Alt-BEAM Archive
Message #12693
To: beam@corp.sgi.com
From: Justin JAF60@student.canterbury.ac.nz
Date: Wed, 29 Mar 2000 23:45:41 +1200
Subject: [alt-beam] To those going to the Robot Games...
I've just been looking at the cool robot gallery on Solarbotics (again).
Many of the robots there are unique to places like Solarbotics - their
creators do not maintain websites or put information about them online,
the only place to see them is places like the Solarbotics gallery.
It sounds like the upcoming robot games will have a lot of this kind of
robot - bots that will only be seen at the games, or that minimal
info/pics will ever become available online. There are so many robots
that the online community has never seen (most of mine included).
With this in mind - a plea to those attending - if you possibly can,
take a camera (take a digital camera!), wander around and talk to other
BEAMers, ask to see their creations (it's not like they'll say "no" :-),
chances are that as well as the competition bots on display, they also
have many from their collection with them and are all too happy to have
people see them. Take photos of the robots. Take detail photos. Ask
about how they work, what they do. (Take some notes if warranted. Find
out if there is a nearby photocopier so you can get a copy of a
schematic if you desire).
Bring the unseen robots into the light! Put your findings online -
perhaps collaborate with other people doing the same thing. Do it for
your own good - compiling a treasure trove of your personal favourites -
all the robots that are so cool they inspire you to make bots, robots
that incorporate neat ideas that might come in handy (provided you
record them somewhere so as not to forget them), the things you aspire
too, the ones you aspire to beat at the next games, etc. Doing this
gives you a _huge_ resource to draw from that you really won't
appreciate until you've tried it. Try it - then dump it online. We'll
love you :-)
Hmmm, I think I shall refer to those robots that are unknown to the
community as "Robot Dark Matter". We know there is a lot of it out
there, but we're not sure what form it takes...
(And vast resources should be spent uncovering it :-)
Disclaimer - I've never been to a Robotgames, so if my guestimation of
how they work is obviously a long way from reality, this is why. Moving
to a related topic with which I have more experience:
If at all possible, come prepared - miniature photography in wildcard
conditions is not something you should expect to just be able to do and
have it come out properly. Practise at home beforehand if need be. Bring
the appropriate equipment if you can (eg sheets of stiff white card are
useful for heaps of things, especially to remove shadows. Perhaps a
tripod, as a narrow aperture gives greater depth of field but requires
longer exposure times. Try bring high-speed film if using film - needing
the flash (rather than being able to choose it) can bring ruin in many
circumstances. If you're extra keen, make a cardboard studio (eg white
box with 2 or 3 sides removes) that you can put bots in to get great
pictures of them. If you have the camera ready - sitting on the tripod
focused on a clean bright surface, with backdrop and adequate light,
people will _want_ you to photograph their robots and be most obliging.
A _lot_ of people don't have the time or equipment to get really nice
shots of their work, but virtually everyone would _love_ really nice
shots of their work. Damn I wish I was going myself... :-)
BTW, If anyone wants some more stuff on the photography thing, I can go
into greater detail, and/or provide some good links. You should note
however that I don't consider myself an expert :-)
12694 Wed, 29 Mar 2000 03:44:35 -0800 (PST) [alt-beam] Re: glider beam@sgiblab.sgi.com Evan Dudzik try this address, you can build your own bs2, cheaper!
http://www.phanderson.com/stamp/bs2_homebrew.html
--- SkavenArmy@cs.com wrote:
> i am goingto make a glider using a basic 2 stamp ,
> carrier board and 2 or 3
> servos ....... i want it to be fully automated so
> that when it hits a certain
> altitude it slams on full flaps and goes into a
> flatspin to the
> ground,landing genlty....also ... i want it to be
> able to be pre programmed
> to do certain things at certain altittudes.like
> maybe start a rocket
> engine(composite carbon rocket motors) etc
> etc.....simple i/o lines would
> work im sure , but the one reason i dont have it
> already, is that i cant
> afford the basic 2 stamp...if anyone has any info on
> maybe warping the
> "sacred" bicore or maybe even dissecting a basic 1
> or 2.......thankx
> jjay-z
> ohio
>
=====
+------------------------+
|http://surf.to/photovore|
|Photovores online! |
|Evandude Dudzik |
+------------------------+
__________________________________________________
Do You Yahoo!?
Talk to your friends online with Yahoo! Messenger.
http://im.yahoo.com
12695 Wed, 29 Mar 2000 03:57:08 PST [alt-beam] stop and start on their own? how? beam@sgiblab.sgi.com "Pat Wehren" Once these things stop walking, what starts them again? Does noise just
start feeding back or something?
______________________________________________________
12696 Thu, 30 Mar 2000 00:53:18 +1200 [alt-beam] Re: G.S.O.A.I.T. and possible RJP market. beam@corp.sgi.com Justin >>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.
I agree about a standardised course, but we can't use time as a useful
measure without scaling the course to fit the bot - compare a 20cm x
20cm photovore with a 2cm x 2cm one. The smaller one has (relatively)
ten times the distance to cover, and it must do this with less solar
energy per gram to work with. Thus using time alone might end up being
more a measure of size than brains. You could divide time by size, or
make a Body Mass Index style chart, but I think there will be a nicer
way.
Time should definitely be involved though.
>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....
I'm inclined to agree on the grounds that it is simple enough to be both
easily and accurate reproducible, and already has some establishing
"weight" behind it (not to mention, there will probably be photovores
actually _designed_ for this platform, thus their creators will like
having it as a standard measure :-).
I have some reservations about this course when looking at, say, a test
for a micromouse (maze navigator). These would be constructed along the
lines of a maze (duh) with various checkpoints. The more checkpoints the
mouse can find, the higher the score - and each checkpoint is checking a
particular ability. Eg, placing a checkpoint in a section of maze that
cannot be found by using the "follow the right-hand wall" method. Many
mice rely on this limited technique, and the course (test) can easily
rate their intelligence in this respect - if they're just
"wall-huggers", they cannot reach all the checkpoints, and not only will
the test results show inability - they will identify which aspect/s the
mouse is lacking in.
Applying this technique (clear-cut objectives that solidly demonstrate a
defined ability) to a photovore course, you could have for example a
"barrier" between bot and the optimal light-path, that consists of a
lowered section of floor (2cm lower than surrounding area). If the bot
wanders into this (ie would walk off a table and die if given the
chance), it incurs a penalty, and is then up-righted if needed, to
continue the course. (The "penalty" would probably just be that it never
makes it to the checkpoint on the "safe" path).
Obviously, a course that is a series of these specialised
ability-testing barriers would make a great test, but it would also be a
complex course to specify for accurate reproduction, and reproduction
itself would require so much work as to preclude it catching on at this
stage. So I'd be inclined to pass on the ideal for now, and instead try
to establish _any_ universal test (ie the Robotgames one). Once we have
a universal test, and people are using it, then start to think about
improvements to it.
Currently, I think the Robotgames course as an _intelligence_ test
leaves a lot to be desired, but in terms of practical reality, is our
best bet.
On the other hand, if an excellent test arena was designed that also
functioned as a very cool RJP (as any excellent test could with minimal
modification), then if one person could manufactured these RJP's at a
good price, and places like Solarbotics stocked them (or forwarded
orders), then I can imagine enough people wanting to buy an awesome
ready-made RJP that _also_ functioned as an intelligence test, that we
might just pull a coup that creates a uniform-quality intelligence test,
makes it universal and available at all robot conventions, and make it
cover all aspects of performance in depth.
That would be cool.
I would be tempted to buy a ready-made RJP myself... Or perhaps even a
kit of the pre-cut parts with Some Assembly Required if it was seriously
cool.
12697 Wed, 29 Mar 2000 08:42:45 EST [alt-beam] Re: G.S.O.A.I.T. beam@sgiblab.sgi.com Ld5253@aol.com Black lights are ultraviolet which is in the 400 nanometer range and below,
infrared is beyond the opposite end of the visible spectrum past 700
nanometers. So barring any anomalies in ir detectors or in the lamp
producing the UV, an ir detecter should not be sensitive to a black light.
12698 Wed, 29 Mar 2000 08:44:44 -0500 [alt-beam] SIMD1/LEDPump Problems beam@sgiblab.sgi.com "Dye, Roy" I am working with Wilf's SIMD1/LEDPump circuit and having a few problems. I
was under the impression that this circuit charges all day and blinks a LED
all night. However, my breadboarded circuit flashes the LED while it
charges and only lasts about a half hour in the dark (with about an hour
charge). I don't think it would make it all night.
To get the LED to flash longer in the dark, I decided to replace the blue
LED with a high efficiency red LED. Now, the LED stays on steady when the
cap voltage is above 2.4v.
I am using a high output solar cell 3766 (5.5V&30ma), a coin type 1F super
cap, and 1N4148 diodes for the SIMD1.
I am not sure if these problems are due to my substitutions, or, I have some
type of a wiring problem.
Has anyone else built this circuit?
Roy
12699 Wed, 29 Mar 2000 10:16:56 -0500 [alt-beam] Re: RC walker! (was : Feelers or Sensors) beam@sgiblab.sgi.com "Sathe Dilip" That's good news. However, the piezo elements can generate high
voltages (certainly higher than what a 555 can withstand) depending on
how much you flex them. The current however is very small. Its
probably the potential divider made of three (5K ohm?) resistors
(internal to the 555) that is loading the transducer and protecting the
555. The internal circuitry of 555 may be capable to withstand some
abuse but just as a good design precaution, I would suggest adding a
zener in parallel to the piezo. Choose the zener value at about the
supply voltage. This should provide adequate protection.
Dilip
------------------------------------------------
Andrew Hooper wrote:
>
> Hi All,
>
> Well im proud to announce that the first test WORKED :)
>
> I built a simple 555 circut, replaced the cap with the pezio transducer
> added a 1m resistor, soldered a wisker on to the pezio and used
> a logic probe to check pin 3 :)
>
> When the wisker was touched or flexed pin 3 would drop low then return
> to a high state :).
>
> Got even funkier when i droped the resistance to 5k and added a 1k
> between pin 8 & 7 on the 555, got a sort of scratchy sound when the
> wisker was flexed.
----------------------------------------------------------------------
The early bird may get the worm, but the second mouse gets the cheese.
----------------------------------------------------------------------
12700 Wed, 29 Mar 2000 08:21:02 -0700 [alt-beam] Re: Constructor beam@sgiblab.sgi.com Flaming Headphones Hi,
If you don't mind not being to load the other "bots" just disconnect once
the page has loaded. You can still construct new bots and do everything else.
Mike, AKA Flaming Headphones
>? i'd kinda like to be able to goof without being online! <
Home