It seems that I'm constantly having problems with my airspeed sensor and the values it reports when I'm at the field. However when I'm sitting in an enclosed room I run tests with a fan and my anemometer and the sensor is almost dead accurate. I have read several posts here mentioned that you need to be sure that the pitot tubes aren't pointed into the wind when you turn on the APM, which I try to do each time, but in the field I get extremely inconsistent results.

Just yesterday in about 1m/s winds the airpseed sensor was reading 17m/s and the ground speed 8m/s while flying into the wind. Curiously though the indicated wind velocity in the mission planner was only 2m/s at altitude, so it seemed to figure that out on its own somehow. The only time I've managed to get the airspeed somewhat accurate is when there was on an extremely rare day where there was absolutely zero wind.

Of course having an airspeed sensor that is completely inaccurate wreaks havoc with the PIDs, one flight could be perfect and the next could result in a stall/crash if the APM thinks the plane is flying at 17m/s and throttles back too far. The only difference being the direction of the wind relative to the plane when it was first plugged in, combined with gusts and wind flow around obstacles it can be difficult to get this right.

So my question is: Similar to the manual level mode for the accelerometers/gyros, is it practical/possible to have a similar manual calibration mode for the airspeed sensor? I realize temperature and air density all play factors, but based on my own experience it seems like it would be more accurate to calibrate it in a controlled environment then taking chances with calibrating it on the field. The manual level mode has been wonderful for me in this regard.

Even better would be some sort of algorithm to automatically adjust the calibration based on in-flight data, similar to the auto-calibration of the compass, but that is probably much more difficult.