Last edit February 4th, 2010 By 8BitMe
I was watching the History Channel and they had a show on potatoes. One part was about how they use GPS in planting potatoes by tractors controlled by GPS and the only thing the driver does is make the 180 degree turns. The GPS helps make the rowes straight so there is less wasted spaces. But given GPS limitations (I think that the government limits the accuracy of GPS) could we use multiple N900 units to ge more precision for GPS (like using 3 N900’s) by having them work together as one unit. I am not sure that they are specically set up to do such a thing but sounds interesting.
Also I was wondering if it would be possible to program N900’s to locate a friend that also has another N900.
Or say if I was at a concert to locate and communicate with others that have N900’s and interact with them at the concert, say share pictures (or video) captured on the fly, or just set up a localized chat for that event. The event sponsor could provide its own web site for the people there by providing a encryption key for those particiapating (it could support other devices as well but to a more limited amount).
What type of interface would one need to have in order to do this. Is the N900 capable of doing this themselves is my question (ie ability to interact with another N900 by itself already). I know there are tons of security issues with this but is it possible to do so.
Again please forgive all spelling / grammar errors………..
by 8BitMe on February 4, 2010 at 12:38 pm
You cannot increase the accuracy of GPS readings by using multiple GPS units (at least not directly). Three units on top of each other (literally) can show differences of 3 meters.
However there is a different way – you can’t increase the precision, but the movement from a base position is highly accurate. it is called real-time kinematics. That has the 10cm accuracy.
http://gpspp.sakura.ne.jp/rtklib/rtklib.htm
You need the “raw” data, but can then say you’ve moved a few centimeters to the north or west because the phase of the satellite signal has changed. Sort of like using a fringe interfereometer.
by tz on February 8, 2010 at 4:34 pm
Thanks for your response. It may be way over my head but I will look at the link and at least should gain some perspective on this matter. The following is just some random speculation and questions for myself regarding your response. So please forgive my ignorance in these matters.
I’m sorry I did’t state that the GPS or N900’s would of course be not on top of each other but in a triangle of one another. Of course my next question regarding this matter is if the variance is random between all the GPS units or is there a way to predict what the difference is between each unit. Is the raw data more accurate due to the reading of the data from the hardware or what causes the variance for the output. Is the weakness due to reception and couls a better antena creat more precise data. What causes the variance of the data, is it on purpose that the data is different on each unit and can this randomness be reduced by any method. I knowlege of mathematical theroy is very limited in this regards so the explanation would need to be in plain english which may make explaination more dificult. As stated previously I am not a coder but hopefully the link will help others. Again thanks for your response. I know my buddy Ryan would have no problems with your link. I know he once told me that he was getting into fluid dynamics and that was a difficult class for him so I would assume that this is similar in theory.
by 8BitMe on February 9, 2010 at 7:44 am
The GPS in the N900 cannot be used for precise measurements. However you can create devices which although not accurate (they don’t know their lat/lon closer than 3 meters), can be precise, i.e. if both are placed right next to each other, they create a 3d “tape measure”. Both will have a phase measurement which will vary equally so cancel out. The error is such that they can’t know there are “on top of each other” unless someone tells them, but once known, if one moves 10cm west, and the other 10cm north, they will have accurate values to that effect – they will know how far they moved and if connected can compare. Or leave one as a reference, so the other will know how far in each direction it moved.
by tz on February 14, 2010 at 1:17 am