PDA

View Full Version : GPS Inaccuracies Question



IRobot2
06-06-2005, 04:07 AM
This question stems from Al’s post on "Telemetry with a Boe-Bot or Two?"

I have heard time and time again that GPS is only accurate to about 10 -15ft. Is this inaccuracy held only at each GPS unit by itself, or is it a random inaccuracy in a given area? It is hard for me to put in words what I am thinking so let me say it differently.

If I had two of the same GPS units and held them side by side… would they give me the same readouts? (Inaccurate to the world) Or would they give me different readouts within say a 10 ft area. (Inaccurate to themselves)

I ask this because if you had a yard and put a GPS “station” at one corner and a GPS “station” in another corner. And let a robot with a GPS unit, able to communicate with the “stations”, roam around the yard. Could you not triangulate where it was at? It may be inaccurate to the world, but accurate to its current location. (You would know where it was at in the yard) But that is only if all three units had the same inaccuracies.

I was really curious as to how the GPS inaccuracies work anyway. Because I know some farm equipment uses GPS to guide gigantic equipment down very accurate paths. Yet as hobbyist we can barely get a robot to realize what city bock it is on.

Any information would be great!




▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Alex Burke

burke@ajbrobotics.com (mailto:burke@ajbrobotics.com)

"The factory of the future will have only two employees, a man and a dog. The man will be there to feed the dog. The dog will be there to keep the man from touching the equipment." ~Warren G. Bennis!

steve_b
06-06-2005, 05:44 AM
Someone will correct me if I'm wrong (or when!! lol).

The GPS system stemmed from the US military. They wanted a way to locate ppl/places to a very accurate degree (don't ask me what it is).
The military opened it up to the public but didn't want ppl to have the ability to use the location system as a means of dropping ordinance, so they put in an "offset".
Back in the day, you're accuracy was 50m or so (3m today). You couldn't get a better accuracy without a "black box" (for lack of a better term) that would calculate out the offset.
Then there's DGPS (differential) that averages a bunch of readings over a given time and determines (by calculation) what the location is.
Today there is WAAS (Wide Area Augmentation System...hope I got that right) that offers a landbased 'truthing' for WAAS enabled GPS's to get more accuracy.
GPS is great for 'out in the open' uses where you don't need more than 20ft resolution (if you're within 20ft of a cliff or a Mcdonalds...you'll hopefully see/smell it) which is why not many mini/combat bots use it (except for playing outdoors).
You could use GPS to 'ballpark' where you are then you'd have to switch to some other telemetry system to inch it in!

And really, close only matters in horseshoes and hand grenades!!

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔


Steve
http://ca.geocities.com/steve.brady@rogers.com/index.html
"Inside each and every one of us is our one, true authentic swing. Something we was born with. Something that's ours and ours alone. Something that can't be learned... something that's got to be remembered."

Warren
06-06-2005, 07:03 AM
The accurate tractor and other precision agricultural implements use gps plus correction factors that are broadcast to the farm implement from a dedicated base station. The cost of the base station is dropping: I believe they were around $40K, but have dropped to $15K (+/-). The advantage of using this hybrid system is location accuracy, generally precision to under one inch. These gps-guided tractors and implements have become very common in the large, flat and typically rectangular open farm fields of central California.

IRobot2
06-06-2005, 07:44 AM
What does this base station broadcast back to the tractor? Is this something that can be imitated on a smaller scale for hobbyist? I would assume not because of the price tag. But then again, we may not need all of the data that they do. I live in Southern Illinois alot of this goes on around me, if not down the road. HaHa I just thought it was really interesting. Any one have any other types of ideas how to do somthing like this?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Alex Burke

burke@ajbrobotics.com (mailto:burke@ajbrobotics.com)

"The factory of the future will have only two employees, a man and a dog. The man will be there to feed the dog. The dog will be there to keep the man from touching the equipment." ~Warren G. Bennis!

PLJack
06-06-2005, 08:20 AM
First read this:
gpsinformation.net/main/poordgps.htm (http://gpsinformation.net/main/poordgps.htm)

Some background:
www.beaglesoft.com/gpstechnology.htm (http://www.beaglesoft.com/gpstechnology.htm)

More information than you can handle:
gpsinformation.net/ (http://gpsinformation.net/)

Playing two economy GPS's off each other will not work.
Even placing them two inches apart, they will never agree.

As for "Inaccurate to the world" or "Inaccurate to themselves" it is neither.
The location is an estimate only, worldly or otherwise. Even then it will vary over time.
If you mark a location in your yard, that position will move over time.

It just don't work.
That is why an accurate GPS cost good money.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Jack

Warren
06-06-2005, 08:53 AM
Here are some non-technical references that would help point you to more technical information, if you are interested.

http://farmindustrynews.com/mag/farming_allinone_autosteer/

http://www.sunbeltexpo.com/press/index.cgi?pr=171

http://www.trimble.com/news/080604a.htm

Forrest
06-06-2005, 09:30 AM
I don't know if you realize it - but GPS requires an unobstructed view of the sky to operate. For smaller distances, you might be better off with an electronic compass such as the $30 model from Parallax that's accuracte to +/-2.7 degrees after calibration. Other users in this forum have suggested a triangulation scheme with IR emitter/detectors at the corners of the field and the robot has an IR emitter/detector mounted inside a tube cut with slots. The tube is controlled by a servo and the position is calculated when the IR beam from the corner beacons is detected.

IRobot2
06-06-2005, 10:38 AM
Thank you for all the info guys. I have learned more than I ever wanted to about GPS. And like many others I am sure, I am disappointed that all the ideas in my head are now diminished. But still good to know!

Since my robot was to be out of doors, would IR emitter/detectors be flooded by sunlight? What is the maximum range that I could go before loosing contact? I am sure it would all depend on the parts that I use but I did not know if it were even plausible to consider something like that for a typical yard length.

Any other ideas for making a robot find its way around a yard?

Thanks again for all of your help. This is really cool stuff!



▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Alex Burke

burke@ajbrobotics.com (mailto:burke@ajbrobotics.com)

"The factory of the future will have only two employees, a man and a dog. The man will be there to feed the dog. The dog will be there to keep the man from touching the equipment." ~Warren G. Bennis!

steve_b
06-06-2005, 05:17 PM
you could use a compass module with some counters on your motor assembly. Count revolutions of a wheel or tread and calculate how many revolutions in a certain direction and you'll get a good idea of where you are from where you were (still need to know where you were to a good degree though).
This all might need a 2nd controller if you have a lot going on on your Bot!

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔


Steve
http://ca.geocities.com/steve.brady@rogers.com/index.html
"Inside each and every one of us is our one, true authentic swing. Something we was born with. Something that's ours and ours alone. Something that can't be learned... something that's got to be remembered."