1

Problem: azimuth value changes when pitch or roll change even if the device orientation is still the same; I need to get stable independent azimuth and pitch: tilt should not affect yaw.

I've already read and tested all solutions proposed here with all sensors available and with or without LPF, remapping, etc. but didn't manage to sort that out.

Is there anyone who can solve this problem? I need to have a stable azimuth for at least + or - 45 deg pitch change wrt the horizontal plane and viceversa. Thanks in advance.

abe65xt
  • 11
  • 2

1 Answers1

0

azimuth is very trickie thing.

first of all remember that you need to calculate change when the orientaton is changed. the second thing you need to know that it is very sensitive for diffrent kind of distortions. For example if you're testing it near some electrical thing, for example your pc, than your results can jump.

I've made some experiments during my masters and I've found out that one of best ways to filter distortions is you count average from some set of medians :) it practice it is very easy: you create set of results, than you pick median from it and pass it to second collection from witch you're counting average. It will allow you to reice preatty acurate results with out distortions but it can make some small delay in actualizations

Fixus
  • 4,631
  • 10
  • 38
  • 67
  • Thank you for the reply. Sure I'm debugging close to the PC, which can cause some interference. But the change in yaw is too high compared to the pitch variation and there seems to be no functional relationship between them. – abe65xt Sep 08 '15 at 14:14
  • 1
    @abe65xt try my method and check if it will help or paste your code because currently we're just talking about some theoretical problem :) – Fixus Sep 08 '15 at 18:26