Alt-BEAM Archive
Message #08847
To: beam@sgiblab.sgi.com
From: Phillip A. Ryals phillip@ryals.com
Date: Wed, 5 Jan 2000 21:31:47 -0500
Subject: RE: Biped Beam Bots
> Not to step on toes, but the idea of a tail isn't really new. The MIT
> leglab comes to mind...
>
> If I understood Nathaniel's explanation though, his use of the tail does
> sound original.
>
> Nathaniel, I was wondering... What exactly do you mean by tilt sensors?
> Are you talking about just normal switches, or some kind of analog device?
> I was curious because it seems like if you just use switches, the
immediate
> on/off/on would tend to throw the bot into overcorrection.
>
> That, and one other thing... At first I though you meant a 'tail' that
> swung around the bot, but then I pictured it holding the 'tail' up in the
> air. (like balancing a yardstick on your finger) Could you clarify?
>
> Cheers,
> phillip
>
>
> -----Original Message-----
> From: owner-beam@sgiblab.sgi.com [mailto:owner-beam@sgiblab.sgi.com]On
> Behalf Of Dane Gardner
> Sent: Tuesday, January 04, 2000 11:23 PM
> To: beam@sgiblab.sgi.com
> Subject: Re: Biped Beam Bots
>
>
> The entire notion of a tail is genius. How did you think that one up?
>
> CYA,
>
> Dane
>
>
>
> ----- Original Message -----
> From: "Nathaniel Martin"
> To:
> Sent: Tuesday, January 04, 2000 5:48 PM
> Subject: Re: Biped Beam Bots
>
>
> > Nice bots, Ian and Evaristo.
> >
> >
> > Here's my idea for a Biped BEAM 'Bot ("BBB"). I'll try to build it after
I
> > build a 2 motor 4-legged BEAM bot (A. Miller type.)
> >
> > Remember, I'm a newbie at this, so tell me if what I'm thinking is
> impossible:
> >
> >
> > Basically, I would keep walking and balancing separate.
> >
> > The robot would be "chicken legged" with the knees angled back. It would
> > have two servos on the hip, one servo on the knee, and an
> > electro-mechanical clutch on the ankle (I'll get to that later.)
> >
> > I would use a few microcores to get a basic walking rhythm:
> >
> > 1. Move left leg in air and rotate forwards (loosen ankle clutch)
> > 2. Straighten left leg (moves foot down) until pressure sensors on the
> left
> > foot read that they have touched something.
> > IF: the leg is completely straight and it hasn't touched anything
> > (like
> > with a cliff) then 1: Bend leg again
> > 2: Move leg left (at hip)
> > 3: go back to number 2
> > 3. Lock ankle clutch (the clutch allows the foot to rotate to work with
> > rough surfaces)
> > 4. Bend knee while rotating leg back and straighting right leg (moves
the
> > body of the BBB forward.)
> >
> > Repeat for right leg.
> >
> > Now this would walk, but it would fall down. Here's how it would
balance:
> >
> > I would have a "tail" to balance it. I put tail in quotes because it can
> be
> > in front of the bot or in back, so it isn't exactly like a real tail.
> >
> > The tail would be on top of the bot. Basically it would be a fairly long
> > rod (compared with the BBB) with a fairly heavy (compared to the BBB)
> > weight on the end.
> >
> > I would have 4 (or more) tilt switches on the BBB.
> >
> > I would have a very simple circuit: If a tilt switch registers a tilt,
it
> > moves the tail to the opposite direction of the tilt to reposition the
> > center of gravity.
> >
> > The tail would have 2 DOF so that it can go anywhere above the 'bot.
> >
> >
> > Did I explain this well?
> >
> > Do you guys think this would work?
> >
> >
> > -Nathaniel Martin
> > nat@lumiere.net
>
>
>
Home