Alt-BEAM Archive

Message #12682



To: beam@sgiblab.sgi.com
From: SkavenArmy@cs.com
Date: Wed, 29 Mar 2000 01:48:41 EST
Subject: [alt-beam] Re: Portable AV monitor for Quadrapod


Electronics goldmine
has a nice sturdy dtmf touchtone keypad for 4.95 each and some nice speakers
and amps/xovers are available from allelectronics.com



12683 Tue, 28 Mar 2000 22:51:35 -0800 [alt-beam] Re: G.S.O.A.I.T. beam@sgiblab.sgi.com Bruce Robinson BUDSCOTT@aol.com wrote:
>
> ... i finally came up with a descent test, the Goal Seeking
> Obstacle Avoidance Intelligence Test (GSOAIT) ...

> Any and ALL comments are greatly appreciated, i kinda want to set
> a standard way to judge general intelligences of phototrophic and
> photophobic bots!

OK, Spencer, some comments.

- photo = "light" (from the Greek, phos or photos: light).
- tropism = "the tendency of an animal or plant to turn or
move in response to a stimulus" (from the Greek,
trope: a turning)
- phobe = "fear of" (from the Greek, phobos: panic or fear)
- phile = "lover of" (from the Greek, philos: loving)

So phototropic means "a tendency to turn in response to light", and
includes both:
photophobia, "fear of light", and
photophilia, "love of light".

You might think this is picky, but if we are to develop a lexicon for
BEAM, let's make it a proper one. Otherwise, why not just call them
"light-seeking" and "light-avoiding"?

Anyway, enough of that. The idea is a useful one, although there will no
doubt be a lot of argument about the details. As your page says, "Heavy
Experimentation".

What you are proposing is the first step of what's know as an
"operational definition". Instead of just giving some measurements and a
formula, you are also including a description of HOW the measurements
are to be taken.

So instead of just ultimately saying, "your robot's score is ...", you
will be saying "your robot's score is ..., and this is how we measure
it." To be generally useful, I think you are going to have to be more
specific about the shape and type of your maze ... you can't just leave
it up to individual testors. Size would be variable, but proportions
would probably remain constant.

You may find it useful to test one robot several times each in several
different maze configurations to see how much variation you can
introduce without affecting the scores.

Whatever scheme you end up with, some people will say "it isn't correct"
, or "it isn't fair." The goal of an operational definition is not to be
correct or fair, it is to be useful.

A good start, I think (once you polish up your Greek :)

Bruce



12684 Wed, 29 Mar 2000 02:03:12 EST [alt-beam] Re: G.S.O.A.I.T. beam@sgiblab.sgi.com SkavenArmy@cs.com i agree.....does a blacklight interfere with ir led sensory>??



12685 Wed, 29 Mar 2000 01:38:38 -0600 [alt-beam] Re: Portable AV monitor for Quadrapod beam@sgiblab.sgi.com Richard Piotter You know if it's crystal controled (a quartz crystal oscilator sets a
base frequnecy)?

Also, is it the 12 or 16 key version. I need to find a couple more of
the military key pads like the ones we have in the telcom lab at the
school. More keys mean more commands. I also need to build a crystal
controled DTMF decoder as well. It's then possible to simply double or
triple the crystal you have installed to increase the command set. All
the tones are completely distinguishable then. (Who cares if you change
the frequency, since you don't need to call Ma Bell anyway! :)

I'll have to test the range of my video transmitter tommorow, and see if
it'll die when in the elevator in the apartment. I'm betting it will,
but it will probably reach ground floor at least.


SkavenArmy@cs.com wrote:
>
> Electronics goldmine
> has a nice sturdy dtmf touchtone keypad for 4.95 each and some nice speakers
> and amps/xovers are available from allelectronics.com

--


Richard Piotter The Richfiles Robotics & TI web page:
richfile@rconnect.com http://richfiles.calc.org

-- Make Money by Simply Surfing the Net or responding to E-Mail!!!
-- Click below!!!

http://www.alladvantage.com/go.asp?refid=ATL147
http://www.spedia.net/cgi-bin/dir/tz.cgi?run=show_svc&fl=8&vid=329630



12686 Wed, 29 Mar 2000 14:20:25 +1000 (EST) [alt-beam] Re: G.S.O.A.I.T. beam@sgiblab.sgi.com Ben Hitchcock Actually I was thinking about this today as well, in response to the new
FRED I built a couple of days ago. We really need to quantify a standard
test to perform and a way to measure the results.

The reason that I was thinking about this is that in reasonable light the
FRED circuit doesn't just head for the brightest side the way a 1381
photopopper does. Because the FLED's aren't phase-coupled, a certain
degree of randomness creeps into the FREDs behaviour. If the side that is
lighter manages to flash just before the darker side, then the lighter
side might win and so FRED turns away from the light. This isn't
necessarily a bad thing. Over time this effect is squashed by the times
that the darker side flashes first, and so as a whole the bot is
phototropic. What this effectively does is widen the 'standard deviation'
of the window where either side could fire. The mean value is still
straight ahead but it is as if there is some noise being added to the
'heading' signal.
In English this means that FRED will wander round a bit more than a
standard photopopper, but the standard popper will be a lot more
'single-minded' than FRED. To demonstrate this I attached a pencil lead
to the capacitor and left FRED on top of a clean sheet of paper on my desk
for half an hour. When I came back FRED was happily popping away in the
center of the sheet - and the pattern on the sheet was a weird 'daisy'
with the center point directly underneath where the bulb was. A couple of
times two lines would start off with the same heading and follow each
other closely, then diverge radically. As a whole the picture showed that
FRED never strayed more than about 15 cm from the bulb. I imagine that a
'normal' photopopper would make a beeline for the light, then draw a
circle in the middle and stay on that circle.

Maybe we could make a 'standard course' with a 100 W bulb at one end, and
a round obstacle in the middle. The robot would be released at the other
end of the course. The time to reach the bulb would be the measure of how
'intelligent' the robot is.

What have you come up with? I checked your page but couldn't find
anything there.

Ben

> Well, it's bugged me for some time that somebody hasn't develouped a good
> 'scoring system' for testing phototrophic intelligence, SO, i finally came
> up with a descent test, the Goal Seeking Obstacle Avoidance Intelligence Test
> (GSOAIT), it's still under construction, but i thought it would be
> interesting to see what kind of response i'd get from its early stage, its
> under tutorials and then intelligence test! Any and ALL comments are greatly
> appreciated, i kinda want to set a standard way to judge general
> intelligences of phototrophic and photophobic bots!
>
> -Spencer
>
> <http://www.botic.com/users/beamstop>
>
> not a robot scientist
> not a college major
> not a grad student
> not a professor
> not a very organized person
> just Spencer (isn't that impressive enough?)


--
Nothing is foolproof to a sufficiently talented fool.



12687 Wed, 29 Mar 2000 03:21:41 EST [alt-beam] Re: Portable AV monitor for Quadrapod beam@sgiblab.sgi.com SkavenArmy@cs.com i am not sure...its in the catalogue..ill look if i can....any words of
advice for a newbie??



12688 Wed, 29 Mar 2000 03:21:16 EST [alt-beam] Re: Portable AV monitor for Quadrapod beam@sgiblab.sgi.com SkavenArmy@cs.com



12689 Wed, 29 Mar 2000 01:43:35 -0700 [alt-beam] Re: G.S.O.A.I.T. beam@sgiblab.sgi.com Dave Hrynkiw At 09:20 PM 3/28/00 , Ben Hitchcock wrote:
>Maybe we could make a 'standard course' with a 100 W bulb at one end, and
>a round obstacle in the middle. The robot would be released at the other
>end of the course. The time to reach the bulb would be the measure of how
>'intelligent' the robot is.

May I suggest the new Photovore rules as listed on the games website
http://www.robotgames.com? I've built two of the platforms, and it works
well for 1 on 1 competition, but also well for just a lonely Photopopper
just to truck around on....

Just a quick thought,
Dave
---------------------------------------------------------------
"Um, no - that's H,R,Y,N,K,I,W. No, not K,I,U,U, K,I,_W_. Yes,
that's right. Yes, I know it looks like "HOCKYRINK." Yup, only
2 vowels. Pronounciation? _SMITH_".
http://www.solarbotics.com



12690 Wednesday, March 29, 2000 11:17 AM RC walker! (was : Feelers or Sensors) beam@sgiblab.sgi.com BUDSCOTT@aol.com
>In a message dated 3/28/00 3:15:30 PM Central Standard Time,
>andrew@best.net.nz writes:
>
><< would it not be possible to use in the RC arangement in the
> bicore circut?. >>
>
>Don't ask me how this made me think of this, but wouldn't that be cool to
>stick in some remotely variable resistors and make an RC walker! that be
>cool, cause you could stop it on command, reverse it, turn it, and even use
>it to pass the salt across the table (we've all done that haven't we, until
>it steps in the gravy bowl and somebody gets pissed!). Anyway just a
though!
>
> -Spencer
>
><http://www.botic.com/users/beamstop>
>
>not a robot scientist
>not a college major
>not a grad student
>not a professor
>not a very organized person
>just Spencer (isn't that impressive enough?)

Home