tgCompoundRigidSensor does not output the orientation of a compound rigid body

[apsabelhaus]

At the moment, orientation sensing is not implemented. That's because this gets complicated: we need to be careful when taking transformations between two orientations.

One way to implement this would be to store the position of one of the compound's components at t=0, and set that position as "upward vertically" at the start: e.g. this position would correspond to a direction along the unit vector (0, 1, 0). In other words, the initial orientation of any compound rigid body is along the vertical Y axis (remembering that in NTRT, "y" is upward, what we in engineering usually call "z".)

Then, we'd take the difference between future orientations of that one specific compound against the t=0 orientation of that compound, and output that difference in orientation.

However, we need to make a decision about what this "difference" should mean. In Euler angles, for example, this might be completely meaningless: just subtracting them will not (I think?) give the transformation we want.

To-Do: research how to take this difference easily. Quaternions?
To-Do: decide what parameters we want to output to the log file. If we do Euler angles, will we get the log output jumping around?