******************************************************************************
BSW Electronic Mail 20-Mar-2007 12:55:01 CET Message Number 0242
******************************************************************************
AUTHOR: BSW team
Subject: Re: No 241: Question regarding correlation strategy for the final
Dear Miltiades Chatzinikos,
Dear users of the Bernese GPS Software,
if you have 8 GNSS stations you can form 7 independent baselines. Each
additional baseline cannot add more information because it is derived from
the same original observations.
If you introduce more than these 7 independent baselines in a network
processing the normal equation becomes singular because of the dependent
observations. This is the case if you have considered the correlations
correctly and it is indicated by the warning messages (MATRIX
REGULARIZED).
If you introduce a set of dependent baselines and you consider only the
correlations within a baseline the system does not recognize this
dependency and it will provide a solution without singluarities. That's
why you did not get the "regualrization" messages in your previous GPSEST
runs.
Of course, a network solution should be (nearly) independent from the set
of selected baselines. To reach this status introduce only independent
baselines to GPSEST and consider the correlations correctly. Adding
redundant (depending) baselines make only sense for global networks (see
also on-line help for "FORMING OF REDUNDANT BASELINES" in SNGDIF, panel
3.1).
(This situation is comparable with the classical measurement of directions
in a triangluation network: if you have measured 8 directions you are
allowed to derive 7 angles for the network analysis from these
observations. Otherwise the normal equation becomes signular because of
introduce dependent observations.)
I hope this helps.
Best regards
The BSW team
On Mon, 19 Mar 2007 bswmail@aiub.unibe.ch wrote:
> ******************************************************************************
> BSW Electronic Mail 19-Mar-2007 19:55:02 CET Message Number 0241
> ******************************************************************************
>
> Author: Miltiades Chatzinikos
> Subject: Question regarding correlation strategy for the final solution
>
>
>
> Dear Bernese users,
>
> I am a PHD student at the Aristotetle University of Thessaloniki at =20
> the faculty of Surveying Engineering and I use the Bernese GPS =20
> Software v.5.0 for research reasons.
>
> I am trying to resolve a network that consists of 8 permanent stations =20
> and when I reach the point where I want to have the final solution, I =20
> choose all (trivial +independent) baselines and the GPSEST program. =20
> What happens is that when I reach the 3.1. panel, at the "correlation =20
> strategy" step, if I choose the "correct" guidepost, the software =20
> provides me with some results, but at the same time an error message, =20
> as follows:
>
> *** SR SYMIN8: MATRIX REGULARIZED IN STEP NO: 38 PIVOT: 44
> *** SR SYMIN8: MATRIX REGULARIZED IN STEP NO: 39 PIVOT: 10
> *** SR SYMIN8: MATRIX REGULARIZED IN STEP NO: 40 PIVOT: 16
> *** SR SYMIN8: MATRIX REGULARIZED IN STEP NO: 41 PIVOT: 2
> *** SR SYMIN8: MATRIX REGULARIZED IN STEP NO: 42 PIVOT: 3
>
> (What you see, is just a part of the message)
>
> Could you inform me as what this message is about, and whether the =20
> results I get are correct?
>
> Thank you for your time
> Looking forward to hearing from you
>
> Yours sincerely,
> M. Chatzinikos
>
>
>
>
>
>
> [Mailed from: Miltiadis Chatzinikos ]
>
-----------------------------------------------------------------------------
Rolf Dach Astronomical Institute, University of Bern
dach@aiub.unibe.ch Sidlerstrasse 5, CH-3012 Bern, Switzerland
http://www.aiub.unibe.ch/ Tel: +41 31 6318593 Fax: +41 31 6313869
http://ftp.unibe.ch/aiub
-----------------------------------------------------------------------------
[Mailed from: Rolf Dach ]