Re: Modify MASK - Ublox GPS

[Olivier ADLER]

Yes i think that it is too low.

8-10 degrees would be better or even perhaps 15 degrees.

It is not possible to set it from the code. You'll need to connect your GPS to a FTDI cable and program it with the Windows software U-center from Ublox.

http://code.google.com/p/ardupilot/wiki/ublox

You'll need to open the configuration view (CTRL - F9) inside U-center to get access to all GPS parameters.

Olivier



Olivier.




On Tuesday, June 4, 2013 7:59:47 PM UTC+2, Jesus Alvarez wrote:

Hi all,

I have been reading the group for a long time but never had anything to say.. until now!

I've seen or read that gps mask for arducopter is set to 5º. It seems to low to me (I work with DGPS for trajectography).

Is there a place where I could increase it a bit? Is it in any .h? I have been looking for it but no luck.

thanks!

[Randy Mackay]

     I'm not very familiar with the setting you're talking about...what's the advantage of increasing it?

-Randy

---

From: Olivier ADLER <cont...@nerim.net>
To: drones-...@googlegroups.com
Sent: Wednesday, June 5, 2013 6:16 AM
Subject: [drones-discuss] Re: Modify MASK - Ublox GPS

--
You received this message because you are subscribed to the Google Groups "drones-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to drones-discus...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[Gareth Owen]

Filters out sats within x degrees of the horizon. Sats close to the horizon provide noisy data. Typically sats less than 10-15 degrees are filtered.

[Olivier ADLER]

The elevation mask is useful to remove weak low horizon satellites from the position solution.


If you rise elevation angle mask, you'll get eventually less visible satellites, but the remaining ones will be strong signal ones, with better signal precision (less atmospheric attenuations and less multipathing).


Generally, consumers GPS for general use have a default setting of 5 degrees, to rise the 3D fix availability during difficult urban canyon reception conditions, for car or pedestrian navigation inside cities....


Survey or timing GPS do have a higher default settings at about 8 - 10 degrees or sometimes more to enhance precision.


For our use, because we are starting on the ground and we store immediately our ground position at take inside the home position memory, i think it's better to remove the low horizon satellites from the solution, because multipathing on those satellites is very strong on those satellites at the ground level, specially when there are buildings around or mountains.

I did rise my LEA6-H setting at 10 degrees.

I did switch it to pedestrian mode as well to filter out large acceleration jumps. Pedestrian mode is filtered with a max 30 m/s speed and modest acceleration. This should filter out large jump errors we can see actually in the GPS output ( i had one with 260 m/s recently ).


I did not tried this in the field yet, only lab tests.


Olivier.

[Randy Mackay]

     I guess nobody's done any testing with the different configurations to prove/disprove that it will help?  If we had that it would be very easy to make the decision.

-Randy

---

From: Olivier ADLER <cont...@nerim.net>
To: drones-...@googlegroups.com

Cc: Randy Mackay <rmac...@yahoo.com>
Sent: Wednesday, June 5, 2013 6:43 PM
Subject: Re: [drones-discuss] Re: Modify MASK - Ublox GPS

[Chris Gough]

I'm not saying the current setting is a good default, but keep in mind that the best GPS mask might be different in different situations. In the southern hemisphere we generally see less satellites than you guys north of the equator, and the highest "mountain" on this continent is ~2000m.

Perhaps this should be a tunable parameter?

Chris Gough

[john...@gmail.com]

Seeing a satellite dosn't really help if the positional  data is garbage..

[Chris Gough]

But low satellite aren't always garbage, It depends on the landscape (and weather). A good one helps with GPS altitude. The ability to tweak it might sometimes be useful. We already have that ability, so it's not a bug deal.

Chris Gough

[Andrew Tridgell]

Hi Chris,

> Perhaps this should be a tunable parameter?

yes, I think it should be. What the optimal elevation mask would be is
not obvious to me. Some factors:

- if we are flying at 100m, the horizon may be a fair bit further away
than usual, and noise from low sats may be less

- when a plane is rolled at 70 degrees in a turn, the patch antenna
will have higher gain towards the horizon. Could we lose some of the
higher elevation satellites?

I think we want a GPS_ELEV_MASK parameter. Unfortunately the GPS object
is dynamically allocated, and a quirk of our current parameter system is
we can't have parameters on dynamically allocated objects. I've been
thinking of re-doing the GPS driver to fix this, basically getting rid
of the separately compilable GPS drivers, and just always compile them
all in. That would save a lot of ram on APM2, plus would also mean we
could have parameters in the GPS driver. I'd also like parameters for
this like SBAS, and the dynamic mode.

Cheers, Tridge

[Olivier ADLER]

Yes having tunable parameters here would be nice.

Default SBAS settings are not optimal because the correlators are scanning for all available SBAS satellites (WAAS + EGNOS) using 3 channels.

It is better to restrict that to only the locally available SBAS satellites, and perhaps use only 1 or 2 correlator channels. This reduce the DGPS fix time.


The GPS dynamics mode is another interesting parameter to tune. I did change mine to Pedestrian, to filter out large position jumps. First couple flights with that mode seems to give good results.


I did rise elevation mask to 10 degrees, and it seems that it gives better precision when landing or loitering at a couple meters altitude in my backyard with buildings around.


Olivier.

[Mark Colwell]

I have set my SBAS for 2 channel search and specify 135, 138 sats. as Other option and use 10° for minimum detection works in Colorado USA 

--