[BC] GPS recommendations

[Jonathan E. Hardis]

On Jun 15, 2012, at 2:29 PM, RichardBJohnson at comcast.net wrote:

> The government ...
> 
> http://www.gps.gov/systems/gps/performance/accuracy/
> 
> ...says that some receivers provide better than the "theoretical" 7.8 meter accuracy. How is this possible? No, it is not done with smoke and mirrors. If your receiver is not moving like an airplane moves, one can obtain the GPS position over-and-over again and average the result.

There are several methods for improving the accuracy of a GPS receiver.

1)  Much of the error arises from variability of the signal delay in the ionosphere and the troposphere.  Good receivers have models to compensate for the variability.  These models are improved by monitoring two frequencies (traditionally, both L1 and L2, though L5 is beginning to be deployed).  While the data on L2 are encrypted, there are "semicodeless" techniques that glean the necessary information from the signal.  (Note that virtually no consumer grade receivers use two frequencies, as that would more than double its cost and complexity.)

2)  Use of a "Real Time Network," such as http://www.theorgn.org/.  This requires an additional data path from the network to the receiver, e.g., a cellular data link, or a commercial service from a geostationary satellite.

3)  Use of "Real Time Kinetic" techniques.  http://en.wikipedia.org/wiki/Real_Time_Kinematic

4)  Post processing the data.  GPS depends on knowing the exact positions of the satellites as a function of time.  The navigation message provides estimates in advance as polynomials, however these estimates can be hours old.  Much better estimates are available in retrospect.

     - Jonathan