Hi,
Most of the GPS modules provide an accuracy of less than 5m. My application requires a better accuracy than this (1m would be great).
Would a Kalman filter help providing a better accuracy? Or are Kalman filters already implemented in most GPS modules? Thus processing the GPS data with a Kalman filter would not be necessary.
Thanks!
For a Kalman filter you need multiple sources of the same data (all it does is sensor fusion). For your application you’d use some accelerometers, since integrating an accelerometer twice gives you displacement just like a GPS, and the kalman filter would work out which of the GPS and accelerometer to ‘trust’ when determining your position.
You could probably do 1m accuracy with some of the MEMS accelerometers that Sparkfun sells - the only way to find out is to try 
Will.
Have you had a look at the resolution of DGPS? I seem to recall that it was on the order of a meter or two, but I may be completely wrong! 
jeanseb:
Hi,Most of the GPS modules provide an accuracy of less than 5m. My application requires a better accuracy than this (1m would be great).
Would a Kalman filter help providing a better accuracy? Or are Kalman filters already implemented in most GPS modules? Thus processing the GPS data with a Kalman filter would not be necessary.
Thanks!
about the best you can get is to ensure that your receiver supports WAAS and that you have that enabled. If you are receiving WAAS differential corrections, your GPS fix types would be so-flagged.
To receive WAAS, you need to see the equatorial plane of satellites since WAAS is sent via geo satellites, not GPS satellites. It also depends where you are on the globe; there are WAAS equivalents in other countries. So in north America, you need a clear line of sight to, say, 34 degrees elevation from the horizon. Free of buildings and trees.
Otherwise, you’ll have to get differential GPS corrections from the Internet and feed them to the GPS receiver’s 2nd serial port per the RTCM standard, or buy/use a DGPS data provider service.
Or spend $$$$$$$$$$$$$$$$$$ on RTK GPS.
1m short term accuracy is a marginally realistic expectation.