| Date: | 1996-12-16 10:00:00 | |
| Sender: | Andrew Sinclair <[Mailed Andy Sinclair <ats@ast.cam.ac.uk>]> | |
| Subject: | [SLR-Mail] No. 36: IRVs for GPS satellites | |
| Author: | Andrew Sinclair | |
| Content: | ******************************************************************************** SLR Electronic Mail 1996-12-16 10:00:00 UTC Message No. 36 ******************************************************************************** Author: Andrew Sinclair Subject: IRVs for GPS satellites IRVs for the GPS satellites, GPS35 and GPS36 A.T. Sinclair 1996 December 16 At the Shanghai Workshop I described a new set of IRVs for the GPS satellites that are now being formed each day by RGO, and placed in the RGO ftp directory. Outline details are repeated in this note. The IRVs are based on the daily predicted orbits for the GPS satellites, produced by the Centre of Orbit Determination (CODE) at the Astronomical Institute of the University of Bern. The CODE predictions, followed by the RGO IRVs, are produced sometime after 16.00 UT each day, and are for the current day and for the following two days. Some comments follow on the CODE predictions and the IRVs : 1. The CODE predictions are excellent, probably better than 1 metre. For the best possible predictions a station should use the rectangular coordinates produced by CODE directly, provided it understands the reference frame and time scale used. (But there are serious dangers in having predictions that are too perfect, of generating noise points parallel to the range gate, which will be impossible to separate from real data.) 2. Representing these predictions in IRV form will degrade the accuracy, due to deficiencies of the IRV model. If no action is taken to remedy this then there is no advantage to be gained by basing the IRVs on the CODE predictions. The usual IRVs plus daily time bias determined from the broadcast elements are just as good. 3. After much consideration I have concluded that best way of accomodating the force model deficiency in order to give simplest implementation for the station is to form multiple sets of IRVs over a day, each valid for just a fraction of a day. For the GPS satellites it is sufficient to give four sets of IRVs for each day, each valid for a quarter of the day. These are presented as a single header line, followed by 4 sets of 3 lines of data. The header line contains the number ´4´ in column 25, to indicate that 4 sets of IRVs follow. All the station has to do is select the set for the fraction of the day that their pass starts in, and use it in exactly the normal manner. It can simply edit out the 9 lines not required. I have actually tuned the GPS IRVs over 8-hour spans with a 2-hour overlap, so they are valid for 0-8, 6-14, 12-20, 18-26 hours. Thus if a pass continues for up to 2 hours outside the quarter that it starts in then there is no problem. 4. Use of these IRVs will not give any significant improvement in pointing accuracy - that is already sub arcsecond. They will just permit a reduced range gate, from about +/- 350 ns to +/- 100 ns. This would be an advantage for a station struggling with a low signal to noise ration, but if a station does not have this problem then there is no point in using these IRVs. [Mailed From: Andy Sinclair ******************************************************************************** |