g.bartoli Posted January 30, 2017 at 09:54 AM Share Posted January 30, 2017 at 09:54 AM Dear TinkerForge, We recently bought a GPS Bricklet unit and we would like to have some more details about the data received in the callbacks. We successfully used the coordinate callback as explained here: https://www.tinkerforge.com/en/doc/Software/Bricklets/GPS_Bricklet_C.html#c-c-gps-bricklet The data contains also the EPE value which we read about here: https://www.nps.gov/gis/gps/WhatisEPE.html What we did not understand is which confidence interval is used: 50%, 68%, 95%, 99.7? We need this information to calculate the position covariance (standard deviation) following this approach: https://en.wikipedia.org/wiki/Standard_deviation#Rules_for_normally_distributed_data Many thanks, Guido Quote Link to comment Share on other sites More sharing options...
borg Posted January 31, 2017 at 11:20 AM Share Posted January 31, 2017 at 11:20 AM I am not sure to be honest . The datasheet does not mention anything about the EPE confidence interval. Quote Link to comment Share on other sites More sharing options...
g.bartoli Posted January 31, 2017 at 11:43 AM Author Share Posted January 31, 2017 at 11:43 AM Ok, looking through the Datasheet by myself, maybe I found it at page 16: https://github.com/Tinkerforge/gps-bricklet/raw/master/datasheets/FGPMMOPA6H.pdf Position Accuracy Without aid:3.0m (50% CEP) DGPS(SBAS(WAAS,EGNOS,MSAS)):2.5m (50% CEP) But I do not know how to relate this fixed CEP specification (which seems the best case) with the variable EPE measurement coming from the GPS Bricklet (often I receive about 1.3m, which is surprisingly low)... can I consider them equivalent measures? Quote Link to comment Share on other sites More sharing options...
andycruce Posted February 20, 2017 at 11:22 AM Share Posted February 20, 2017 at 11:22 AM Based on your question about getting covariance values for the GPS errors, it sounds like you are working on a Kalman filter to fuse GPS and IMU data. Is this the case? I am interested in the same thing but need someone to talk to about the actual kalman filter implementation. If you are working on the kalman filter I'd like to converse with you. My email is andycruce@comcast.net. Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.