Machines and Models

Samuel Saunders[950110:1845 EST]

Following up on Runkel and Abbots comments on robots- there was a news
item in Science in mid-December (I don't have the reference handy) about
use of a robot model of a female cricket locating a conspecific male by
auditory cues. The robot was built from lego, using wheeled
locomotion, with custom auditory sensors. The intent was to test the
notion that a simple system, with tuned hearing and comparison of inputs at
the two ears, could locate and navigate to a conspecific in an environment
with acoustic reflectors and other complexity. The robot was used instead
of a computer model because of the complexity of representing the acoustic
environment. According to the article, a number of features of the behavior
of the real female cricket were reproduced. This was a news item, so
technical details are missing, of course. As I recall there was no
reference to a publication. It struck me at the time that this would be
interesting from a PCT perspective, and Bruce and Phil jogged my memory.

It could also be interesting to set up a robot with a few intrinsic
references, and some sort of reorganization system, and a few sensory
inputs, and see what it takes to build a control system.

Sam