Question

Hi All,

I'm trying to use GPSTk to solve solutions. I'm using SP3 orbits and the pseudoranges taken from a Sirf Star III configured to work in Sirf Binary mode that gives pseudoranges.

I have used SolverLMS^? and SolverWMS^?. When I run the solvers with data from a base station, obtaining the pseudoranges from the RTCM frames, I obtain a very good solution. Standard deviation under two meters.

However, if I use the pseudoranges from the chip SiRF^? Star III, I obtain very bad positions, with a std deviation of hundreds of kilometers. I have checked several times that I'm taking the pseudoranges in the good way.

Maybe should I filter in some way the data before processing? Maybe I'm not using the correct solver?

By the way, I wanted to make differential corrections, but I really haven't been able to find if I have to substract the corrections (from the RTCM file) to my raw pseudoranges or instead, add them.

Thanks for your comments and forgive my english...

Jose Luis Martín joseluismartin (en) or (at) gmail. com

-- JoseLuisM - 26 Feb 2009

Answer

If you answer a question - or have a question you asked answered by someone - please remember to edit the page and set the status to answered. The status is in a drop-down list below the edit box.

Hi Jose Luis!

Sorry, but I have to answer your question with another question:

When you process your SiRF^? data, what kind of ephemeris are you using?: I guess that your data has around 1 sample per second rate. In that case you should be using broadcast ephemeris, because satellite clock information in SP3 files comes at 1 sample per 900 seconds rate. Otherwise, you should insert a properly tuned 'Decimate' object into your data process chain, and expect just one correct fix each 15 minutes.

Another option is to use SP3 orbits, and mix them with broadcast satellite clocks (writing you own class that achieves that). Either way, you must always use appropriate satellite clock information.

Regarding your RTCM-related question, if I remember well what is sent in an a RTCM message is the pseudorange correction (PRC), which should be substracted from your pseudorange observation.

However, be aware that synchronization issues are very, very important here, and that PRC values are satellite-dependent.

Regards,

Dago

-- DagobertoSalazar - 04 Mar 2009

From my ignorance, I thought that the algorithm interpolates the satellites positions from the SP3 corrections. I'm understanding from you that if I use SP3 files the algorithm is going to know the satellites positions only once for each 15 minutes. So how does it perform the solutions in the epochs in between?

I have used PRSolver with sp3 files and the solutions looks very good.

Could I write the questions in Spanish? :/ It hard for me to explain in english, xDD

Thank you very much for your help Dago

-- JoseLuisM - 10 Mar 2009

Hi Jose Luis:

You are RIGHT regarding SP3 satellite POSITIONS: They are interpolated and the results of these interpolations are very good.

However, you are WRONG regarding SP3 satellite CLOCKS: If you interpolate them, your results will be bad, because clocks tend to drift in unexpected ways.

This is the reason because, in order to obtain the best results, you must either decimate SP3 data to 900 s, or look for a good substitute for satellite clocks (for instance, 30 s precise satellite clock information from CODE processing center).

Coming back to your specific setup, if you are using a simple, consumer-grade GPS receiver, then your pseudorange observations are probably single frequency, and very noisy. That is the reason you get good results: The noise from your observations masks SP3 satellite clock interpolation errors.

If the former is the only setup you are interested in, then go ahead. I must remind you, however, that the gains achieved using SP3 orbits data instead of broadcast ephemeris will then be rather modest.

Finally, I believe it is better to keep this discussion in English, in order to reach a wider audience. However, you can write me directly in Spanish to dagoberto Punto jose Punto salazar Arroba upc punto edu .

Best regards,

Dago

-- DagobertoSalazar - 13 Mar 2009