Re: GPSD

[Gary E. Miller]

Yo Anton!

On Mon, 5 Sep 2022 15:17:56 +0200
Anton Strydom <agstrydom@gmail.com> wrote:

> Firstly the "code" is what I found online and not mine it appeared
> easy to adapt but proved not to

It looked familiar, and was the motivation for :
    https://gpsd.io/gpsd-client-example-code.html

> I have very little knowledge of Linux and even less when it comes to
> Python

Pretty much any language can be used with gpsd.  C is most common.

> I have had lots of success navigating the vehicle to within 100mm
> from the targets that I have to my disposal

The trick is getting it right 99.99% of the time.  GNSS is accurate
50% of the time, but 5% of the time it can be way off.  That is wy
you need to run gpsprof for 12 hours and see your worst case.

> Below is a comparison between a marker GPS2 and my device being static
> 
> X Y LABEL LAYER ATTR_1 TIMESTAMP
> 
> -82266.7 -2895988 GPS2 Unknown Point Feature GPS2
> 
> Difference
> -82266.7 -2895988 1 Waypoint 02/17/22 10:42:26 0.024698
> -82266.7 -2895988 2 Waypoint 02/17/22 11:32:12 0.020616
> Average

Sure, it was good, once.  But when human safety is involved you
need to get better than 99.99%.  Sunspots can do a real number
on GNSS accuracy.


> No I won't be using gpsprof, I have had a look at it and it is great
> but not for what I am doing

It is always hard for us to convince people to run it for 12 hours.
But all are surprised by the results, and no one said it was a waste
of time.

> I am using a wide array of gps receivers, all of them precision
> timing and raw data units.
> 
> UBlox LEA 4T single frequency
> UBlox LEA 6T single frequency
> UBlox NEO 8 P single frequency
> UBlox F9P dual frequency
> Hemisphere dual frequency
> Novatel dual frequency
> Leica SR500 dual frequency
> Trimble R2
> To name a few

Some samples from the Hemispehere and Leica would be good for the gpsd
regressions.

> Apart from accurate navigation I am also busy designing systems that
> monitor structures to predict possible failures

Cool.

> I can see the "Status" value in the GGA string on all the units I use

Yes, but sadly, no two models set "status" the same way.

> The "Datum" would be nice to have as the final product will have to be
> programmed with for instance  Schwarzeck datum used in Namibia

If you eeceiver send it, then gpsd reports it.  Sadly the datums in
receivers are rarely accurate.

> WGS84 would be the best if everybody used it, especially the mines
> etc.

Which WGS84?  So many to choose from.  And then, at least in the USA,
you need to convert to NAD84 (or similar) as used by USGS charts.

> The units used are all configured in a very special way, the accuracy
> achieved is sub decimeter without any outside influence such as NTRIP
> using a Moving Baseline configuration.

I believe it when you show the gpsprof scatterplot that shows that.

> The "Track" value returned is accurate to less than a third of a
> degree compared to a total station measurement.

With the more than occasioanl perverse case.

> ECEF is for the survey industry and would be nice to have later on as
> they all work using Northing and Easting. It also brings it own set
> of issues but that is
> for the future.

Several misunderstandings there.

All GNSS compute in ECEF space, not just for "survey".  So not nice to
have, essential to the internals.  Once you have ECEF you can convert to
any datum you desire.

Like everything else, Northing and Easting are computed from ECEF,
but are not ECEF.  You can think of them as yet another way to denote
distances.

> I have version 3.17 installed

That is 5 years old!  Many bugs and a few CVE fixed since then.

> The following Python versions are installed
> Python 2.7.16
> Python 3.7.3

Any from 2.7 on up works with gpsd, for now.

> If you are referring to Example2.py, I have studied it yes and am
> busy with it trying to make sense of it all thank you

Feel free to ask questions.  The example can be improved from your
questions.
 
> The threading and everything is not mine but I hear you and it is all
> making sense.

Good.

> I have asked the question on other sites but have not had any
> worthwhile reply, yours is the first making any sense at all

A lot of bad advice out there.  So hard to know which is correct.
 
> Any advice will be appreciated as I am looking at future expansion
> moving beyond the Raspberry and L298HBridge too a more commercial
> type vehicle.

A lot of very expensive stuff has a RasPi hidden inside...

RGDS
GARY
---------------------------------------------------------------------------
Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
        gem@rellim.com  Tel:+1 541 382 8588

            Veritas liberabit vos. -- Quid est veritas?
    "If you can't measure it, you can't improve it." - Lord Kelvin

pgpXFvbCi3ImM.pgp
Description: OpenPGP digital signature