On Mon, 16 Dec 2002 18:41:50 -0000 "Jim Wilson" <[EMAIL PROTECTED]> wrote:
This is exactly the issue: as we all know, the rotations are fine (a rotation is a rotation - the *orientation* of the vehicle always ends up correct). It's the translation issue that is cloudy. I think what we will do (and I think Tony is agreeable to this, and Andy sounds like he'll be able to do this if he is not already) is to provide the lat/lon/alt (LLA) of the nose/prop_hub_tip for the aircraft in our coordinate system. To restate what our coordinate system is:
Yes and no. It should greatly improve the situation as far as keeping the
gear above the pavement in an external view. The other issue is the one I described before where it looks odd because the camera tracks the nose as it pitches up and down. That needs to be dealt with in the viewer
code.
X positive aft
Y positive right
Z positive up
origin - now irrelevant given that we would supply the lat/lon/alt of the NRP (Nose Ref. Pt.). I'm going to have to think for a moment how to supply that. We'll have the LLA of the CG, and we'll have the body to local transform matrix. I guess I can calculate the offset from current CG to NRP and then transform to Local frame and sum with the LLA of the CG.
There probably needs to be a translation there, too. It would not really be noticable except in proximity formation flight or near the ground - but these are important visual cues and they need to be modeled correctly. Perhaps we should provide the pilot eyepoint location (LLA) as well?The pilot's view is always kept in sync with the 3D model's rotation (we actually start with the same matrices). Except as I noted above, when the user turns the pilots head additional rotations for that are applied.
Jon
_______________________________________________
Flightgear-devel mailing list
[EMAIL PROTECTED]
http://mail.flightgear.org/mailman/listinfo/flightgear-devel
