[FOSS-GPS] rtklib: simple DGPS post-processing experiment, improvements with rtklib 2.4.0

[Timo Juhani Lindfors]

Hello Tomoji,

I upgraded my internal Debian package of rtklib to version 2.4.0 and
reran all the commands I mentioned in my earlier email. Here's a
summary of changes in output I got:

Timo Juhani Lindfors <timo.lindfors at iki.fi> writes:
> Single mode (-p 0)
> ==================
> http://iki.fi/lindi/gps/experiment/2/p0-neo1-scatter.png
> http://iki.fi/lindi/gps/experiment/2/p0-neo2-scatter.png

http://iki.fi/lindi/gps/experiment/2/p0-neo1-scatter-2.4.0.png
http://iki.fi/lindi/gps/experiment/2/p0-neo2-scatter-2.4.0.png

=> no changes

> DGPS mode (-p 1)
> ================
> http://iki.fi/lindi/gps/experiment/2/p1-neo1-scatter.png

http://iki.fi/lindi/gps/experiment/2/p1-neo1-scatter-2.4.0.png

=> no changes

> Kinematic mode (-p 2)
> =====================
> http://iki.fi/lindi/gps/experiment/2/p2-neo1-scatter.png

http://iki.fi/lindi/gps/experiment/2/p2-neo1-scatter-2.4.0.png

=> remarkable improvement! Scatter with radius of 0.00005 degrees is
now gone and most of the points are very tighly packed in the middle
with some slow random walk.

Here's the same scatter in UTM coordinate system so that the units are
meters in various zoom levels:

http://iki.fi/lindi/gps/experiment/2/p2-neo1-scatter-utm5-2.4.0.png
http://iki.fi/lindi/gps/experiment/2/p2-neo1-scatter-utm1-2.4.0.png
http://iki.fi/lindi/gps/experiment/2/p2-neo1-scatter-utm0.30-2.4.0.png
http://iki.fi/lindi/gps/experiment/2/p2-neo1-scatter-utm0.10-2.4.0.png
http://iki.fi/lindi/gps/experiment/2/p2-neo1-scatter-utm0.03-2.4.0.png

It really seems that a huge number of points stay inside a radius of
just 2.5 centimeters. Can this really be true?