I know that is typically how a flight computer computes wind, even on a real aircraft. BUT:
1) What if you aren't flying forward. Mag heading wouldn't be the direction it is pointing on a copter.
2) There is no airspeed sensor.
Hi P2P, I've wondered this. I'm hoping to write some code to get a 3d wind vector estimate from a quadcopter for a research application. I was thinking that there should be a way to hack the new EKF code to get this if you assume that differences between observed and expected behaviour are due to wind.
Pretty cool... I'm looking for something applicable to multirotors, though.
That was way too much math. I stop doing math when letters and other funny looking symbols pop up. I'll take your word for it. Does this apply to or not apply to multirotors?
Is This a relative wind with respect to APM or is it true wind prediction ?
As far as I know, the method referenced by Sam does not work for multirotors. Perhaps someone from the APM team can comment on what they are using for multirotors.
The method computes true wind vector in earth frame of reference. It has been used routinely for fixed wing aircraft by several autopilots, including APM and MatrixPilot. It does provide a reasonably accurate estimate of both the airspeed and the wind speed without an airspeed sensor, provided the aircraft makes a few turns once in a while. The details of the turn are not important, as long as the aircraft changes direction.