I'm trying to obtain raw double difference carrier phase residuals from GPSTk using DDBase output.

I'm having problems with residuals detailed within the RDD output file. They appear to increase over time, suggesting another error.

Graphs show non-linear relationships between epochs, differing to those formed with the same data within Leica Geo Office. The GPSTk raw DD's wander off considerably, some 10m from zero. Many journal articles and graphs of raw double differences show a simple linear pattern over epochs, varying around a mean of zero.

I understand Leica Geo Office forms double differences following the least squares adjustment phase, resulting in post-fit DD's. However the output from DDBase appears to exhibit some other error and thus increase and decrease largely (1-10cm over 2-3 hours). This increase can be seen by looking down the list of RDD resL1_m.

Basic double differencing using a fixed master (e.g. G22) within Excel, using the GPSTk RAW output data, has shown differences exists between these calculated DD residuals (DDL1 - DDER) and those given in the RDD output.

Looking at the source code, the Excel process follows same used by the GPSTk algorithm.

Any ideas?

I am happy to provide graphs of what I am trying to explain smile

Thank you for your help,


-- MichaelRigby - 24 Oct 2007


ALERT! 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.

Michael, the change in double differences over time will be determined by the apriori receiver positions that you give it. If the receiver position is off by, say x, then the DDs will tend to wander with changes in satellite geometry with an amplitude that is approximately x. This change is exactly what DDbase uses to compute the position; if you look at post-fit residuals they should be flat - only noise and multipath. Brian

-- BrianTolman - 24 Oct 2007

Thanks very much Brian smile

Makes perfect sense!

I will try recomputing better estimates site coordinates today and recheck the raw double differences later.

Thanks again

-- MichaelRigby - 25 Oct 2007

Thank you Brian,

Here in Australia our datum (GDA94) is static and consequently all coordinates have been moving with the tectonic plate, whilst WGS84 and IGS are always being updated.

Therefore it's necessary to convert GDA94 to ITRF 2000 (IGS, with +7 year adjustments for 2007)in order to be more accurate.

Thank you very much smile

-- MichaelRigby - 13 Nov 2007

-- MichaelRigby - 13 Nov 2007 No such template def TMPL:DEF{PROMPT:supportquery}

Topic revision: r5 - 27 Nov 2007, RickMach

This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Foswiki? Send feedback