ILRS recommended procedure for tuning IRVs

This note was compiled by Roger Wood & Graham Appleby (NERC Space Geodesy Facility, UK) on November 2002

Andrew Sinclair described the IRV force model and reference system in some detail. The principal attractions of the IRV (Inter-Range Vector) system for laser ranging predictions are that it uses an exceptionally compact transport format, and for SLR observing allows recovery of a very good approximation to a precise orbit using a simple integrator and simple gravity field model, but only if the IRVs are correctly "tuned". This supplementary note gives a step-by-step description of the tuning procedure and point out how observing stations should use tuned IRVs to get the best results.

The key feature of the IRV system is the IRV integrator used to recover a precise orbit from state vectors (satellite positions and velocities) for 0hr UT. The integrator uses a particular gravity field model GEM10N (high satellites, Lageos and "above", using degree and order 7, and lower satellites using degree and order 18) together with a simple model of lunar and solar perturbations. It explicitly does not take account of any other forces on the satellite. Thus stations using IRVs for observing do not have to know anything about gravity fields or other sources of perturbation: that has all been modelled as accurately as possible by the prediction centres; but to get the best predictions they must use the same standard integrator and named gravity field that was used in the tuning process.

Converting a precise orbit to tuned IRVs

1. For a given satellite form a precise orbit from positional data for, say, the previous 5 days, using a sophisticated model which might include, as appropriate, the latest, high-resolution gravity field, detailed modelling of atmospheric drag, solar radiation pressure, magnetic effects and other surfaces forces.
2. Extrapolate this precise orbit forward for 5 days (or more) ahead to produce a "precise" prediction of the future behaviour of the satellite.
3. Form the state vector for 0hr UT for the first day of this prediction and input it to the IRV integrator with GEM10N to propagate the orbit forward over the next 24 hours to produce an "IRV" prediction.
4. Tune the IRVs for this 24 hour period by comparing the "IRV" prediction, point by point, with the "precise" prediction. Examine the differences between them, make small changes to the initial state vector and repeat step 3 to make a fresh comparison.
5. Iterate steps 3 and 4 until the differences over the full 24 hours are minimised (in a "least squares" sense).
6. Adopt the final state vector as the tuned IRV for that day.
7. Repeat for each day of the prediction set.

Using tuned IRVs for observing predictions

It is clear from the above that tuned IRVs achieve optimal results only when processed with the standard integrator in combination with the GEM10N gravity field of the appropriate degree and order. Any other procedure will lead to (unpredictable) degradation of the prediction. In practice the non-gravitational forces on a satellite often differ from the best estimates modelled at prediction time. Their consequences may be effectively corrected using timebias functions (based on near real-time feedback from observations) and, for satellites in the lowest orbits, drag functions. Most satellites now have IRVs recomputed and tuned at least once a day. However, if across-track errors remain small, long duration IRV sets, together with up-to-date timebias information, can continue to be used with confidence, even when timebias values are large.

Responsible Government Official: Carey Noll
NASA's Privacy Policy and Important Notices

Send us your comments

Last modified date: Thursday, December 29, 2005
Author: Roger Wood
Maintained by: Carey Noll