Improve realism of aircraft simulation

[adegroote]

This set of (relatively disconnected) commits allow to improve the realism of Morse simulation for aircraft. The first one allow to choose if we control the attitude of the aircraft in yaw_rate or in absolute yaw, a common choice in nowadays autopilot. The second one allow to simulate the drag force, i.e. air resistance. I choose to simulate it using an "uncontrollable" actuator, but I'm opened to other design idea. The last one extend (and break) the force/torque interface to allow to add force in the global frame. It will be useful to simulate (at a "high-level") wind.

[flixr]

Just keep in mind that you can't easily have multiple force/torque actuators at the same time (without changes). The force/torque set by the last one will actually be applied and not the sum of all...

[severin-lemaignan]

@flixr this is a very good point, indeed! @adegroote did you think of it?

[adegroote]

Where do you get this information from ? I checked the blender / bullet code, and clearly the sum of force is applied (and my tests seems to confirm).

CcdPhysicsController::ApplyForce calls btRigidBody::applyCentralForce which does

m_totalForce += force*m_linearFactor;

m_totalForce is reset once after each integration step.

[severin-lemaignan]

Then good! Would it be possible to write a test (for instance with and without wind) to ensure we are correct here?

[adegroote]

Done (but it looks like there is a bug in accelerometer (but it is an another story :D))

[severin-lemaignan]

Alright! Let's merge it then!

[flixr]

Sorry for the noise then... Either it changed in Blender or I am not remembering correctly... in any case I had to work around some similar problems 2 years ago...

[adegroote]

Merged after tests improvement and changelog addition. :) Next task: review #673 :)