Hello, this is my first post.

With FPV drones, I noticed that the camera looks forwards and the drone is skilfully flown so that it points its axis at the subject while changing its flight direction relative to the body of the drone. Flying past an object becomes a balance of controlling the angle of drone axis vs the flight direction from approximately the same angle, to 180 degrees apart. Difficult, and impressive, if accurate.

All drone operators, that I have seen, are basically stationary during flight. No head or body movement.

That could change! In the near future, their heads could be looking around...

The idea that I have is to measure the 2 angular differences between the joystick controls, and the goggles (Simply relative to gravity doesn't give you left/right)
Alternatively, 3-axis inertial measurement in the goggles could give all the information that is required, but these sensors tend to drift over time.
The measured goggle angle is copied by the angle between the drone camera and the drone body.

Obviously, a drone operator hasn't quite evolved to "Owl status" (that'll take a few more years), and 180 degree control is not possible, but up to 90 degree movements could be controlled. <Edit: move hands and head in opposite directions to get more than 90 degrees!>

This is particularly useful, because the subject can be more easily maintained in the centre of view by movements of the head. Also, when turning, the operator could look where he wants to go, rather than along the current axis of the drone.

Head trackers have been around for years, just mainly used on fixed wing.

mc54 Posted at 1-30 04:03
Head trackers have been around for years, just mainly used on fixed wing.

Yes, but as the OP says, they have inherent problems. They use IMUs, which do drift and often don't do what you really want them to do.

I've tried to use the head tracking in my DJI goggles with my MP2 and I start getting airsick almost immediately after turning it on . It's the most awful, and unrealistic, thing in the world.

Instead, what we really need is a duplicate of what we actually do with the naked eye and our heads: a full 180+ degree FOV, which is stationary, and our heads free to move relative to it. That's how it is in manned aviation also. When you turn the aircraft, the view turns with the aircraft, but you're still able to turn your head relative to the view, usually to keep your eyes fixed on a spot on the ground, for example when shooting a landing, etc.

But that technology is years, probably decades, away unfortunately.

OP here. Thanks for the comments. I hadn't previously seen the head trackers for remote controlled aircraft, but armed with knowing what they are generally named I was able to do some searching of various patents and applications.

It seems they fall into two types:

* Camera based - usually military applications. Camera is on the helmet, and tracks marker points around the cockpit. This could be used by DJI easily, since they already have advanced vision tracking, so don't even need the markers. I think this would be OK up to a point. It needs to be fast and accurate or, as luciens mentions, nausea rapidly arises. Furthermore, the measurement must be "head relative to something", and that something could be the general environment, OR the handheld controller (only if line of sight is maintained, for which it probably will not).

* Accelerometer type. Cheap, but there will need to be a calibration routine and elimination of drift over time. It's relative to itself (but it integrates acceleration twice to get distance - hence the drift). I didn't check how the angle calculation is made from x,y,z distance...I guess there must be a system where two accelerometers are balanced around an axis, and they are subtracted one from the other, or something like that, and then integrated twice over time. I think it just means it gets more difficult to get stability.

But there is a fairly new technology (at least in more common usage, which means it's getting cheaper) - check the specs of iPhones and the latest Samsung Note and 21 Ultra, for example - it's called Ultra Wide Band, UWB. Some manufacturers (not Apple or Samsung) are claiming mm accuracy, and able to detect direction with error of max 6 degrees but only in a hemispherical target range. 150Hz refresh rate. Max range 200m. Interference free. Maybe not all of these things at the same time. There is an advantage to this type (probably some disadvantages, too - cost?). Line of sight not needed. This could reference between the headset and the hand controls. It means the hands and head could effectively rotate 180 degrees from each other by looking one way, and pointing the controller the other way.

Dave