Alt-BEAM Archive
Message #10472
To: beam@sgiblab.sgi.com
From: JVernonM@aol.com
Date: Sat, 19 Feb 2000 14:46:26 EST
Subject: [alt-beam] Re: reconfigurable Robots: Severing the master/slave bicore
In a message dated 2/19/00 11:37:16 AM Eastern Standard Time,
dibst11+@pitt.edu writes:
> I don't know about that. I think we can manage emulating insects. The basic
> Mobile robot's work to emulate insects through subsumption architecture,
yet
> with BEAM subsumption architecture tends to be built into the design, as
> BEAM reacts immediatly to changes in stimuli beyond that, it also acts in
> accordance to the degree of change. Very Subsumption like. But if you look
> at some of Mark T's solar bot's they easilly act like insects. Sure insects
> may in actuality be much more complicated, but Mark T's bot are capable of
> surviving and seeking out new food sources. Besides reproduction the do
seem
> to fufill the requirements of life. They seek out a food source, they
change
> their environemnt, and the excrete, (heat).
I don't think so, not yet anyway. We can make them LOOK like insects and we
have pretty much mastered getting them to find a light source, but that's
about it. Compared to the humblest insect, they pale by comparison. On the
survival issue. Again, I ain't so sure. If you build an environment
specifically suited to the bot, with no cracks and minimal obstacles, you can
get some semblance of "survival". Fact is, that is it. If you throw them into
an even slightly more complex system, like a living room rug, you quickly see
just how unable to survive that they are. One of the problems I see in the
BEAM community is claiming more than is actually there. Work on the level Bob
was talking about, namely a one celled animal complexity. There you may
actually be able to emulate a reasonable facsimile of actual life forms. The
Turbot was an excellent example.
I've run head on into this problem with Aquabots. It is very hard to get
reasonable behaviors out of a bot that lives under the surface of the water.
What does the bot do to survive? If it is solar powered, it needs the sun to
feed. But if you tune the bot to seek the brightest light, it doesn't want to
dive. If you configure it for negative phototropism to encourage diving, it
can't feed. Sure, you can make the thing dive until it runs out of juice, and
then float back up to feed, but that isn't emulating a lifeform, it's simply
a system that keeps the bot moving. And that is often mistaken for emulating
life. Many options have been offered to bias the core, thermisters, depth
sensors, temperature sensors, etc. But, these are difficult to wire into
present BEAM tech. The easiest, and therefore most widely used response, is
to light. I think Bicores are not enough. A rider will be needed to break
BEAM past the stage of singular sensing. Infra red, sonar, sound, taste,
smell. These can be accomplished through a horse and rider system. When you
build that, you will be closer to the lowest insect on the chart.
See ya,
Jim
http://www.geocities.com/SoHo/Exhibit/8281/beamart.html
ICQ# 55657870
10473 Sat, 19 Feb 2000 14:52:15 EST [alt-beam] Re: BEAM: Tendency toward miniaturization beam@sgiblab.sgi.com JVernonM@aol.com In a message dated 2/19/00 2:18:18 PM Eastern Standard Time,
dibst11+@pitt.edu writes:
> I built a two motor one, which did work, but it
> was aggrivating as hell.
That is the nature of the two motor. Perhaps if you built that upscaled
walker on a 5 or more motor platform, you would get better results.
See ya,
Jim
http://www.geocities.com/SoHo/Exhibit/8281/beamart.html
ICQ# 55657870
10474 Sat, 19 Feb 2000 14:58:53 EST [alt-beam] Re: FYI: ANN: fire-fighting robot contest at Penn State (3-25) beam@sgiblab.sgi.com CIRCITZ@aol.com Is there an official website with the rules and regs that you could post for
us to look at. This sounds like it could be interesting!
Dan
<< I'm only an hour or so away from that competition. I'm seriously thinking
of
creating a BEAM and CPU TEAM of robots to conquer the course. I read in the
info page that a team of robots is allowed this year. So I was thinking,
why not use 4 smaller bots to jet off really fast to the 4 rooms in the
course. The only function they would have is to find the candle, emit an RF
beacon, that triggers the bigger CPU bot to respond and put out the fire.
>>
10475 Sat, 19 Feb 2000 15:17:30 -0500 (EST) [alt-beam] Re: reconfigurable Robots: Severing the master/slave bicore connection. beam@sgiblab.sgi.com jester96beam@iname.com What about Mark T's Spyder? I've never seen it move, but I'm sure it would be able to navigate quite complex terrain. I think the problem is that we're making the bots too small to be able to find their way through anything more than a smooth surface, and also most people are making 2 or 3 motor designs which really don't display much of the behavior that more complex walkers do.
That Spyder looks as though it would be able to get around most environments in the average house hold, and maybe even the front yard.
I could be wrong - I've only read about the Spyder and seen a few pics. 5 motor designs are also much more capable than simple 2 motor walkers.
Richard what ever happened with your 8 motor walker?
Now photovores and rollers are a diferent story, but they're only designed for smooth table tops etc. The only "survival" they show is that they find there own food in an environment built specifically for them. We may as well sit them beneith a 1000 watt bulb in a 1 x 1 foot RJP!
I don't think BEAM has really evolved that much recently. Everyone seems to be building the same robots and most are 2 motor walkers (or rollers), of which we really can't expect too much other than for "proof of concept".
This is all just my oppinion - please don't flame me TOO bad (Though I'm interested in any arguments, because I hope I'm wrong about the evolution part)
Chris
---------------------------------------------------
Get free personalized email at http://www.iname.com
10476 Sat, 19 Feb 2000 12:51:35 -0800 [alt-beam] Re: Microcore controller for Bicore aps... alt-beam@eGroups.com "Droidmakr" stick-@aol.com wrote:
> Is it possible to have multiple bicores regulated by say a four
neuron
> microcore? Is it useful to do such a thing? Would the microcore still
need > an auto PNC circuit, because the bicores can regulate
themselves? Has anyone
> out there tried this?
>
> stick
Check this out:
www.geocities.com/droidmakr
Its an idea I had that runs along the same vain as your question.
Let me know what you think.
Home