retoweiss
commented
6 years ago Together with such a major enhancement one should also consider a new definition of X,Y and Z-accuracy, currently just defined as HSDV, VSDV resp..
HSDV and VSDV are not sufficient, since they are calculated on a very short observation interval, typically 10 seconds.
Instead of just averaging HSDV and VSDV one could also take into account
- multiplication with HDOP and VDOP (horiz./vertical dilution of precision)
- "reliability bonus" in case of multiple measurements spread along a larger time period, e.g. 1 hour
- theoretical precision of the GNSS unit as maximum accuracy
ubx
Until now we claimed that all measurements belonging to the same GCP are continuous, that's not always true, as it turns out.
Example: GCP5 can only be measured once and maybe with a a signifficant HSDV/VSDV, hence we continue with other GCPs and come back later.
In order to recognize these cases, we can no longer process the measurement one by one. Instead, we have to temporary store them in an array (or so) . Upon identifying all the measurements for a given GCP, we then can consider every sinle measurement, not only the next one.