[Avery Andrews 940902.1311]
Today I came across the Aug 27 issue of New Scientist, which contains
a review of a book called 'out of control' by someone called Paul
Kelly, arguing that the classic 'top down' approach to AI control
of robots doesn't work, and proclaiming the virtues of an allegedly
biologically-inspired approach reminiscent of Brooks, etc.
The reviewer finds the general picture presented to be vague and
unsatisfactory, but says interestingly that the task of holding
a drink-glass straight & not spilling it while moving it around is
still beyond the bounds of what current robot-technology can do. So
maybe it would be worth somebody looking at the book to get specific
references to this work - like this is a problem we ought to be
able to solve in real life, as well as in computer simulation.
(I have been working away on nu14v2, but am a ways off from anything
I'd care to post).
Avery.Andrews@anu.edu.au