [phone]

CEP 19060.900 Presidente Prudente, SP, Brazil, e-mail: [e-mail]

²Institute of Engineering Surveying and Space Geodesy

The University of Nottingham

University Park, NG7 2RD Nottingham, UK.

The permanent GPS network of the International GPS Service for Geodynamics (IGS) provides a global data set of precise GPS measurements, which is invaluable in the development of algorithms and software for very high accuracy GPS positioning. The intensive EPOCH'92 campaign expanded the core network to include densification in many regions of the world. This project has involved the analysis of data from the Brazilian GPS network, with selected connections to the IGS Global Network. In this preliminary report we discuss the initial data processing, demonstrating a level of relative positioning accuracies of better than 10 ppb in height differences, typically the worst positional component. Besides the normal standard used in the processing of a GPS network, it involved an Earth body tide model.

Key words: GPS; Precise ephemerides; Earth body tide model.

Rede GPS de alta precisão usando efemérides precisas e modelo de maré terrestre - A rede GPS do IGS (International GPS Service for Geodynamics) proporciona uma série de dados GPS global preciso, os quais são de grande valor para o desenvolvimento de algoritmos e programas computacionais para posicionamento GPS de alta precisão. A campanha intensiva denominada EPOCH '92 expandiu o núcleo da rede permitindo densificar muitas regiões do mundo. Neste projeto analisou-se os dados de uma rede brasileira com conexões para a rede global IGS. No artigo discute-se o processamento dos dados que demonstraram um nível de precisão relativa melhor que 10 ppb (partes por bilhão) para diferenças de altitudes, normalmente a pior componente. Além dos procedimentos normais do processamento de uma rede GPS, incluiu-se um modelo para Maré Terrestre.

Palavras-chave: GPS; Efemérides precisas; Modelo de Maré Terrestre.

INTRODUCTION

The last 10 years have seen a steady rise in the level of relative positioning accuracies achieved by the Global Positioning System (GPS). A fundamental development was the adoption of fiducial GPS techniques, and the subsequent methods for ambiguity resolution over very long baselines. Since then, refinements in data processing techniques and physical modelling have gradually been introduced. Only a few years ago, the goal of many research groups was to achieve, reliable and repeatable accuracies at the one part-per-million (ppm) level. Current results are now demonstrating relative positioning capabilities at the level of few parts-per-billion (ppb).

The permanent GPS network of the International GPS Service for Geodynamics (IGS) provides a global data set of precise GPS measurements, which is invaluable in the development of algorithms and software for very high accuracy GPS. The intensive EPOCH'92 campaign expanded the core network to include densification in may regions of the world. This project has involved the analysis of data from the Brazilian GPS network, with selected connections to the IGS Global Network. In this preliminary report we discuss the initial data processing, demonstrating a level of relative positioning accuracies better than 10 ppb in height differences, typically the worst positional component.

To achieve the level of precision described above, the fiducial GPS technique was applied, and a set of models had to be taken into account. The main aim was to investigate the influence of Earth Body Tides (EBT) on the processing of very large network.

IGS EPOCH '92 CAMPAIGN: BRAZILIAN STATIONS

IGS was established in 1993, by the International Association of Geodesy (IAG), to provide support to geodetic and geophysical research activities with GPS. The first IAG test campaign, Epoch '92, was held from 14 June to 23 September 1992 and showed the near real-time capability of the global GPS community to retrieve data and compute products such as satellite ephemerides and Earth rotation parameters (Noll & Dube, 1993).

The Brazilian participation in the Epoch '92 campaign involved the Regional Center (Bergamini, 1993). It included several organizations, namely the Federal University of Paraná (UFPr), Brazilian Institute of Geography and Statistics (IBGE), the University of São Paulo (USP), the Paulista State University (UNESP) and the National Institute for Space Research (INPE). They cooperated with the international scientific community and provided GPS data to the geodetic community in Brazil for the establishment of new GPS stations. Three stations (Paraná, Presidente Prudente and Brasilia) were continuously occupied during 14 days of the IGS Epoch '92 Campaign.

In order to connect the stations occupied in Brazil during the Epoch '92 campaign to an international reference frame, e.g. IERS Terrestrial Reference Frame (ITRF), the GPS data were processed with other stations outside Brazil, which were also occupied during this campaign. These stations were Santiago in Chile, and Goldstone and Richmond in the United States. The stations of this network are shown in Fig. 1.

Figura 1 -
Brazilian Epoch '92 Campaign.
Figure 1 - Participação Brasileira na Campanha Epoch'92.

The three Brazilian sites belonging to the IGS Epoch '92 campaign (BRAS, UEPP, PARA), and the other stations included in this processing are listed in Tab. 1. The receiver at station PARA was a Trimble 4000 SST (L1 and L2 carrier-phase with P-code). The other two receivers were Trimble 4000 SD with C/A-code (squaring type). The stations outside Brazil (SANT, RCM2, GOLD) are permanent IGS stations equipped with Rogue RN8 receivers.

Table 1
- Stations of the Brazilian Epoch '92 Campaign.
Tabela 1 - Estações Brasileiras na Campanha Epoch'92.


Although more data were available for the stations belonging to the Epoch '92 campaign, only six days (208 - 213) were included in the processing. This may provide a representative sample for the purpose of this study.

SOFTWARE AND PROCESSING STRATEGIES

GPS Analysis Software (GAS)

The data were processed using GPS Analysis Software (GAS) developed at the University of Nottingham (Stewart et al., 1994). GAS process either double-difference carrier phase measurement or pseudo-ranges, formed with different base satellites for each baseline processed. This provides a capability to process very large GPS network or even a global GPS network.

The GPS network processing by GAS can be carried out either with the full fiducial technique (Ashkenazi et al., 1989) or by using the broadcast or the precise ephemerides for the satellites. In all cases, corrections for the effects of the Earth body tide (EBT) and Ocean tide loading (OTL) can be included in the processing. For the former case, the model recommended by the International Association of Geodesy was used (McCarthy, 1992). The effects of the ionosphere can be greatly reduced, by using the ionospheric free observable (Ffoulkes-Jones, 1990). For the tropospheric delay, GAS provides a series of modelling options. For example, one could define zenithal scale factors as unknown parameters in the adjustment. GAS also provides the option of 'fixing' the double-difference carrier phase integer ambiguities to their integer values. This is achieved by one of the two different techniques, namely the ambiguity search technique and the sequential bias fixing (Ffoulkes-Jones, 1990).

Processing Strategies

In order to determine the coordinates of the Brazilian Epoch'92 stations consistent with the ITRF, the effects of the body tides of the Earth were also investigated by considering their influence on the repeatability of the solutions, when processing different periods of data. To provide the connection to the ITRF, the coordinates of stations GOLD, RCM2 and SANT were held fixed at their known values, once they had been transformed to the epoch of the campaign. The coordinates were taken from a Jet Propulsion Laboratory (JPL) global, fiducial-free, GPS solution, which had been mapped onto the ITRF (Blewitt et al., 1993). The precise ephemerides of the satellites were also taken from JPL, and were held fixed in the subsequent processing at Nottingham. Tab. 2 shows the 5 baselines processed, together with the respective lengths.

Table 2
- Length of the baselines processed.
Tabela 2 - Comprimento das Linhas de Base Processadas.


The atmospheric modelling options used were based on standard processing techniques and experience at Nottingham. The ionospheric free phase observable was used and the double-difference ambiguities estimated as real values. The tropospheric refraction was modelled by applying the Magnet model. Additionally, two zenithal scale factors per-day-and-station were modelled, as a first order polynomial and estimated as unknown parameters in the network adjustment.

Each of the six days was divided into two sessions of 12 hours. The data was processed by adopting four different scenarios. Each day of the data set was either processed as a full 24 hour period, or as two consecutive 12 hours sessions. In both cases the effect of the Earth body tide (EBT) was investigated, by repeating the solution both with and without the correction implemented. Tab. 3 summarizes these four scenarios.

Table 3
- Tested Scenarios.
Tabela 3 - Experimentos Testados.


RESULTS AND DISCUSSION

The repeatability of each of the baseline components of the different solutions was computed. This simply represents the scatter of the estimated daily values about their weighted mean value, for each individual baseline (Overgaouw et al., 1994) and provides an indication of the consistency of the results.

Fig. 2 (a) shows the repeatability for scenario A, which was obtained from 12 solutions of 12 hours each. The height component, which is typically the worst one, shows rms (root mean square) of about 70 mm. In Fig. 2 (b) , which shows scenario B, i.e. the 24 hour solutions, the same value has been reduced to 50 mm. Neither scenario A or B includes the Earth body tide corrections. However, as solution B is computed over a 24 hour period, this reduction may be attributed to the Earth body tide and ocean tidal loading effects partially cancelling out, over the longer period.

Figure 2(a)
- 12 hours solution;
Figure 2(b) - 24 hours solution;
Repeatability without Earth Body Tides corrections.


In scenarios C and D, Earth body tide corrections have been applied. The resulting rms of the height component of scenario C reduces from 70 mm (in scenario A) to about 35 mm, as illustrated in Figure 3(a). This also represents an improvement over the 24 hour solutions without Earth body tide (scenario B), but predominantly for the longer baselines, over 1000 km.

Figure 3(a)
- 12 hours solution;
Figure 3(b) - 24 hours solution;
Repeatability with Earth Body tides corrections.

Once Earth body tide corrections are applied to the six 24-hour solutions, the rms in the height component reduces to only 20 mm, as shown in Figure 3(b). This constitutes an improvement with respect to the 12-hour solution (in scenario C) of the order of 10 mm for all baselines. It might be due to the ocean tide loading cancelling out over the longer period. No corrections for ocean tide loading were applied during these preliminary experiments. To summarize, the final scenario, involving six 24-hour solutions with Earth body tide corrections applied, provides an average rms of the order of 8 ppb for the worst component (height). The repeatability of the length component is better than that of the height by a factor of 2, i.e. an average rms for lengths of the order of 4 ppb applied in this experiments.

CONCLUSIONS

It has been shown that one can achieve very high position and height accuracies, by using an IGS Precise Ephemerides, such as that provided by JPL. This has been tested by using data from the Brazilian Epoch '92 GPS Campaign, supplemented with a number of other IGS sites, thus creating long baselines extending up to 9000 km.

In this preliminary set of tests, the intention was to determine the accuracies which could be achieved by using the IGS Precise Ephemeris, and the simplest of modelling options in current practice. The results showed that for the highest precision, the corrections derived from an Earth body tide model have to be taken into account in the processing. At such case, precision at the ppb level can be achieved.

ACKNOWLEDGMENTS

This work was carried out with the financial support of CNPq- of the Secretary for Science and Technology of Brazil and UNESP - Campus de Presidente Prudente. The authors wish to acknowledge the organizations in Brazil which took part in the Epoch '92 Campaign, namely IBGE, INPE, UNESP, UFPr and USP. Special thanks are also due to Dr. Denizar Blitzkow of the University of São Paulo, for his support and useful discussions during this project.

Submetido em: 01/05/97

Revisado pelo(s) autor(es) em: 19/11/97

Aceito em: 25/11/97

REDE GPS DE ALTA PRECISÃO USANDO EFEMÉRIDES PRECISAS E MODELO DE MARÉ TERRESTRE

A rede GPS do IGS (International GPS Service for Geodynamics) proporciona uma série de dados GPS global preciso, os quais são de grande valor para o desenvolvimento de algoritmos e programas computacionais para posicionamento GPS de alta precisão. Um dos algoritmos de fundamental importância envolve a técnica fiducial bem como a solução de ambigüidade. Durante a fase de validação do IGS, realizou-se a campanha intensiva denominada EPOCH '92 que expandiu o núcleo da rede permitindo a densificação de muitas regiões do mundo. Durante esta campanha, três estações foram ocupadas no Brasil. Neste projeto analisou-se os dados destas estações com conexões para a rede global IGS, via três estações, duas nos Estados Unidos da América e uma no Chile. Para o processamento utilizou-se o programa GAS (GPS Analysis Software), desenvolvido na Universidade de Nottingham, UK. GAS permite o processamento de redes GPS globais, usando a técnica fiducial com correções ou não de maré terrestre. A observável básica pode ser a dupla diferença de fase da portadora ou da pseudo-distância, com várias opções de modelos para a troposfera. Fundamentalmente, as estratégias para o processamento dos dados, incluíram efemérides precisas e a aplicação ou não de modelos de maré terrestres, considerando intervalos de 12 e 24 horas. Utilizou-se a combinação linear que reduz consideravelmente os efeitos da refração ionosférica (ionospheric free observable). Alguns dos resultados obtidos demonstraram nível de precisão relativa da ordem de 10 ppb (partes por bilhão) para diferenças de altitudes, normalmente a pior componente. Além dos procedimentos normais do processamento de uma rede GPS, incluiu-se um modelo para Maré Terrestre.

NOTES ABOUT THE AUTHORS

NOTAS SOBRE OS AUTORES

João Francisco Galera Monico

Formação Acadêmica: Engenheiro Cartógrafo, Mestre em Ciências Geodésicas, PhD em Engenharia de Levantamentos e Geodésia Espacial.

Áreas de Interesse: Controle de Qualidade em Geodésia, Posicionamento Geodésico de Alta Precisão, Fundamentos de GPS e GLONASS, DGPS e GPS para GIS.

Vidal Ashkenazi

Formação Acadêmica: Engenheiro Civil, PhD em Ciências Geodésia e Ciências Fisicas

Áreas de Interesse: Wide Área DGPS. Controle de Deformações. Posicionamento Geodésio Espacial.

Terry Moore

Formação Acadêmica: Engenheiro Civil, PhD em Geodésia Espacial

Áreas de Interesse: Posicionamento Geodésio com GPS e SLR Wide Área DGPS, Navegação

• ASHKENAZI, V., MOORE, T., FFOULKES-JONES, G.H., WHALLEY, S. & AQUINO, M. ž 1989 - High Precision GPS Positioning by Fiducial Technique. In Global Positioning System: An Overview (I. Mueller, ed.). IAG Symposia 102, 1: 195-202.
• BLEWITT, G., HEFLIN M. B., VIGUE Y., ZUM-BERGE J. F., JEFFERSON D. & WEBB F. H. ž 1993 ž The Earth viewed as a deforming polyhedron: method and results. In 1993 IGS Workshop (G. Beutler & E. Brockmann, eds.). 165-174.
• BERGAMINI, E.W. ž 1993 - IGS Service: Epoch '92 Regional Centre in Brazil. In 1993 IGS Workshop (G. Beutler & E. Brockmann, eds.). 25-31.
• FFOUKES-JONES, G.H. ž 1990 - High Precision GPS by Fiducial Techniques, PhD Thesis, University of Nottingham, 132pp.
• McCARTHY, D. D. ž 1992 - IERS Standards (1992), IERS Technical Note 13, Central Bureau of IERS - Observatoire de Paris.
• NOLL, C.E. & DUBE, M.P. ž 1993 - Developing of a GPS Data Catalogue for the IGS Epoch '92Campaign. In 1993 AGU Meeting.
• OVERGAOUW, B., AMBROSIUS, B.A.C. & WAKKES, K.F. ž 1994 - Analysis of the EUREF-89 GPS data from the SLR/VLBI sites. Bulletin Géodésique, 68:19-28.
• STEWART, M. P., FFOULKES-JONES, G. H. & OCHIENG, W. Y. ž 1994 - GPS Analysis Software (GAS) Version 2.2 User Manual, IESSG Publication, University of Nottingham, UK., 145pp.

Publication Dates

• Publication in this collection
  08 June 2001
• Date of issue
  July 1997

History

• Reviewed
  19 Nov 1997
• Received
  01 May 1997
• Accepted
  25 Nov 1997