Special Study Group of the IAG 1.157 - Ambiguity Resolution and Validation
Paul de Jonge
Scripps Institution of Oceanography
University of California, San Diego, USA
Since the founding of the special study group at the IUGG meeting in Boulder in 1995, GPS ambiguity resolution and validation continued to be an important research topic. Of the 325 papers collected in the on-line bibliography on ambiguity resolution and validation (http://www.geo.tudelft.nl/mgp/people/paul/amb.html) 177 are from the years 1995-1999:
|1981 1||1982 0||1983 0||1984 0||1|
|1985 4||1986 1||1987 3||1988 6||1989 10||24|
|1990 13||1991 14||1992 30||1993 27||1994 39||123|
|1995 34||1996 35||1997 59||1998 38||1999 11||177|
This large interest for ambiguity resolution is partly driven by the interesting scientific aspects, and partly by the large commercial interests in the area.
Traditionally, algorithms for resolving the GPS ambiguities have been developed for two different applications. On the one hand methods have been devised for applications where a multiple of stations are occupied for several hours until several days, and maximum inter-station distance can be of the order of thousands of kilometers.
At the moment (June 1999), four of the seven IGS processing centers are, in their global networks, resolving a varying percentage of the ambiguities for baselines up to several thousands of kilometer.
On the other hand methods have been developed for rapid-static and navigation applications, where usually only two stations are involved, the maximum distance is some tens of kilometers, and time of occupation is of the order of seconds to minutes, or the receiver is moving.
Successful instantaneous ambiguity resolution (using only one epoch of dual frequency carrier phase and pseudorange data) has been reported for baselines of tens of kilometers. It is not always clear however, if these are isolated cases, or that the results can be reproduced at any place on the Earth at any moment during the day. In this respect, considerably more is known about the long baseline, long time span application, since the IGS processing centers operate on a daily basis.
It can not be stressed enough that conceptually there are no differences between the two applications, and that research directed to one application can benefit from research conducted for the other application.
This SSG was founded with the idea to make the field of ambiguity resolution more transparent. Citing form the terms of reference:
"Despite the large effort spent by many groups from all over the world in devising various schemes, knowledge about their theoretical foundation, and how the schemes are related to each other, is still lacking. Different terminology is used and comparisons between methods are rare. Due to a lack of knowledge about the various methods, the implementations used in the comparisons (if made at all) are not always complete, thereby making the test results unreliable. Moreover, results reported of one particular method, are often difficult to relate to the results of another method, due to lacking knowledge of the characteristics of the data and the type of computer that was used."
After four years, many aspects of the problems mentioned above are still present. Although progress has been made, and even more methods for ambiguity resolution and validation were proposed, detailed descriptions remain rare. To be able to reproduce results and compare methods, either the data set and the model for the float solution has to be completely specified, or a low-level (preferably some sort of source code) of the method used for the ambiguity resolution should be available. At this moment there is only one method (LAMBDA) for which the source code and a detailed description is freely available(http://www.geo.tudelft.nl/mgp/lambda/).
Validation methods are usually described in more detail and comparisons can be found in the literature. There is however still no general agreement on its theoretical foundations.
An new interesting approach has been the proposal for a new diagnostic quantity for assessing the probability of correct ambiguity resolution (ADOP, or Ambiguity Dilution of Precision). It is, as the existing DOPs, based solely on the model at hand and the expected satellite configuration (so it can be computed before any measurements are made, i.e. in the design stage).
We are currently entering a period with, due to the rise of the solar sunspot number to its next maximum (expected to take place somewhere between June 1999 and January 2001), an highly active ionosphere. It will be the first time we are experiencing this since the GPS became operational, and it will likely provide us with some interesting problems with respect to ambiguity resolution.
At the site
http://www.geo.tudelft.nl/mgp/people/paul/amb.html a bibliography on ambiguity resolution and validation can be found. It is intended to keep this bibliography up-to-date.
List of Members:
Hasanuddin Abidin (Indonesia)
Barbara Betti (Italy)
Simon Corbett (United Kingdom)
Mattia Crespi (Italy)
Hans-Juergen Euler (Switzerland)
Shaowei Han (Australia)
Paul de Jonge (The Netherlands/USA)
Hansjoerg Kutterer (Germany)
Herbert Landau (Germany)
Barbara Marana (Italy)
Manuel Martin-Neira (The Netherlands)
Daniela Morujao (Portugal)
Joao Galera Monico (Brasil)
Benjamin Remondi (USA)
Stephan Schaer (Switzerland)
Wolfgang Werner (Germany)
Pascal Willis (France)
Gerhard Wuebbena (Germany)
Ming Yang (Taiwan)
Zuofa Li (Canada)