Alt-BEAM Archive
Message #10562
To: beam@sgiblab.sgi.com
From: "FCO Enr." fco@total.net
Date: Mon, 21 Feb 2000 00:05:01 -0500
Subject: [alt-beam] Re: BEAM: Tendency toward miniaturization]
Greetings,
Oh my poor choice of words may have made
my paragraph below was surely misunderstood. What
I meant was that given my electronic development skill
level (I mainly look at NvNu net's designed and optimized
by others) I wouldn't be able to spend time coming up with
more advanced NvNu ideas. I'd rather USE those existing
and very fascinating ideas. Not only to build a robot that
can walk on smoothe surfaces, but rather be able to survive
in a more realistic and hostile environment. It was about
setting goals in terms of the most efficient, fast and agile
and bullet proof microcore walker :) I'd like to see how
motors would fare in real weather, see microcore walkers
react to real life obstacles that threaten it's existence if
not handled correctly. Hence the beauty and simplicity
of using an analog NvNu net brain and sensor array interface.
Andy
fco@total.net
On 20/02/00, at 2:26 PM, Steven Dang wrote:
>To me, it seems that along the path that you have planned for yourself,
>you will encounter what you are seeking to avoid, walker neural nets and
>nervous nets. These nets are the basis of the walker's survival in=
addition
>to sensors. Without these the walker is just a paperweight. So although=
your
>goal is very valid, you will find that avoiding advanced control systems=
is
>impossible. That is without moving to microprocessors
>
>Steven
>>fco@total.net wrote:
>> challenges and meet them. So for me it isn't going to be the
>> walker microcore optimizing due to my lack of electronic
>> skills. But rather building a walker that can survive in that
>> environment for long periods of time.
Home