RTKLIB Timestamp Post Processing

Could anybody point me in the right direction for some info on post processing a position from a timestamp and raw data with RTKlib?
I would like to start looking into this while waiting for my Reach.

Thanks!

Sam

2 Likes

The basic algorithm works like this:

  1. Log Ublox data stream with raw data and time-stamp messages.
  2. Convert raw data to Rinex.
  3. Post-process Rinex with RTKPOST.
  4. Extract time-stamps from the log.
  5. Interpolate points using solution data and time-stamps (would work great for things moving in straight lines).

To further improve point calculation we should also use IMU data (e.g. direction and speed).

1 Like

Nice one! Sounds great.

@mikhail.avkhimenia is point 5 implemented into RTKPOST in the mean time or do I have to do that myself?

It is implemented in our fork of RTKLIB:

That is great!

Is there any research about how different frequencies (1 or 5 Hz) influence the position on a copter flying with a speed of 5 m/s?

Depends on the frame, environment, usage scenario etc.
Higher frequency is expected to be better, but may not be necessary.

14Hz GPS only vs 1or 5Hz GPS+GLONASS should be interesting to know !
End-user product like ebee are just at 1 Hz between base and flying wing

Do you know how fast the update rate on the ebee with rtk is? Watched one flying today in moderate wind (4m/s) at 10m/s and interpolating from 1Hz cannot result in a precise position.

Do you know how the images in the ebee/sequioa are tagged and whether it is possiple to interface the sequoia with PixHawk or Reach?

Ebee is capable of 100Hz but has really 20hz record for navigation, 10Hz for logging and get corrections @1hz.

Sequoia is another story, I have one and it have his own GPS on sunshine sensor. Not reliable, you could not expect accuracy from it…