Help with guiding a tractor

Hey guys,

i was approached by a friend of mine to help with his tractors, what he needs is me to make him something that can show the driver where to go,
how it works currently is that his whole field is sprayed/seeded etc by having some people stand on one side of the field and then also on the other and the tractor needs to drive from one person to the other person, needless to say this is very inefficient as the people don't measure correct and the tractor sometimes doesn't really drive straight to them and misses large pieces..

so im thinking before i think of connecting a stepper or something to the wheel to start is to do a LED strip that has red on the outer edges then yellow then green LED in the middle. the Arduino will read form a GPS and magnetometer to establish position speed and heading, and then light the LEDs to show the driver which way to turn to stay on course.

can anyone tell me if this could keep the tractor within about a meter of a designated track?(what distance error can i expect from a GPS ) where can i find a GPS that's accurate to within a meter?
Will the magnetometer be necessary?

ill save some way points on a SD card and have him work them in order....

anyone have any ideas on how to make it nice and simple? dont think a LCD will work so well in the sun.. maybe a servo with a arrow to show the way??

any ideas welcomed!

Willie

For that kind of accuracy with GPS you will need Differential GPS: Differential GPS - Wikipedia. There's a reference in the wikipedia article about uses in agriculture.

Thanks WildBill,

only prob im in Zambia so don't have the the land based transmitters to use with a differential GPS.

So i read that the GPS is accurate to 5m. i also have an idea of how the gps works but does anyone know what tipe of accuracy i will get with a moving tractor? could i expect to get the tractor to follow a line within like 2 meters? or 3? if i make way points 7 meters apart will that mean the way point can move like 5 meters more?? or could it be pretty stable?

plz anyone with some experience help :slight_smile:

regads
w

Check out this robotic combine harvester.
http://www.robotliving.com/agriculture/robotic-combine-harvester/
He was able to get 1.5 inch accuracy.

only prob im in Zambia so don't have the the land based transmitters to use with a differential GPS.

You can provide your own, I think.

Our reliance on GPS scares the crap out of me... what happens if Uncle Sam decides to pull the plug for whatever reason, eg security. I know there are other systems - the Chinese or is the Russians?- or Europe?- (not sure if implemented or mooted?) but whoever's satellite location system you use, you're at the whim of the owner keeping it active.

The low budget approach would be to use a "foam marker". You just need an air compressor, a small fuel tank, and some soapy water. The compressor bubbles air into the soapy mixture and forces blobs of foam out that you use as a reference to see where you've been.

Chagrin:
The low budget approach would be to use a "foam marker". You just need an air compressor, a small fuel tank, and some soapy water. The compressor bubbles air into the soapy mixture and forces blobs of foam out that you use as a reference to see where you've been.

ROFL!!!!!!!!!!!
i have to say this made my night!!!

im reading about GPSes and looks like 3m is about the closest i could get im of to my frend now to disscuss with him if that would be an acceptable tollerance.

from what ive read about GPS protocolls ... i think with a 3m "possible" error and a magnometer it could actually work.. maybe!

I was wondering if anyone could think of any other ways of doing this? without GPS?

without GPS?

GLONASS.

willievmobile:

Chagrin:
The low budget approach would be to use a "foam marker". You just need an air compressor, a small fuel tank, and some soapy water. The compressor bubbles air into the soapy mixture and forces blobs of foam out that you use as a reference to see where you've been.

ROFL!!!!!!!!!!!
i have to say this made my night!!!

I was being serious; this is the standard way that it is done professionally, albeit with more complex foam generating equipment.

I mean, that's no toy tractor, and at the end of the booms those are the foam droppers.

The foam markers are a common and standard way to prevent overspray for tractor spray rigs (as well as most truck mounted systems).

The other method farmers use is GPS but not a standard uncorrected unit. The professional units use corrections from a variety of external sources such as WAAS, Omnistar, Starfire, etc.

For greater accuracy (<2cm) you can use RTK GPS which uses corrections from a second GPS unit. This second unit may be purchased by the farmer and hosted on his property (costs an extra >$15,000). The other way is to subscribe to companies that provide a subscription service where the corrections are sent to your GPS receiver by mobile phone or radio network.

To do it on the cheap, try using a 'normal' GPS module (without corrections) and put the output through a Kalman filter and see how much deviation you get over a few days.

Is the world ready to take geo-spatial advice from someone called lemming? Personally I'd follow him to the ends of the earth.... Oh, wait...

:smiley:

Chagrin:
I was being serious; this is the standard way that it is done professionally, albeit with more complex foam generating equipment.

Sorry no disrespect intended i just had this image of a tractor racing around the field with a foam party behind it!
that looks like a good idea havin the foam things at the end of our spray booms...

lemming:
The foam markers are a common and standard way to prevent overspray for tractor spray rigs (as well as most truck mounted systems).

The other method farmers use is GPS but not a standard uncorrected unit. The professional units use corrections from a variety of external sources such as WAAS, Omnistar, Starfire, etc.

For greater accuracy (<2cm) you can use RTK GPS which uses corrections from a second GPS unit. This second unit may be purchased by the farmer and hosted on his property (costs an extra >$15,000). The other way is to subscribe to companies that provide a subscription service where the corrections are sent to your GPS receiver by mobile phone or radio network.

To do it on the cheap, try using a 'normal' GPS module (without corrections) and put the output through a Kalman filter and see how much deviation you get over a few days.

Thanks Lemming i was thinking about using a Kalman on the GPS and seeing what the deviation would be over different times.
Also to use the mag to filter out any wayward readings from the GPS. this could increase accuracy yes?

i looked at GLONASS looks interesting and will read some more :slight_smile:

I think the GPS + Mag wil increase the streightness of the lines already as at the moment the tractors drive from person to person by eyeball :smiley:
if i could get a reliable way for the tractor to start at the same places then it will work perfect :slight_smile:

w

Is the world ready to take geo-spatial advice from someone called lemming?

Fair comment, but I do a fair bit of work wiring and setting up RTK GPS systems for linear move and swing arm corner pivot irrigators. I have also programmed an autonomous agricultural vehicle (RTK GPS) using the Haversine formula for steering.

(I adopted the lemming moniker after pragmatism forced me to use Windows (XP Embedded) for a project when I would have liked to use Linux. Felt like a lemming following the crowd.)

simply use laser and recievers...

Khalid:
simply use laser and recievers...

"Simply ..." XD

willievmobile:
I was wondering if anyone could think of any other ways of doing this? without GPS?

You could lay out the initial straight line using inexpensive laser gizmos and manpower, then the tractor driver follows the first plow track like has been done in the old days.

willievmobile:
how it works currently is that his whole field is sprayed/seeded etc by having some people stand on one side of the field and then also on the other and the tractor needs to drive from one person to the other person, needless to say this is very inefficient as the people don't measure correct and the tractor sometimes doesn't really drive straight to them and misses large pieces..

can anyone tell me if this could keep the tractor within about a meter of a designated track?(what distance error can i expect from a GPS ) where can i find a GPS that's accurate to within a meter?
Will the magnetometer be necessary?

I think it is more feasible (in comparison with DGPS approach) to use an IMU (Inertial measurement unit) preferably with an internal GPS implementation. Such devices are commercially available and in general produce reasonably good localization. A list of IMU models can be found at below address:

http://damien.douxchamps.net/research/imu/