Hey all. I'm new here. I've thought about experimenting with an Arduino before, just haven't.
Got to thinking about a project that could save me thousands of dollars, and really shouldn't be that hard.
(You can skip this paragraph if you wish)As some of you may know, farmers employ GPS equipment for better accuracy, to gather information, and in the case of auto steer, improve accuracy, and greatly reduce operator fatigue. It's very accurate equipment that carries a very heavy price tag. The most basic unit starts at $1000 for a light bar only, and goes up to about $9000(screen only). Auto Steer is extra, section control is extra. You can upgrade the accuracy of the GPS to several accuracies, including RTK. To get to RTK, you'd have to buy two CODES for about $3500 each. Point being it's very expensive.
The thing I'm interested in is auto steer, simliar to an Trimble Ez-Steer.(which is $3500) I have a Trimble CFX-750 which has a built in light bar. Basically, on a light bar there's usually 3 lights lit up. Ideally you want the 3 green lights to be on, which means you're right on target. If you're in the red on either side, you want to steer back, to get back into the green. If I took an input from each light(or every other light) (which would require modification to the GPS) and used them as inputs to the Arduino, (with the end goal being to only have input from the 3 middle green lights) it could control a motor to turn the steering wheel. There are other variables that would have to be worked out, such as not over steering.
What are your thoughts? Sorry if this isn't very well articulated. If you want me to clarify anything, just ask. Thank you
On the face of it, it seems doable is all I'm saying. I never said that code was easy to write, but the concept, in my limited understanding of it, seems viable. I just wanted input from people that do have experience with it.
You're paying VERY dearly for the R&D when you buy GPS equipment. $3500 for a motor and wheel that can be told what to do by the terminal is outrageous, but that's how it is.
joelvv:
The most basic unit starts at $1000 for a light bar only, and goes up to about $9000(screen only).
I bet if someone knows how to make a competing system for $800 / $8000 s/he would be busy working on it right now.
There is a huge difference between writing an Arduino code that can light coloured LEDS for Christmas lights and a robust reliable commercial system that is sold with a warranty and after-sales service.
A tractor on a farm must be about the most hostile place you could put a computer.
I think you may have partially misunderstood what I'm wanting to do. I don't want to build a GPS. I want to take signals from the lightbar, and operate a motor with it. I want to do this for myself only, partly because it may save me some money, but mostly because I'm curious about the whole thing. I like challenges, but like I said, I know very little about Arduino's, and was looking for some guidance.
There is nothing wrong with a DIY approach to a problem. I thought you were thinking of a commerical product.
If you have the device that lights up leds to help you steer my guess (and it's no more than that) is that it would provide a too-coarse signal for any steering system. Remember an Arduino (or a PC, for that matter) does not have the computing capability of your brain. When you are steering I suspect you are basing your decisions on a lot more than the LEDs. For example could you steer a straight line across a field in the dark with nothing but the LEDs to guide you? My guess is it would be wobbly line in the correct general direction.
I think you would need to take data directly from the GPS - and even then it would probably be a challenge. For example, what level of GPS accuracy is needed? How do you relate the speed of response of the tractor steering system to the speed of response of the GPS system.
it could control a motor to turn the steering wheel. There are other variables that would have to be worked out, such as not over steering.
The hard part is not the Arduino or the inputs from the light bar, it is the electromechanical components of the steering and the failsafes/human override that you have to build.
I suggest to get that done and working to your satisfaction, and we can help with the Arduino part.
Certainly don't take this as me not needing help or anything, I have given this some thought. I just don't know how capable an Arduino is, so let me know is what i propose here is unrealistic.
On the override side of the system, in a typical hydraulic auto steer system, driver input to the steering system is detected and auto steer disabled until reactivated. In the case of the friction wheel driving the steering system, pulling the wheel away from the steering wheel is how it's typically done, or in this case, is more than realistic.
On the electro-mechanical side of it, I was thinking about using a low speed 12V motor, that could be controlled via PWM. Could it possibly be programmed to treat each input from the lights differently? Meaning the lights furthest from the center of the lightbar, would cause the highest duty cycle to the motor, to speed up correction. This way a light near the center would not cause the motor to race to correct the steering, which would likely cause it to over steer. This would have to be fine tuned of course.
joelvv:
Could it possibly be programmed to treat each input from the lights differently? Meaning the lights furthest from the center of the lightbar, would cause the highest duty cycle to the motor, to speed up correction.
That is certainly doable ...
This way a light near the center would not cause the motor to race to correct the steering, which would likely cause it to over steer. This would have to be fine tuned of course.
... but the devil is in the detail.
Writing code to avoid under- and over-correction can be a very complex business.
I should add that it is not sub inch accuracy I'm after. I'm unlocked for RTK, but am only running WAAS, which with an AG15 is good for 6-8" accuracy. I'd only be running auto steer for spreading fertilizer and spraying. Normally you'd use foam markers, or with smaller booms, eye ball it, which is as accurate as it sounds. In the case of the fertilizer spreader, you have to eye ball your tracks that are 40-50' away, which again is as accurate as it sounds. If I could be within 1', I'd be more than happy with it's performance.
If I wanted sub inch accuracy, I'd have to buy hydraulic auto steer and get an RTK subscription.
Thank you for your help so far.
So when you say that the navigation is easy, steering is not, what do you mean? Are you saying that using the lights as inputs is easy enough, but controlling a motor without under steering or over steering would be difficult?
I meant what I said in reply #6. You are focused on the easy stuff -- reading the light bar, Arduino, GPS navigation, etc.
The really, really hard part is the electromechanical steering gear, with failsafes and overrides. Unless you can buy this off the shelf, mechanical engineering expertise and custom machining will be required. For example, look at what this guy did to simply steer the trolling motor on his canoe.
Perhaps this machinery seems trivial to you, but that has to be done before the Arduino stuff can be designed and debugged.
joelvv:
I should add that it is not sub inch accuracy I'm after. I'm unlocked for RTK, but am only running WAAS, which with an AG15 is good for 6-8" accuracy.
Steering or directional accuracy is not measured in inches or millimeters. It is measured in fractions of a degree - probably tiny fractions of a degree if you wish to stay +/- 6" either side of a straight line along a 100yd field.
I don't know what is the wheelbase of a tractor but even within that distance I suspect a 6" error would represent a very small angle.
Humans find it remarkably easy to steer a vehicle but that disguises the extraordinary complexity that is involved.
If you can do the experiment with complete safety I suggest you try manually steering with no sensory input except the lightbar.
joelvv:
Hey all. I'm new here. I've thought about experimenting with an Arduino before, just haven't.
Got to thinking about a project that could save me thousands of dollars, and really shouldn't be that hard.
(You can skip this paragraph if you wish)As some of you may know, farmers employ GPS equipment for better accuracy, to gather information, and in the case of auto steer, improve accuracy, and greatly reduce operator fatigue. It's very accurate equipment that carries a very heavy price tag. The most basic unit starts at $1000 for a light bar only, and goes up to about $9000(screen only). Auto Steer is extra, section control is extra. You can upgrade the accuracy of the GPS to several accuracies, including RTK. To get to RTK, you'd have to buy two CODES for about $3500 each. Point being it's very expensive.
The thing I'm interested in is auto steer, simliar to an Trimble Ez-Steer.(which is $3500) I have a Trimble CFX-750 which has a built in light bar. Basically, on a light bar there's usually 3 lights lit up. Ideally you want the 3 green lights to be on, which means you're right on target. If you're in the red on either side, you want to steer back, to get back into the green. If I took an input from each light(or every other light) (which would require modification to the GPS) and used them as inputs to the Arduino, (with the end goal being to only have input from the 3 middle green lights) it could control a motor to turn the steering wheel. There are other variables that would have to be worked out, such as not over steering.
What are your thoughts? Sorry if this isn't very well articulated. If you want me to clarify anything, just ask. Thank you
Joelvv, are you still part of the arduino family?
I'm curious if you finished this project , or give up.
This is exactly what i want to try with my Trimble 500 monitor
???
