Alt-BEAM Archive
Message #13917
To: alt-beam@yahoogroups.com
From: Your Friendly Hogfather hogfather@earthlink.net
Date: Tue, 11 Dec 2001 22:47:04 -0700
Subject: Truth or Dare
The proper subtitle for this message would be: "Just how hard do you have
to hit an obstacle before you've got to admit that you DID NOT avoid it?"
To be heretical about it, the phrase "obstacle avoidance" is literally
slathered all over BEAM literature. I've got a huge problem with
that: the sensing mechanism for pretty much all of this kind of
functionality is the "universal direction wire-and-spring" thing that
Tilden put together.
Look, folks, if the sensor isn't activated until the wire gets jostled by
some object (something that is not the robot) then there's been a hit ("a
palpable hit!" for you Shakespeare types). The object has NOT been
avoided, it has been hit. No avoidance has occurred. Yes, it has been
sensed, yes, the body of the robot -probably- won't hit the obstacle -- I'm
not arguing that half of the issue.
It is kind of like Bill Carlin's set of jokes about the airline industry
and the phrases they use. "Near-miss", he says, "is total bullsh*t A
near-miss is a hit! A near-hit is what they are really talking about and
in either case it scares me to h*ll". (sorry, that was paraphrased somewhat
and the bowdlerizing I've done here is exceptionally mild 'cuz the guy
swears up a storm).
These robots aren't avoiding obstacles. They are plowing right into them
with their touch sensors.
Oh well -- I guess it is a question of degree, however. If you were using
some kind of IR system instead, you'd have to take into account photon
pressure: after all, you are throwing out that IR energy: it HITS something
and you get a signal as a result: what can I say? Nearly the entire
industry would consider _that_ to be non-contact, but to the heretical
purist (oh dear, I appear to be one. sigh.) you don't get to observe
anything without modifying it. Hitting it, as it were.
This being your quarterly keep-alive rant,
YFH
Home