Best Juggling Robot Yet?

[n8rae]

Hi Again.

Just updating you with the latest juggling robot video. I think this is
the best "real arm" robot juggler so far though it seems to be only able
to track and throw balls in two dimensions so far...

Video Here: http://youtu.be/vnDwvsVk_nc

What do you think?

Nathan


--
----== posted via www.jugglingdb.com ==----

[Dave Barnes]

n8rae wrote:
>
> Hi Again.
>
> Just updating you with the latest juggling robot video. I think this is
> the best "real arm" robot juggler so far though it seems to be only able
> to track and throw balls in two dimensions so far...
>
> Video Here: http://youtu.be/vnDwvsVk_nc
>
> What do you think?
>
> Nathan
>

It looks amazing, but I couldn't help tutting at its technique!

Dave

[Jason Perry]

Dave Barnes wrote:

>
> n8rae wrote:
> >
> > Just updating you with the latest juggling robot video. I think this is
> > the best "real arm" robot juggler so far though it seems to be only able
> > to track and throw balls in two dimensions so far...
> >
> > Video Here: http://youtu.be/vnDwvsVk_nc
> >
> > What do you think?
>

> It looks amazing, but I couldn't help tutting at its technique!

What are you talking about? I didn't see it's feet move once ;-p

Jason.

[capricornwhite]

I think it's still some way off, there needs to be 2 of those bad boys and
i know you mentioned it being 2 dimensional but some shoulder movement
would clinch the deal to be able to throw them on a front plane.

You can see it's potential though

[pumpkineater23]

n8rae wrote:
>
> Hi Again.
>
> Just updating you with the latest juggling robot video. I think this is
> the best "real arm" robot juggler so far though it seems to be only able
> to track and throw balls in two dimensions so far...
>
> Video Here: http://youtu.be/vnDwvsVk_nc
>
> What do you think?
>
> Nathan
>
>

Thanks for the link, I do love watching the progress of juggling robots.
Is there a reason why robots juggle balls, not beanbags? Is it because a
ball is slightly more accurate to throw? Or something else?


pete

[Adam Rowney]

I imagine it is more accurate to throw. Perhaps harder to catch, has to be
more exact, but certainly there should be less give when throwing, so less
error.

I would have thought anyway.

[Varkor]

Most juggling patterns only require two dimensions - it's just using the wrong two :P
It's the most human-like juggling robot I've seen so far - thanks for sharing this!

[Peter Bone]

n8rae wrote:
>
> Hi Again.
>
> Just updating you with the latest juggling robot video. I think this is
> the best "real arm" robot juggler so far though it seems to be only able
> to track and throw balls in two dimensions so far...
>
> Video Here: http://youtu.be/vnDwvsVk_nc
>
> What do you think?
>
> Nathan

I disagree. I think this one is still the best.
http://dce.fel.cvut.cz/juggler/
http://www.youtube.com/watch?v=gIheAjvEiBE
It also uses visual feedback to correct errors and can juggle 5 balls
instead of just 2. I agree that the robot you linked to uses a more human
movement but I doubt that the human morphology is optimum for juggling. It
also appears to be more human because it's making larger errors. If we're
talking about pure juggling ability then the CTU robot wins easily.
Pete

[CamStradeski]

I have to agree with Peter here, while the 1st one may resemble a human
more it's nowhere near as accurate, and like Peter says if it's purely on
juggling skill the CTU one wins as it eliminates the variables of shoulder
and arm movement.

I guess it depends what the goal is, to replicate human-style juggling or
to create near perfect juggling patterns...

[lutkus]

CamStradeski wrote:
> I guess it depends what the goal is, to replicate human-style juggling or
> to create near perfect juggling patterns...

I'm still waiting for a robot I can pass clubs with.

[CamStradeski]

lutkus wrote:
>
> I'm still waiting for a robot I can pass clubs with.
>

Haha! That could still be a ways off, though in theory what the
"human-like" robot is doing wouldn't need to be adapted a huge amount for
clubs (I would assume since the throwing motion is quite similar, but I'm
no mechanics/robotics expert), send him an e-mail!

[n8rae]

lutkus wrote:
>
> CamStradeski wrote:
> > I guess it depends what the goal is, to replicate human-style juggling or
> > to create near perfect juggling patterns...
>
> I'm still waiting for a robot I can pass clubs with.
>
>
>

You need a system that fires clubs at you in the pattern and tempo that
you require like a tenis practice machine. You'd need a huge magazine of
clubs though...

Oh and my question was if this was the best "real arm" robot. I'm fully
aware of the servo juggling robot...

Nathan

[CamStradeski]

>
> Oh and my question was if this was the best "real arm" robot. I'm fully
> aware of the servo juggling robot...
>

Yar so you did...

The club passing would be hard because of the added variables of 2
participants (either 2 bots or 1 bot 1 hume), so there would need to be
some serious real time object tracking, also it's amplifying the
requirements on the z axis (If we assume that y is height and x is
lateral), which means it will need some artificial depth perception
(though I do remember those 2 hovering ones that were passing 1 ball back
and forth).

[lutkus]

n8rae wrote:
>
> lutkus wrote:
> >
> > CamStradeski wrote:
> > > I guess it depends what the goal is, to replicate human-style juggling or
> > > to create near perfect juggling patterns...
> >
> > I'm still waiting for a robot I can pass clubs with.
> >
> >
> >
>
> You need a system that fires clubs at you in the pattern and tempo that
> you require like a tenis practice machine. You'd need a huge magazine of
> clubs though...

No, that would not be acceptable to me. I realize that my human throwing
abilities may not be up to par here. So as a compromise, I'd like to see
two (presumably of the same design) juggling robots that can pass clubs
with each other. The important part would be that they are 100%
independently controlled. Presumably, they would need to communicate with
each other to synchronize -- if they have sufficiently developed optical
sensors, up-down-pass should be enough to work with.

Passing aside, just juggling clubs would be an incredibly impressive feat.
Particularly, juggling them in the sort of pattern a human would use (eg:
3 club cascade with single spins).

[Warren]

On Thursday, May 17, 2012 10:15:58 AM UTC-4, lutkus wrote:
> CamStradeski wrote:
> > I guess it depends what the goal is, to replicate human-style juggling or
> > to create near perfect juggling patterns...
>
> I'm still waiting for a robot I can pass clubs with.
>

I think Vova is in San Francisco! :-)

[Miika]

n8rae wrote:
> lutkus wrote:
> >
> > CamStradeski wrote:
> > > I guess it depends what the goal is, to replicate human-style juggling or
> > > to create near perfect juggling patterns...
> >
> > I'm still waiting for a robot I can pass clubs with.
> >
> You need a system that fires clubs at you in the pattern and tempo that
> you require like a tenis practice machine. You'd need a huge magazine of
> clubs though...
>
> Oh and my question was if this was the best "real arm" robot. I'm fully
> aware of the servo juggling robot...
>

Perhaps it's the best if you require the hand to grab the juggled ball? Like
this one
http://www.youtube.com/watch?v=6p9ZXueZeJc looks pretty good but doesn't grab
the balls in any way.

As for a robotic juggling partner, that would be cool! Maybe you could get
away with a club feeder that can be refilled by throwing clubs in a big
funnel. So you might use something like 20 clubs to feel like you're passing
six clubs, while the machine is holding most of them. And of course if the
clubs are too difficult, why not build one for balls to begin with?

-Miika

--

btw, tämä on ensimmäinen viestini tältä sivustolta

--
----== posted via www.jugglehub.com ==----