[FOSS-GPS] High accuracy positioning with low cost GPS devices: a FOSS project

Jaykob jraible at gmx.de
Wed Dec 17 08:32:49 EST 2008


Hi Eugenio and all the others!

I'm currently writing my diploma thesis about pretty much the same stuff.
My aim is to develop a realtime GPS solution for robotic outdoor
environments with consumer grade single frequency receivers.
I have two AEK-4T (LEA-4T) U-Blox receivers as well.
I'm pretty much at the beginning and I'm still in the phase of reading a
lot. I came about this mailing list and your goGPS site as well as
coyotebush's Kinematic site (precision-gps.org).
Because I don't want to invent the wheel again as it would be probably too
much for my thesis anyway I want to ask you guys for some hints.

I'm working with two netbooks (eeePCs), one as base station and the other
one as rover. Communication will be over bluetooth or WLAN so pretty short
baselines (not much more than 100m) have to be considered.
I'm aiming for an accuracy of about 50cm. Is that possible with the LEA-4Ts
and the standard ANN-MS antennas? What techniques / algorithms would you
suppose? Can anybody tell me how accurate Kinematic (precision-gps.org)
would be taking the recent Snapshot #8? Does anybody have an advice how one
could improve this program or where the problems are?

I also had a look for the GPSTk and wrote a RINEX generator for the LEA-4Ts
for it. I took the existing rtAshtech program and modified it for the U-Blox
receivers. It's still under development but it's already generating the
observation and navigation files. I also tried out the DDBase program (also
in the GPSTk) for a static situation with a very short baseline (20cm). When
I entered a pretty good position estimate for the rover the result was
really good but taking the pseudorange rover estimate didn't give good
solutions for relatively short observations. It's not what I need because
it's a post-processing program but does anybody know if I could take parts
out of it and make it realtime (or near realtime)? I also read about kvecsol
in a paper which is vecsol (also from the GPSTk) but modified for rtk. But
vecsol only seems to work with dual frequency data...

In short: I'm very thankful for any ideas and hints that could help me for
that purpose.

Eugenio: Do you publish your thesis and software when you're done?

Thanks in advance!
Jakob

-- 
View this message in context: http://n2.nabble.com/High-accuracy-positioning-with-low-cost-GPS-devices%3A-a-FOSS-project-tp1312818p1667877.html
Sent from the Open Source GPS-related discussion and support mailing list archive at Nabble.com.



More information about the FOSS-GPS mailing list