GPS Repeatability
RogerInHawaii
Posts: 87
I'm wondering about the "repeatability" of GPS location detection. I understand that commercially available GPS units, including those from Parallax, have a resolution of about ten feet. That means that if my GPS unit is showing some particular lat/long location and I move a few feet left or right, back or forward, the unit will continue to show that exact same lat/long position, UNTIL I move sufficiently far away (about ten feet) from my initial position, at which point it will then show the next adjacent lat/long position.
Now, suppose I configure my GPS unit so that it clicks each time it detects a unit change in its lat/long position. I slowly move the unit until it clicks and I place a marker on the ground where the transition occurred, i.e. where the unit detected that I moved from one detectable position to the "next" detectable position. I move a few feet further and then turn around, moving back again over that position where the click occurred. And here's my question: will the unit once again click at the exact same position? Is the position detection "repeatable" so that going back and forth over the same ground will result in clicks in the same position? If not, how much variation should I expect to get from the unit? Is it always within ten feet, or five feet, or what?
Now, suppose I configure my GPS unit so that it clicks each time it detects a unit change in its lat/long position. I slowly move the unit until it clicks and I place a marker on the ground where the transition occurred, i.e. where the unit detected that I moved from one detectable position to the "next" detectable position. I move a few feet further and then turn around, moving back again over that position where the click occurred. And here's my question: will the unit once again click at the exact same position? Is the position detection "repeatable" so that going back and forth over the same ground will result in clicks in the same position? If not, how much variation should I expect to get from the unit? Is it always within ten feet, or five feet, or what?
Comments
There are many factors/variables in play with a GPS based locating system. In certain conditions, it is completely possible to stand completely still and have your "current location" bounce around inside a 15 foot radius circle. This is simply in inherent factor of the GPS system. Your GPS is attempting to pinpoint it's location from GPS satellites that it currently can receive properly. These signals are affected by many variables (atmosphere, obstructions, etc) and your GPS receiver's quality and antenna integrity adds to that variability. These are some the problems of the accuracy of the GPS system.
However, there are many ways to compensate for these variables to increase the repeatability of your GPS. One is WAAS, which helps compensate for some atmospheric issues among other things. My Garmin GPS uses WAAS and when I use it for geocaching, I am usually within 5 feet of any geocache's actual location when using it's coordinates. Other methods are built into the GPS itself. These involve sophisticated algorithms such as Kalmann Filtering to reduce the affect of varying signals received from the satellites. There are some Prop threads with Kalman filtering code, just search the forum.
So, with the same WAAS enabled GPS, excellent weather/atmospheric conditions, no obstructions, etc, you should be able to walk a 30 foot line back and forth, place markers each direction at the changes you indicate, and end up with markers within 3 feet of each other.
I know companies use radio tags to keep track of stuff, but I think that only tells when something enters/exists a room and not it's place in the room.
If this could be developed it could spread out like cell phone and wi-fi coverage.....just don't know if the market for a better GPS system would be there to cover the cost.
DGPS requires a fixed instrument sending correction data to the moving one. Range limit is 10-20 km.
There are solutions involving an internet connection, multiple bases, a central server offering this resolution over a wider area.
I use it for surveying tasks, and you can get impressive results. Hardware and software are completely different, and rather expensive.
Moreover in a room you have not satellite view, so a 10-50k$ solution would still be useless.
Massimo