AgOpenGPS - Page 51 - The Combine Forum
 2000Likes
 
LinkBack Thread Tools
post #501 of 3741 (permalink) Old 03-18-2017, 11:56 PM Thread Starter
Senior Member
 
Join Date: Aug 2012
Location: Vermilion Alberta Canada
Posts: 5,426
Mentioned: 5 Post(s)
Quoted: 2372 Post(s)
Only problem is you need special connector for access to I2C lines and it would be about 130$ + tax, duty, gst... landed in canada. So for about 80$, cut and paste a few lines of code is pretty easy to use the exact same chip.

Without calibrating, its only off a degree or two off level, easily zeroed out.

Was playing around with magnetometer/heading and had the IMU sitting on a cloth/wood stool by a swivel chair i was sitting in. IMU sitting still, all calibrated, shifted a few degrees heading just by swiveling the chair a little bit. As far as using it to augment GPS heading, i sure don't see it. Basically its an inclinometer, but a good one.

BrianTee is offline  
Sponsored Links
Advertisement
 
post #502 of 3741 (permalink) Old 03-19-2017, 11:17 AM
Senior Member
 
Join Date: Sep 2009
Location: S. AB
Posts: 3,368
Mentioned: 5 Post(s)
Quoted: 828 Post(s)
I would imagine that in UAVs, the 9dof sensors can be used to improve and smooth the heading. The magnetometer and the gyro can help filter out changes in GPS heading a little bit. Also they can be used to predict the heading before the GPS heading is even calculated, such as when standing still before the GPS resolves. And also it detect detect reverse, which could be very useful (the one thing still missing from AgOpenGPS as far as coverage mapping is concerned).

Could be wrong.

Obviously the inclinometer sensors are most important to correct GPS position for terrain.

torriem is offline  
post #503 of 3741 (permalink) Old 03-19-2017, 11:35 AM Thread Starter
Senior Member
 
Join Date: Aug 2012
Location: Vermilion Alberta Canada
Posts: 5,426
Mentioned: 5 Post(s)
Quoted: 2372 Post(s)
Quote:
Originally Posted by torriem View Post
I would imagine that in UAVs, the 9dof sensors can be used to improve and smooth the heading. The magnetometer and the gyro can help filter out changes in GPS heading a little bit. Also they can be used to predict the heading before the GPS heading is even calculated, such as when standing still before the GPS resolves. And also it detect detect reverse, which could be very useful (the one thing still missing from AgOpenGPS as far as coverage mapping is concerned).

Could be wrong.

Obviously the inclinometer sensors are most important to correct GPS position for terrain.
The good thing about UAV's is they are so forgiving, they wobble around and don't need to be accurate. In field tractors, half a degree and the implement overlaps itself and looks horrible.

Oh ya, the reverse thing. So if magnetometer heading and gps heading are out by more then pi/2, tractor is going backwards? Fair assumption?

So at that point, turn sections off and stop updating fixes until they are back in line ie < pi/2 delta. Only problem, if magnetometer goes whack, everything goes whack.

It is easy to get heading, its the x in the xyz of euler from the IMU.

Torriem or Kyler, do you guys actually know how mathematically the magnetometer value could be used to smooth the GPS readings? If the GPS is relatively stable, how can "correcting" it with another variable that knowingly bounces around going to improve the GPS heading? My lacking math ability says no matter what, it would only make it worse.
BrianTee is offline  
Sponsored Links
Advertisement
 
post #504 of 3741 (permalink) Old 03-19-2017, 07:36 PM
Senior Member
 
Join Date: Sep 2009
Location: S. AB
Posts: 3,368
Mentioned: 5 Post(s)
Quoted: 828 Post(s)
Yeah, good points. Simplistically, yes, I'd say that if the GPS heading changes by some great amount, and the magnetometer heading is within some small amount of the previous heading, then we could say the tractor is reversing.

But actually gyro readings could probably do this better, even with drift and no calibration. If the heading changed by a significantly different amount than the gyro reading did, that's a change tractor direction. I don't have a board to play with but it sounds like I should order one.

Of course I think you can do direction determination with just GPS heading. If the GPS heading changes drastically in a short period of time, it's most likely the tractor changing direction. I had put in some logic to do that at one time but I never got a chance to test it. If nothing further was done in your branch, I was planning to try this change again with the latest version.

As to how to integrate the magenetometer dating into a smooth heading, that I don't know either. Perhaps it would be most useful at a dead stop before any GPS heading fixes are available, to give an estimate display purposes. In general I think the magnetometer is most used to correct for and calibrate gyro drift on the fly. I suspect that's the main purpose of the magnetomters on the 9dof boards.

If the gyros and magnetometers were needed on agopengps, you would want to mount it on the roof, not in the cab.

Last edited by torriem; 03-19-2017 at 07:41 PM.
torriem is offline  
post #505 of 3741 (permalink) Old 03-19-2017, 08:11 PM Thread Starter
Senior Member
 
Join Date: Aug 2012
Location: Vermilion Alberta Canada
Posts: 5,426
Mentioned: 5 Post(s)
Quoted: 2372 Post(s)
What is the problem if the direction isn't determined properly while backing up? Is it just visual?

The other oddity, for us non flat land operators, is side draft. when a tractor is on a sidehill the back of the tractor, pulled sideways down the hill by the implement, will go "straight" forward at an angle. In this instance the single antenna is far more accurate the angle created is not measured in any way, its just a series of points. The magnetometer however will show up to 10 to 20 degree difference then the GPS heading. Also unsure what one would do for that as you couldn't just use hill slope as you have no idea the traction conditions, weight of draft.....

Its like putting in so many conditions with corrections maybe causes more trouble than preventing.
BrianTee is offline  
post #506 of 3741 (permalink) Old 03-19-2017, 09:31 PM
Senior Member
 
Join Date: Sep 2009
Location: S. AB
Posts: 3,368
Mentioned: 5 Post(s)
Quoted: 828 Post(s)
Yeah you're right. There are a lot of things you *could* account for at the expense of complexity.

So the reverse detection thing is merely something that would be nice, but not essential. It would help me more than just the visual aspect. I do have to regularly back up when spraying around pivot points, which is why it would be nice to detect it. To prevent issues right now I just am in the habit of shutting off the SmartBoom or AgOpenGPS coverage when I reverse so that when I move forward again, the booms will come on again when I reach what I haven't done. Without doing that, I get spurious coverage recordings when the outfit flips around from the 180 degrees change in heading. If that makes sense.

Probably the easiest way to account for side draft is to put GPS receivers on every implement. That's what some guys do with their tow-between air drills.

As it stand right now, AgOpenGPS has everything I need right now, actually, so I'm pretty excited to get it running on a real sprayer in a few short weeks. Tomorrow I'll do another test and see if the coverage gap issue is still appearing.

Last edited by torriem; 03-19-2017 at 09:47 PM.
torriem is offline  
post #507 of 3741 (permalink) Old 03-19-2017, 10:08 PM Thread Starter
Senior Member
 
Join Date: Aug 2012
Location: Vermilion Alberta Canada
Posts: 5,426
Mentioned: 5 Post(s)
Quoted: 2372 Post(s)
Well you may as well have a fresh copy!

v2.1 March 2017
*
* NMEA Logging capability
* BNO055 IMU drivers for angular velocity, roll, pitch
* tilt and roll GPS compensation
* completely functional with or without IMU or relays connected, no setup reqd
* ability to set low speed section turn off
* Parse to tryParse - much faster code for all parsing
* 30 % frame speed improvements by reducing multiple OpenGLreadpixel calls
into a single.
* section minimum distance trigger no longer affected by turnOn time.
* very peculiar bug caused the sections to turn on at 26 km/hr. Individual section lookaheads
had no limits set causing overflow
* Program will now do lookahead successfully up to 36 km/hr. Sufficient for agriculture.
* Will map at any speed, sufficient for aircraft speeds.
* various bug fixes as usual
* Only Roll is GPS compensated. Have not done sufficient testing of pitch, code done, just commented for now

Up on the git..... https://github.com/farmerbriantee/AgOpenGPS
Jmay.david likes this.
BrianTee is offline  
post #508 of 3741 (permalink) Old 03-19-2017, 10:12 PM
Senior Member
 
Join Date: Sep 2009
Location: S. AB
Posts: 3,368
Mentioned: 5 Post(s)
Quoted: 828 Post(s)
Okay I was replaying some test GPS tracks I had recorded back in november with the latest git version of AOG. The skipping problem is showing up with this data set. I'll have to run it again but it looked to me like the skip was occurring when AOG switched between multiple sections and one purple section. I'll have to add some print outs to the code to dump out some debugging info to find out what's going on.

I can post my GPS data if you want to look at it (just an ascii file of the RS232 output). On Linux I made a little python program to replay the file at the right frequency (5hz) at piped it through a utility called socat that sent it to the UDP port.
Attached Images
File Type: png bug1.png (170.4 KB, 25 views)
torriem is offline  
post #509 of 3741 (permalink) Old 03-19-2017, 10:21 PM
Senior Member
 
Join Date: May 2011
Location: Manitoba
Posts: 317
Mentioned: 0 Post(s)
Quoted: 78 Post(s)
Quote:
Originally Posted by BrianTee View Post
What is the problem if the direction isn't determined properly while backing up? Is it just visual?
On our 20/20, it flips the planter around and thinks it's on unplanted ground after backing up, then once it flips back it's on area planted while you were pushing the planter. I have to switch to manual on and remember to switch back to auto once it's all straightened out. Sometimes it thinks the planter is jack knifed and that also causes trouble with coverage. On our GS3, it sometimes shows the combine moving backwards at an angle after backing up. There at least the combine hasn't disengaged automatically. I don't know why Deere hasn't figured that one out, I expected them to have the sensors for it.
m_elias is online now  
post #510 of 3741 (permalink) Old 03-19-2017, 10:36 PM Thread Starter
Senior Member
 
Join Date: Aug 2012
Location: Vermilion Alberta Canada
Posts: 5,426
Mentioned: 5 Post(s)
Quoted: 2372 Post(s)
Quote:
Originally Posted by m_elias View Post
On our 20/20, it flips the planter around and thinks it's on unplanted ground after backing up, then once it flips back it's on area planted while you were pushing the planter. I have to switch to manual on and remember to switch back to auto once it's all straightened out. Sometimes it thinks the planter is jack knifed and that also causes trouble with coverage. On our GS3, it sometimes shows the combine moving backwards at an angle after backing up. There at least the combine hasn't disengaged automatically. I don't know why Deere hasn't figured that one out, I expected them to have the sensors for it.
It is really not an easy problem to solve completely at all! I think most manufacturers have said, not going to do it all because of the possibility of a sharp turn or something else be interpreted as backing up and messing up whe it shouldn't

The most fool proof is 2 antennas in trans mode like wvca does.

BrianTee is offline  
Sponsored Links
Advertisement
 
Reply

Quick Reply
Message:
Options

Register Now



In order to be able to post messages on the The Combine Forum forums, you must first register.
Please enter your desired user name, your email address and other required details in the form below.

User Name:
Password
Please enter a password for your user account. Note that passwords are case-sensitive.

Password:


Confirm Password:
Email Address
Please enter a valid email address for yourself.

Email Address:
OR

Log-in










Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page



Posting Rules  
You may post new threads
You may post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On

 
For the best viewing experience please update your browser to Google Chrome