Alt-BEAM Archive
Message #00167
To: beam@corp.sgi.com
From: Michael Kennan kennanm@efn.org
Date: Thu, 04 Feb 1999 12:27:17 -0800
Subject: [alt-beam] Re: 'Schmitt trigger'
>>>In my personal opinion, I think that BEAM as a 'new technology' is pure
hype.
>If not for the "hype" would you have heard about this "new technology"?
>Would you be on this list?
Yes, I would be, since my introduction to BEAM resulted from a research
project I did in college on simulating a biological central pattern
generator (CPG) using neural-equivalent electronics. Take a look at a
perceptron and try to think of a way to implement one in analog
electronics... you'll discover very quickly that the best way to implement
a threshold in a neuron is to use a Schmitt trigger with some sort of RC
circuit. I talked to a number of people about the idea, even built a chain
of 6 neurons that would cascade when the first was triggered (an open loop
microcore, it turns out) and mentioned in my paper that this type of
circuit had some promise in controlling legged bots, such as centipedes who
can't possibly be figuring out where to put all those legs every time they
want to take a step. Got my paper back with a note telling me to look up
Mark Tilden and BEAM, as I might 'find it interesting'.
So, to answer your question, yes, I would be here. But the hordes of
people who don't know which end of a battery is which wouldn't be. See the
semi-recent flood of letters regarding splitting the list into a pro list
and a newbie list for more info.
>>>BEAM is a tool that can teach us many things, but it should not
>>>become a fanatical movement or thought of as a revolution in
>>>robot design.
>Unfortunately, there are "fanatics" among us. It is a revolution
>tho...in that it has made people realize that a CPU is not nessary
>to controll certian functions. I would like to state for the record that
>I LIKE CPU's... Computers have played a big part in my life.
Also for the record, I seem to recall that Tilden himself states that the
primary _usefulness_ of the microcore will be a result of the integration
of this type of circuit with conventional circuits. Even in biological
critters, you can make a distinction between the 'processing' core and the
CPG. Almost all insects and swimming creatures have some sort of CPG that
controls their primary locomotion in addition to a more complex 'brain'
whose job, among other things, is to influence the CPG when needed. The
microcore addresses a need for a locomotion controller that doesn't require
all of the clock cycles of a computer to work... so now with that out of
the way, we can work on getting the bot to *do* something.
As for the patenting issue, I was a little shocked at first: these
circuits are in electronics books going back 30 years plus. But they had
never been applied to robotic locomotion, and in that lies the essence of
Tilden's patent.
I also know a fellow who patented a shape. (I'm not kidding; one of the
projects I did in college was to derive the mathematical description of his
shape.) My first thought was 'surely the fact that you can patent a shape
must be a sign of the end of our so-called civilization', but my second
thought was 'so what, he patented a shape, people will just use other
shapes when they find out that the shape of their product is covered by his
patent'... but it turns out there are plenty of people who are licensing
the shape from him, including a perfume company who uses it for their
bottles, an arts society who uses it for their awards, and a museum who
built a hallway based on the shape.
What strange times we live in.
------------------------------------------------------------------------
eGroup home: http://www.eGroups.com/list/alt-beam
Free Web-based e-mail groups by eGroups.com
Home