Welcome to the Parallax Discussion Forums, sign-up to participate.

RogerInHawaii
Posts: **87**

GPS units have a resolution about about ten feet, which means that you can move it around a bit and it will keep showing the same latitude/longitude until you get about ten feet away from your "start" location and only then will it show its new position. This is fine for applications like driving but not so useful for a small robot trying to cut te grass in the back yard without also cutting the flowers in the garden.

So I'm wondering about the sensitivity and repeatability of GPS units.

Suppose I have a GPS unit displaying lat/long, at some position in my yard. I mark the "initial position" on the ground. I then slowly move the unit, inch by inch, in one direction. When the unit "transitions" to a different lat/long position I make a mark on the ground. I bring it back to the initial position and again move it, inch by inch, in the same direction. Will it transition at the same point that it did the previous time? If I do this a hundred times and each time marking where the transition occurred will all the marks be in precisely the same place, or scattered about? If they're scattered how large a scatter "area" will they occupy? If I do this in many directions away from my initial point will I get something comparable to a circle (or a square?) drawn by all the dots I've drawn?

·

So I'm wondering about the sensitivity and repeatability of GPS units.

Suppose I have a GPS unit displaying lat/long, at some position in my yard. I mark the "initial position" on the ground. I then slowly move the unit, inch by inch, in one direction. When the unit "transitions" to a different lat/long position I make a mark on the ground. I bring it back to the initial position and again move it, inch by inch, in the same direction. Will it transition at the same point that it did the previous time? If I do this a hundred times and each time marking where the transition occurred will all the marks be in precisely the same place, or scattered about? If they're scattered how large a scatter "area" will they occupy? If I do this in many directions away from my initial point will I get something comparable to a circle (or a square?) drawn by all the dots I've drawn?

·

## Comments

106Your question does not have a simple answer - it is a function of the speed that you're moving at and the sensitivity and repeatability of the GPS chipset and the design of the receiver. You can purchase a unit for ~$10k that has a dual receiver,·differential correction and high refresh rate that is accurate within a few centimeters. The newer consumer-oriented modules that use the SiRF Star III 20-channel chipset with a·5Hz refresh rate have an error of approximately 1.5 meters due to the slower refresh rate, lack of correction and a single receiver topology.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔

Craig Eid

www.TriadRD.com

87Thanks for the info. I had read about those more expensive units with the super high resolution. My understanding was that they really only work at that resolution when the unit is either stationary (e.g. using it for surveying) or moving very slowly (e.g. a very slow farm tractor). I was also under the impression that they don't just use the satellte GPS but that they also had to be within range of some sort of ground-based government transmitter.

But knowing the higher resolution doesn't really address my question. Whatever the resolution, I'm curious about repeatability. If I move the unit around, as in my description in the original posting, how repeatable are the readings? Will it always transition to the "next" location at the same position as I move it around, presuming I'm moving it at a reasonably slow rate?

Roger

531