| Date: | 1997-05-20 18:00:00 | |
| Sender: | A.T. Sinclair, Royal Greenwich Observatory <slr@gxvf.rgo.ac.uk (SLR at Herstmonceux)> | |
| Subject: | [SLR-Mail] No. 62: The use of RGO multiple IRVs for GPS (and other) satellites | |
| Author: | A.T. Sinclair, Royal Greenwich Observatory | |
| Content: | ******************************************************************************** SLR Electronic Mail 1997-05-20 18:00:00 UTC Message No. 62 ******************************************************************************** Author: A.T. Sinclair, Royal Greenwich Observatory Subject: The use of RGO multiple IRVs for GPS (and other) satellites The use of RGO Multiple IRVs for GPS (and other) satellites This note describes a method of significantly improving predictions from IRVs by tuning over spans shorter than 24 hours, and providing four sets of IRVs per day. The IRV force model is fairly simple, and omits drag and solar radiation pressure forces. The result is that for very low or very high satellites the IRV orbit cannot be fitted to a precise orbit over a 1-day span to high accuracy. The residuals of mis-fit can be up to about 100 metres. A possible solution, as adopted by GFZ for GFZ-1, is to represent these mis-fit residuals by an additional, daily time-bias correction, the drag function. This works satisfactorily for low satellites, for which the mis-fit error is nearly all along track, but it does not work for high satellites (e.g., GPS), for which there is a large radial component of the error. Another solution is to fit the IRVs over a shorter span than 1 day, and to generate several sets of IRVs to cover the whole day. This works well and we have found that four sets per day each covering 6 hours, give worthwhile improvements for both high and low satellites. For the GPS satellites we use the daily predicted orbits computed by the Center of Orbit Determination (CODE) at the Astronomical Institute of the University of Bern. These predictions are computed sometime after 12:30 UT of each day, and are for the current day and the following two days. Within an hour these are converted into IRV form by RGO, and placed in the FTP directory at RGO in pub/slrirv/current, with file names such as spa_9705.14 and spb_9705.14 for GPS35 and GPS36 respectively. The files are also placed in the EDC, in directory pub/laser/predictions/RGO_PRED, with file names such as g35_9705.14 and g36_9705.14. The files can be sent by e-mail daily from EDC for those who request it from EDC (e-mail address: edc@dgfi.badw-muenchen.de). The CODE predicted orbits are very accurate, to better than 1 meter in range. The IRV model over 6 hour spans degrades this slightly, but range errors should not exceed 10 metres. A typical set of these multiple IRVs, for GPS35, is: COD09053 G35 4 1997 5 14 0 0 0.0 -7686149.213148 -20908896.593516 -14450389.494331 3535 053 1 1535.759068366 1047.843637196 -2326.264789429 0 0 0 5605.0 -43045435.300996 257.337916132 1997 5 14 0 0 0.0 -7686213.978476 -20908880.712369 -14450312.613319 3535 053 1 1535.759284666 1047.838512497 -2326.274580282 0 0 0 5605.0 -43045407.304163 257.323216881 1997 5 14 0 0 0.0 -7686081.782310 -20908857.160139 -14450415.065586 3535 053 1 1535.769555952 1047.848261001 -2326.261723572 0 0 0 5605.0 -43045354.008035 257.356093382 1997 5 14 0 0 0.0 -7686275.511105 -20908861.559965 -14450240.472992 3535 053 1 1535.761387774 1047.833765655 -2326.284527933 0 0 0 5605.0 -43045377.544062 257.310625496 The identifier COD09053 is the identifer used for the CODE orbit, and refers to GPS week 0905, and day of the week 3, where Sunday is day 0. The number 4 in column 25 of the header line implies that there are four sets of IRVs following. For the GPS satellites these sets of IRVs have been tuned over 8-hour spans with a 2-hour overlap, so the four sets of IRVs are valid for the spans: 0-8 hours, 6-14 hours, 12-20 hours, and 18-26 hours. Note that the epoch of each set is at 0 hours, and so each set has to be integrated from 0 hours up to the time of the pass. So all the user has to do is to select the set for the quarter of the day that the pass starts in, and use it in exactly the same way as a normal set of IRVs. The two-hour overlap means that there is no problem if a pass starts near the end of a quarter, since it can be continued for 2 hours into the next quarter without the errors becoming large. These 4x IRVs are an improvement over the 1-day IRVs even for satellites at around 1000 km height, as they give smoother time bias values. So it is likely that all the IRVs generated by RGO will change to using the 4x system. It is suggested that stations wishing to make use of 4x IRVs should generalise their IRV handling to deal with *any* number of IRV sets per day, according to the indication in columns 24 and 25 of the header line. (It was agreed at the CSTG Laser Ranging Subcommission meeting in Shanghai that only the first 22 characters of the header line should be used for labelling purposes, and that characters 24,25 should contain an integer to indicate whether multiple IRV sets follow, with a value of 1 or 0 indicating just a single set.) From: slr@gxvf.rgo.ac.uk (SLR at Herstmonceux) ******************************************************************************** |