Alt-BEAM Archive

Message #08045



To: beam@sgiblab.sgi.com
From: "John A. deVries II" zozzles@lanl.gov
Date: Mon, 29 Nov 1999 00:11:30 -0700
Subject: [alt-beam] Consider looking at this site


http://www.robotbooks.com/


Zoz



8046 Mon, 29 Nov 1999 08:58:48 +0100 Re: photowalker video & Quadrapod/Hexapod II legs beam@sgiblab.sgi.com Richard Piotter [mailto:richfile@rconnect.com]
Nope. Not in this dinky little farm town. They don't even have
plexiglass! asked 3 places!

Richard Caudle wrote:
>
> Your local hardware store should have Lexan sheet in stock.
>
> Richard (the elder)
>
> ----- Original Message -----
> From: Richard Piotter
> To:
> Sent: Thursday, November 25, 1999 9:18 AM
> Subject: Re: photowalker video & Quadrapod/Hexapod II legs
>
> > Sounds interesting! Where can I order this material? I have to travel
> > out of town just to get plexiglass!!! I'm considering using a combo of
> > transparent material and the yellow sheets from Lynxmotion. I've worked
> > with Plexiglass and it is a little af a pain to mess with. If I can find
> > some, I'll have to try this Lexan material. Thank you!
> >
> > Of course, with the new saw, I might try to go with a few metal pieces,
> > if I can get the metal and the blades for it!
> >
> > Richard Caudle wrote:
> > >
> > > > > >If anyone knows anything about the Lynxmotion design, send it on
> over!
> > > > > >$200 is too much for me, and if I can do the same with less than
> $20 of
> > > > > >Plexiglass (CLEAR!!!), then I think it's work the extra work and
> > > designing.
> > >
> > > Don't use Plexiglass! It's too brittle and hard to work with. Lexan
on
> the
> > > other hand is a miracle substance. You can even bend it without
> heating
> > > it! Just leave the paper on it until after you've drilled and bent
it!
> > >
> > > Richard
> >
> > --
> >
> >
> > Richard Piotter
> > richfile@rconnect.com
> >
> > The Richfiles Robotics & TI web page:
> > http://richfiles.calc.org
> >
> > For the BEAM Robotics list:
> > BEAM Robotics Tek FAQ
> > http://people.ne.mediaone.net/bushbo/beam/FAQ.html
> >

--


Richard Piotter
richfile@rconnect.com

The Richfiles Robotics & TI web page:
http://richfiles.calc.org

For the BEAM Robotics list:
BEAM Robotics Tek FAQ
http://people.ne.mediaone.net/bushbo/beam/FAQ.html



8047 Mon, 29 Nov 1999 13:18:43 GMT [alt-beam] Different approach to ground feelers. Useful? beam@corp.sgi.com "J Fisher"
format="flowed"
content-transfer-encoding: quoted-printable
x-mime-autoconverted: from 8bit to quoted-printable by sgi.com id FAA07620

I was playing with a laser pointer, and wondering if they would be useful =

for BEAM, and had an idea. It appears feasible, but the exact difficulty of=
=

getting to work is vague, so I'm throwing it up for discussion. It's =

basically a plan for a simple, short-range light-based range-finder, =

suitable for detecting that the bot is about to walk off a cliff and that i=
t =

should back up instead... The concept is very simple (see 2k gif =

attachment), but implementing it might be deceptive. (It's a laser dot =

projected onto the ground at an angle such that the position of the dot =

relative to the bot hull indicates the height of the ground immediately =

ahead. It's a little bit like the two spotlights method used by the =

dambusters.

Anticipated cons:

* Laser diode shouldn't be on continuously (power waste) - the system shoul=
d =

pulse-check the terrain instead. (adds to circuit complexity)
* The dot can disappear from view (ie very high drop), thus a no-dot readin=
g =

should trigger a "pull-back!", when in some cases, it will simply be a =

highly light-absorbent ground surface.
* It may well need lenses on the photodiodes, depending on the distance to =

the ground (adds to complexity).
* Potential circuit complexity could be greater than anticipated.
* More than two photodiodes would be preferable.

Anticipated pros:

* A bicore seems potentially useful as a circuit to interpret the photodiod=
e =

data. (nice and easy start)
* The laser could be reflected off a small mirror on a feeler, thus =

mechanically using the one sensing circuit to detect both ground distance =

and feeler collision - both inputs would require the same response... (This=
=

would also allow for far greater feeler sensitivity than the simpler =

omni-switch, which might be a bad thing, but might not... :-)).
* No ground feelers to drag (or even get snagged) on the ground, thus net =

power gains (assuming a pulsed check).
* No moving parts, thus potentially greater durability in any environment =

that won't cover the lenses with dirt :-)
* A false reading that it is safe to proceed when it isn't, is quite hard t=
o =

achieve via local environmental conditions.

Any enthusiasm?

Seeya
Justin

(BTW, If anyone plans to reply to me rather than the list, send it to =

jaf60@student.canterbury.ac.nz as I won=92t get mail sent to the hotmail =

account that I mailed this message from)

______________________________________________________


Attachment: range.gif

Home