Alt-BEAM Archive
Message #03197
To: beam@corp.sgi.com
From: Jean auBois aubois@trail.com
Date: Tue, 11 May 1999 11:45:50 -0600
Subject: [alt-beam] Re: Neural Neurons (again)
At 10:04 AM 5/11/99 , George Rix wrote:
>Perhaps you forgot to write back, but I only got one reply about my nervous
>neurons question. If you'd like me to re-post the entire question, I'd be
>happy to, but for now this is what there is (a reminder, in other words).
Forgive me for seeming peevish, but really George -- haven't you seen the
multiple references in people's signatures about where you can get answers
to questions like this? In any case, Brian Bush's "B E A M - R o b
o t i c s - T e k" (http://people.ne.mediaone.net/bushbo/beam/nvnet.html)
is a good place to look. I suggest that you look in the "Nervous Networks"
section, particularly the topics:
Anatomy of a Nervous Neuron
Pulse Neutralization Circuits
A simple oscillator PNC to control an Nv network
[1.3] Sensors and Nv Networks
realizing, of course, that a PNC is an Nu neuron with its input connected
to the power supply and its output connected through a diode to some Nv
network. I'm of the opinion that the diode ought to be considered more a
part of the Nv neuron it is connected to because it strongly inhibits
pulses from "passing through" the neuron by holding its bias point high. I
think that the analogy of the PNC circuit "eating" processes is a somewhat
faulty.
I have no doubt that other sites also have answers as well.
Very generally (!), however, Nv neurons provide a time delay that, when
arranged in a loop, allows control of some device to be stepped in some
sort of sequential manner. Also, it only takes one short falling-edge at
the input of an Nv to start its operation. Nu neurons are often used to
'remember' things, like a hit on a touch sensor. They can take several
input transitions (if these pulses are short enough) to cause the output to
change state.
I've long wondered about a claim that Mark has made:
>Q: I have read up on Tilden's Nervous Net. It's a novel idea. But it
doesn't parallel nature as much as one might expect.
>
>A: (contributed by Tilden) ... In biology there are many
>autonomous pulse circuits called Central Pattern Generators (CPGs) that have
>similar Nv behavior, but only at the cost of complexity. It requires three Nu
>integrators to make one Nv neuron equivalent (two feed-forward excitatory,
one
>feedback inhibitory)...
Does anyone have a clue what sort of circuit he is hypothesizing?
---------------------------------------------------------------
jab
Do you have a BEAM question? Try one of the following sites -- they really
help!
BEAM Robotics Tek FAQ http://people.ne.mediaone.net/bushbo/beam/FAQ.html
BEAM Mailing List Archive http://www.egroups.com/list/alt-beam/
BEAM Online http://www.beam-online.com/
Chiu-Yuan's BEAM Page http://www.geocities.com/SouthBeach/6897/
BEAM Heretics http://www.serve.com/heretics/
------------------------------------------------------------------------
eGroup home: http://www.eGroups.com/group/alt-beam
http://www.eGroups.com
- Simplifying group communications
Home