im looking for a GPS module with the highest possible horizontal accuracy. The device will be used for navigation so it will be on the move (slowly, < 10 mph). Time to fix is not a concern. Just browsing though spec sheets ive found this one
that claims 3m w/ standalone CPE. This module is not offered by sparkfun, but is there a comparable one offered by sparkfun? Are these numbers provided by the spec sheets real world or ideal test conditions? What can I expect in real world? Ive never done any GPS embedded work before so this will be a new area of embedded for me.
Real world GPS accuracy is about 5-7m depending on how many sat’s you have in view. Ideal accuracy on normal GPS is 3m by design (this is a limitation of the satellites, set forth by the government).
There are two “solutions” to this…
Use two GPS receivers mounted a known distance from each other. Take the two readings and compare them, then average out the distances to get more accuracy.
Use a system called DGPS, or differential GPS. This is GPS in conjunction with shore based stations serviced by the US Coast Guard. DGPS gives sub-meter accuracy in ideal conditions.
the u-blox ones look nice but they dont look like their avaliable from normal supply houses (digikey, mouser), also the cheapest is $100. I was hopping to stay around half that. Is it asking too much to get <1 meter
accuracy for ~$50?
Shifted, Im thinking DGPS might be the best choice. Do I just look for a module that can handle DGPS? Also does accuracy vary on how many of these land stations are available? I assume im in a best case scenario being on the west coast of FL, Tampa bay area. Theres a coast guard air station probably 5 miles away from me.
DGPS typically has shore stations that reach 120 miles out into the gulf of mexico. You’ll have to look at the USCG Navigation operations page to see more details though. Some DGPS are subscription based, some free. It really depends on what your budget is.
ehh maybe im taking the wrong approach. The end device will be an autonomous robot. its not going to be traveling the world or anything, and in maybe about half the cases will be used in doors. Maybe its a better idea to create a beacon type navigation system? Like DGPS, only my own beacon that will be in a peppermint location. I don’t even know where to start for that. what should I look up to get started?
Novatel has recievers for RTK which requires a base station. It uses GPS in conjunction with your own beacon in a known location, it can derive position to within 2cm. I spoke with Novatel for an application that I have, the RTK beacons can work at up to around 40 miles in optimal conditions. The receivers are around $1500, not sure what a base station costs, and then you have to add an antenna on top of that and the RTK receiver radio.
I suppose I should ask… what kind of budget do you have? Making something industrial, or just a home project?
really im looking to make an wheeled robot that will be able to intelligently map an area and preform repeatable tasks within that area. Something that could be used indoors to vacuum and outdoors to cut grass. Ide like something more intelligent thats a roomba which is biased on random movement (i think, ive never owned one) . For a (simple) example i want to give it coordinates for 4 corners of a box and just have it run laps progressively moving to the other side of the box with each pass till it completes. Indoors the range to the beacon (or beacons) will only be like 20 feets. but outdoors it would need to be ~50’ and may be going out of LOS of the beacon. GPS can be completely drooped from the equation, I don’t really care about its position on the earth, just that it stays on route.