Page 107 - Introduction to AI Robotics

90
3 Biological Foundations of the Reactive Paradigm
Stark and Bowyer represented sittability as a reasonably level and contin-
uous surface which is at least the size of a person’s butt and at about the
height of their knees. (Everything else like seat backs just serve to specify
the kind of chair.) Stark and Bowyer wrote a computer program which ac-
cepted CAD/CAM drawings from students who tried to come up with non-
intuitive things that could serve as chairs (like toilets, hanging basket chairs,
trash cans). The computer program was able to correctly identify sittable
surfaces that even the students missed.
It should be noted that Stark and Bowyer are hesitant to make claims about
what this says about Gibsonian perception. The computer vision algorithm
can be accused of some inference and interpretation (“that’s the seat, that’s
the right height”). But on the other hand, that level of inference and interpre-
tation is significantly different than that involved in trying to determine the
structure of the legs, etc. And the relationship between seat size and height
could be represented in a special neural net that could be released whenever
the robot or animal got tired and wanted to sit down. The robot would start
noticing that it could sit on a ledge or a big rock if a chair or bench wasn’t
around.

3.4.4 Neisser: Two perceptual systems
At this point, the idea of affordances should seem reasonable. A chair is a
chair because it affords sittability. But what happens when someone sits in
your chair? It would appear that humans have some mechanism for recog-
nizing specific instances of objects. Recognition definitely involves memory
(“my car is a blue Ford Explorer and I parked it in slot 56 this morning”).
Other tasks, like the kind of sleuthing Sherlock Holmes does, may require
inference and interpretation. (Imagine trying to duplicate Sherlock Holmes
in a computer. It’s quite different than mimicking a hungry baby arctic tern.)
So while affordances certainly are a powerful way of describing perception
in animals, it is clearly not the only way animals perceive. Neisser postulated
that there are two perceptual systems in the brain (and cites neurophysiolog-
ical data): 110

DIRECT PERCEPTION 1. direct perception. This is the “Gibsonian,” or ecological, track of the brain,
and consists of structures low in the brain which evolved earlier on and
accounts for affordances.

RECOGNITION 2. recognition. This is more recent perceptual track in the brain, which ties in
with the problem solving and other cognitive activities. This part accounts