How do I read the standard deviation of positon error in UBX protocol?

0 votes
Hi Forum,

I'm using an M8P to track the position and velocity of a vehicle (RTK fixed). I recently moved from NMEA to UBX since UBX has support for reading the velocity vector in the NED frame (UBX-NAV-PVT) and NMEA seems to be lacking this. I'm using the NMEA $GPGST message to obtain the standard deviation for latitude, longitude and altitude. I noticed that UBX lacks support for the standard deviations or am I overlooking something? The closest I could find were the hAcc en vAcc attributes in UBX-NAV-PVT. I'm guessing these are standard deviations (1-sigma) but I could not confirm this from documentation. hAcc seems to be a combined latitude longitude deviation so, hAcc should be roughly sqrt(sigma_lat * sigma_lat + sigma_long * sigma_long). Could someone please confirm this for me, or point me to a resource? Thanks!

by dtecta asked 6 days ago
My guess that hAcc should be roughly sqrt(sigma_lat * sigma_lat + sigma_long * sigma_long) seems to be supported by experiments. Also, I noticed that sigma_lat can differ substantially from sigma_long, which is probably due to a limited view of the sky. Still not sure why UBX does not give me sigma_lat and sigma_long separate.
0 votes
Your conclusions are correct.  Note that in all cases, the values are self-estimated with no separate check against "truth" of reality.  Only comparison against a known, surveyed reference gives a more realistic statistic.
by grampy answered 5 days ago
website banner