[Avery Andrews 920828.1209]
As a sort of comprehension check on my understanding of the arm demo
(old version), is there a reason why it was done the way it was
done, rather than the way I would have done it, having position
error-signals set reference-levels for joint-angle velocities (I don't
claim any advantages for my way--it's just what would occur to me
first). Is there any known relevant empirical evidence?
Also, I'm unclear on why there needs to be a computation for a
kinaesthetic vertical angle (A + B/2, where A is the vertical
shoulder angle and B is the elbow angle). Why couldn't the
vertical shoulder angle reference level be derived directly from the
visual system?
Avery.Andrews@anu.edu.au