Prediction Format Study Group
Meeting at 13th International Laser Ranging Workshop (10/7/2002)
Randy Ricklefs, Chairman
Minutes by Jan McGarry

The new Prediction Format was presented by Randy Ricklefs (U.Texas/MLRS). The new format is a tabular (ASCII) format, short enough to be e-mailed, and contains all of the information now contained in IRVs. The format is made up of various subsections (chunks) that can be used or left out as needed to minimize the size. The data is now designed to be interpolated (not integrated) to increase the accuracy. The expected prediction accuracy (determined by the data center) will be included in the data. The data entries have variable spacing to allow the analysis centers to optimize the data for a highly elliptical orbit.

Werner Gurtner (U.Bern/Zimmerwald) presented his work on the method of interpolation that would best be used for the new Prediction Format. Werner looked at both Lagrange and Hermite interpolation. Both can handle the variable data spacing and are easy to program. Lagrange uses "n" points to generate a polynomial of degree "n-1". Hermite interpolation is similar to Lagrange but requires the first derivatives of the data. The quality of the Hermite polynomial interpolation depends heavily on the quality of the first derivatives. Using the first derivatives and Hermite interpolation does not substantially increase the allowable interval spacing. While Lagrange interpolation does not explicitly give the polynomial coefficients, and is not optimized for speed, it may be the best choice for the interpolator since it can potentially provide the velocities (implying no need to explicitly include the velocity in the format).

The following spacing (depending on the polynomial degree) provides for interpolation of the data to 20 cm accuracy using Lagrange interpolation. The spacing is given in minutes.

It was asked if the new Prediction data could be used for the Normal Point generation. Since normal points are determined from a low order polynomial fit to the residuals from the measurement minus prediction data, this implies that the Lagrange polynomial used for interpolation must be smooth in order to work properly. Werner agreed to look into this and to also look at how accurately the Lagrange interpolator gives the velocities.

Dave Rowlands (NASA/GSFC) presented his work on transponders (using GEODYN) in support of the new Prediction Format. Dave has been working with GEODYN for over 20 years. In 1988 GEODYN was given interplanetary capability and has been used for many missions since then. To support this new format, Dave has been testing the accuracy of generating pointing angle, range and range-rate using range vectors (transmit and receive) provided in the prediction format's geocentric frame, updated with corrections provided in the format. A comparison of this calculation against the calculations done rigorously provides the accuracy. He has been using MGS data for this testing.

His tests have highlighted some additional fields required for the Prediction format, but in general, the current Prediction format looks like it can generate transponder predictions to the accuracy required. Still being worked on is the generation of range-rate from range data only (so that range-rate does not need to be included in the format). Dave's tests with MGS data indicate that the 15 cm / sec accuracy required for transponder work can be achieved using two range vectors spaced 0.05 seconds apart. It is possible to generate these range vectors using Hermite polynomial interpolation of vectors spaced 20 seconds apart in the prediction format.

Following the presentations there was a general discussion on how to incorporate tracking restriction information into the format. There are two future missions that may require tracking restrictions: GLAS and ADEOS II. The GLAS instrument can be harmed if the SLR laser beam is placed within ~ 1 degree of the detector. Since the instrument is usually nadir pointing, the chance of this happening is small since this is in the keyhole of most tracking mounts. There is currently no plan to ask for tracking restrictions from the GLAS instrument, however, should future missions have this same situation, we would want to be able to include a maximum elevation restriction in the Prediction Format. ADEOS II has tracking time restrictions - there are certain periods when stations can track and when they can't. We hope to put this type of information somehow in the header.

Randy will distribute the new format to the full ILRS community sometime before the end of the year and send a special e-mail to the prediction centers to get their agreement on this format. Randy will also be working on developing sample code (FORTRAN) for the stations. This will doubtless take several months but must be available before any field tests can be done. Randy is looking for a volunteer to convert his FORTRAN sample code to "C".

The following people have kindly volunteered to generate prediction data in the new format for field testing: Graham Appleby (SLR), Randy Ricklefs (Lunar), and Dave Rowlands (Transponder). The following stations will perform field testing of the new prediction format: Herstmonceaux (SLR) and MLRS (Lunar). There is a possible transponder experiment with the Mercury Laser Altimeter (MLA) onboard after launch in 2004. The tracking station will probably be SLR2000.

[an error occurred while processing this directive]

Last modified date: Thursday, September 27, 2012
[an error occurred while processing this directive]
[an error occurred while processing this directive]