Alt-BEAM Archive

Message #08883



To: beam@sgiblab.sgi.com
From: JVernonM@aol.com
Date: Fri, 7 Jan 2000 01:31:43 EST
Subject: [alt-beam] Re: SIMPLIFIED D1 CIRCUIT


Thanks Wilf! I'll give it a spin.

See ya,
Jim
http://www.geocities.com/SoHo/Exhibit/8281/beamart.html
ICQ# 55657870



8884 Thu, 06 Jan 2000 23:13:57 -0700 [alt-beam] Re: Bi ped beam@sgiblab.sgi.com Senior Wow that is amazingly wonderful, helpful, and well written. Good job! I
love the quote "the higher up the mass is, the less you have to lean
over to the side in order to pick up a foot." You've enlightened me. I
just need some cash now...

-Kyle

Bruce Robinson wrote:
>
> Well, since everyone else is talking about it, I may as well get in my 2
> cents worth. This is a subject I actually know something about (unlike
> my robotics and electronic skills).
>
> First, human leg joints, from top to bottom. Begin with the hip joints;
> each has 3 DOF (ball joints) separated by a relatively wide pelvis. That
> pelvis is very important. Next the knee joints, 1 DOF (hinge joint).
> Skiping down to the ankle joint, 2 DOF (double hinge joint). Then the
> feet, 5 flexible joints inside, not used for motion, but makes a
> wonderful triple arch suspension system. Finally, the toes, not to be
> underestimated for smooth walking. You don't need all 14 toe joints to
> make a walker (humans don't use most of them anyway); you can
> approximate the toes with a 1 DOF hinge joint, simulating the toes with
> a stiff but flexible material.
>
> OK, now you all want to object about my description of the ankle joint
> :) You sit there, twisting your foot from side to side. The knee's not
> twisting (I hope), so there's got to be a third degree of freedom down
> there. The answer: as you twist your foot, look down at those two bumps
> that stick out from either side of your ankle: those are the ends of the
> two bones in your lower LEG ... they aren't part of the foot. So where's
> that extra degree of freedom? Can't be in the ankle, or those bony
> protrusions wouldn't twist like that. Remember, TWO bones in the lower
> leg, side by side.
>
> so, does it matter anyway, exactly where between the knee and the foot
> that rotating motion comes from? Read on.
>
> If you look at human musculature, you'll discover that the muscles that
> control a walking limb are always as high up as they can be. Thigh
> muscles are near the hips. Lower leg muscles are just below the knee.
> Most of the muscles that produce large foot motion are also just below
> the knee, connected by a network of tendons to the foot joint.
>
> There's a reason for this. If you keep all the mass low down, when you
> walk, you'll have to accelerate it to get your limb moving, and stop it
> when you want to reverse the limb direction. That takes a lot of energy,
> and will tend to upset your balance. To complicate things, when the feet
> move, they accelerate to twice the speed the upper body is moving (think
> about it).
>
> With the weight up high on the limbs, there's a lot less force involved
> to move the legs. Try swinging a baseball bat or golf club at your
> favourite robot; stop the swing just before you hit the 'bot (oops,
> sorry about that). Now grip the bat or club at the wrong end and swing
> it. You can stop it a lot faster.
>
> That's why it's important in a robot where you put the servos. The lower
> they are, the more unbalancing momentum you'll have to deal with.
> Furthermore ...
>
> Humans are top heavy -- we ought to fall over, but we don't (because we
> have really great muscle control). Now look at Evaristo's biped -- very
> top heavy with that lead-acid battery up there. Look at Ian's
> description of his walker motion -- top leans to one side, shifting the
> center of gravity over one foot. The thing is, you can't pick up one
> foot if most of the mass is down near the ground -- you need enough mass
> up high to counterbalance the foot you're lifting. More to the point,
> the higher up the mass is, the less you have to lean over to the side in
> order to pick up a foot.
>
> So the rules for effective biped design are: separate the hips, lots of
> joints (but not to begin with ... learn how to c-ordinate the joints
> first), keep the mass in each limb high, keep the non-moving mass
> (batteries, controller, etc) up at the top, and cushion the foot
> slightly.
>
> -----------
>
> Here's a couple more tidbits.
>
> You don't need an motion in the upper body to walk. Just a solid lump
> from the hips up. I spend a good deal of time teaching people to move
> this way: torso erect, hands on hips, no leaning or swaying to the side,
> no up and down motion. Fast & very efficient.
>
> You don't need arms or a tail to get up off the floor. You can get up
> with your arms crossed on your chest, or held out in front of you. I
> teach people to do that, as well. It IS pretty hard to stand up without
> using your arms if you don't have flexible toes on your foot.
>
> -----------
>
> The great thing about using the human body to study pibed motion is,
> there's always one handy to examine.
>
> Regards,
> Bruce



8885 Fri, 7 Jan 2000 03:48:55 -0800 (PST) [alt-beam] Re: Biped Beam Bots beam@sgiblab.sgi.com Evan Dudzik aaah, I did it again, forgetting to say outright that
I hadn't built it yet... I'd love to though, just
need to buy servos! I will try to make a drawing of
it tonight.

--- Sparky wrote:
> Can we see a picture?
> sounds kewl.
> BTW,
> i think Ian's assembly is one of the coolest things
> i've seen lately.
> Now I want one.
>
> -Sparky
>
> At 12:04 PM 1/6/00 -0800, you wrote:
> >well, i have a 5-motored 2 legged bot, in which one
> >motor would be like a knee, only facing backwards,
> and
> >the other would be a hip... like those big two
> legged
> >walker things in star wars (remember the ones that
> >hunted them in the woods???)... thats only four
> motors
> >(servos)... the fifth??? it goes at the top, and
> faces
> >straight up. affixed to the horn is one of those
> 4X
> >AA battery holders, in which two rows of 2 AA's are
> >side by side, so it is long... this way, no weight
> is
> >added to shift the center of gravity left and right
> >save that of the servo... no extra weight to lug
> >around...and the battery pack could be shifted
> closer
> >to being centered over the servo... if centered,
> when
> >rotated, it wouldnt shift the center of gravity
> left
> >or right... but if all the way out as you could put
> >it, it would practically tip the bot over. This
> would
> >make it easy to fine-tune the shifting of weight.
> the
> >feet would be little tripods of coathanger wire,
> with
> >rubber nubs on the end of each foot. The feet
> would
> >be wide enough so that they ALMOST touch in the
> middle
> >of the bot, so as to make as little weight as
> possible
> >necessary to shift... I might do that LATER when i
> >actually HAVE 5 servos...
> >
> >--- Dane Gardner wrote:
> >> > When you place the tail between the legs then
> it
> >> will be impossible to let
> >> > the walker stand on one leg because the center
> of
> >> gravity will remain
> >> > between the legs.
> >>
> >> I didn't mean to leave it there. I meant for it
> to
> >> be able to swing outward
> >> (either forward or behind) from in between the
> legs.
> >> If you leave enough
> >> room...and don't want a head...you could have the
> >> tail swing completely
> >> around the center of the body, in a vertical
> fasion.
> >>
> >> > Suppose the tail is behind the legs so it can
> >> swing
> >> > freely to the left or right. No then the whole
> >> center of gravity would be
> >> > so low that it will take a lot of efforts to
> get
> >> the robot stand on one
> >> > leg. The whole issue is center of gravity. The
> >> higher it is the easier it
> >> > is to shift it.
> >>
> >> Yes but a top heavy two leged bot would never be
> >> able to just sit still. I
> >> would have to be able to move constantly. And
> >> unless you build something in
> >> that allows the bot to get up from a fall, you
> would
> >> be condeming the bot to
> >> a lack of sleep.
> >>
> >> > Besides isn't a tail designed to prevent an
> >> animal/robot
> >> > from falling when the body bends forward
> towards
> >> the ground?
> >>
> >> Not only that. It helps it to get up without the
> >> help of arms and hands.
> >>
> >> > A can remember
> >> > a walker which had a rod mounted on the waist.
> The
> >> rod pointed upwards and
> >> > at the end of it there was a weight attached.
> By
> >> moving the rod the robot
> >> > could control the center of gravity. I have
> seen a
> >> video of this robot
> >> > walking and you could see very clearly how the
> rod
> >> controlled the balance.
> >>
> >> I'm just saying that unless you want a HUGE bot
> with
> >> a lot of body mass,
> >> you're going to end up with a really top heavy
> >> system.
> >>
> >> CYA,
> >>
> >> Dane
> >>
> >>
> >>
> >
> >
> >=====
> >+------------------------+
> >|http://surf.to/photovore|
> >|Photovores online! |
> >|Evandude Dudzik |
> >+------------------------+
> >__________________________________________________
> >Do You Yahoo!?
> >Talk to your friends online with Yahoo! Messenger.
> >http://im.yahoo.com
> >
> >
>
>


=====
+------------------------+
|http://surf.to/photovore|
|Photovores online! |
|Evandude Dudzik |
+------------------------+
__________________________________________________
Do You Yahoo!?
Talk to your friends online with Yahoo! Messenger.
http://im.yahoo.com



8886 Fri, 07 Jan 2000 14:01:16 +0100 [alt-beam] Re: Pressure Sensors beam@sgiblab.sgi.com Evaristo Westplate Hi,

I used foam when I started with the biped and it works great except after a
while the foam gets compressed due to the weight of the robot. We had to
give the robot a break and get it of its feet in order to let the foam
restore itself. We also looked at piezo sensors but they only give a
voltage when the pressure changes. In other words they only give you a
relative indication. You will need a maximum voltage detector in order to
get an idea how big the force was which generated that voltage. Although
simple to build it will require some electronics and isn't suited for BEAM
(I think).

If you can then go for the Force Sensor Resistances or similar device which
will give a continues signal about the pressure. It will save you a lot of
frustration in the end.

Enjoy your weekend



Evaristo

Gizmo homepage: http://www.crosswinds.net/~evaristo
Biped photos : http://www.crosswinds.net/~evaristo/biped.html

Home