Hi
Warning: This may by a silly question(!) :o
How should GPS accuracy information (in meteres / centi meters) be understood?
Example: When my RTK facet says accuracy (vertical) is currently 1 meter, does that mean accuracy is within a 1 meter diameter circle or 1 meter radius?
Regards
Roman
Hi Roman,
“There are no silly questions. Only silly answers…”. I hope this isn’t a silly answer:
I believe the accuracy estimates are “Circular Error Probability (CEP) CEP refers to the radius of a circle in which 50% of the values occur, i.e. if a CEP of 5 meters is quoted then 50% of horizontal point positions should be within 5 meters of the true position.”. (Note: the module does not know the true position, it can only estimate it!)
Best wishes,
Paul
Hi Paul
Thank you, that makes things a bit clearer.
Regards
Roman
Paul gave a great answer. Especially that it’s an error estimate!
The reported accuracy estimate is based on a statistical model that’s presumed to have a Gaussian (bell curve) distribution. The hard facts of life include that if 50% fall within the reported error, then 50% fall outside the reported error. So if the CEP is 5 meters, then 50% of the coordinate reports have an error greater than 5 meters. Yes I’m stating the obvious, just wanted to put it out there. This is not a knock on the SparkFun units, this is how most all GNSS receivers work.
Also, generally the reported error estimate only includes the errors the GNSS receiver can model. There potentially are error sources that the receiver can not model (eg multipath, jamming), and are not included in the error estimate.
I generally triple the reported accuracy to get a good sense of what I might be able to depend on. If things are important, I might take multiple shots hours apart and compare coordinates.
Tony.
PS I might quibble if it’s a 50% or 68% (one std dev) confidence interval, but I forget and would need to dig out the u-blox specs.