Origami Robot

May 12, 2005 3:05 PM

Robots that can fold origami has been an open research problem -- the holy grail of manipulator dexterity. This Carnegie Mellon Ph.D. candidate came up with a pretty elegant methodology.
posted by stacyhall1 (12 comments total)

I, for one, welcome our new paper-folding mechanical overlords.
posted by mystyk at 3:33 PM on May 12, 2005

I fail to see how this is much different than existing CAD/CAM technology, other than the fact that it uses paper as its base material. If you created a pair of robot hands that could fold origami, I guess I'd be more impressed.
posted by monju_bosatsu at 3:55 PM on May 12, 2005

It doens't move the paper around? Call me when there is no human intervention.

Not that I want you to call me about origami, that is.
posted by Four Flavors at 3:58 PM on May 12, 2005

Well, I guess that means that we'll start seeing assembly line manufactured paper cranes sold in stores, soon. whoopee!

[/luddite]
posted by shmegegge at 4:08 PM on May 12, 2005

It doens't move the paper around? Call me when there is no human intervention.

It does, it picks it up and rotates it in the movies.
posted by swordfishtrombones at 4:54 PM on May 12, 2005

Baby steps towards the Butlerian Jihad?

I mean, it's not like you should stop people who want to research such a project, and it's not like the problems and results aren't really cool. But there's something about this, something dehumanizing. It's different than when robots do work for people for some reason.

Or maybe I'm just having a very pessimistic day...
posted by ontic at 5:10 PM on May 12, 2005

Well, that was incredibly underwhelming. Perhaps of interest from an industrial automation point of view, but utterly primitive. Then again, simple an primitive but really-really good at repetition is exactly what industrial automation is all about I guess...

wake me up when they demonstrate multi-purpose mechnical manipulators that can do as many different things, or more, than the human hand.
posted by C.Batt at 9:04 PM on May 12, 2005

jeez, harshness!

Since the CMU RI is my school, I know (from taking Matt Mason's class, no less!) that this kind of manipulation problem is no walk in the park.

As I see it, Devon's work is cute for being origami, but it's actually a pretty useful exploration of some pretty tricky mathematics behind what seems like a very natural task for us. If I remember some of the issues correctly, there are some constraints you have to follow to keep the paper bending in a proper way, e.g. without flexing in between creases, and expressing these constraints is a challenge. This is to say nothing of the challenge of planning the folds to get from a flat sheet of paper to a finished piece, all of which must satisfy the constraints.

Even though the origami robot looks like a paper chopper, the research using it will likely lead to more inspiring manipulators in the future.

Devon got his Ph.D. last year; he's a prof at Dartmouth now.

And now for a rant: there are two things that are annoying about working in robotics in this decade:

1. You spend ages designing a system to explore some exotic locale. Mars. Antarctica. The Atacama desert. The sea beneath the Arctic ice caps. Late nights, sweat, tears, and blood (sometimes literally!) go into the machine. Then, the day of the test. You push the button and... stay put. The robot has all the fun, and can't send back enough data to make you really feel like you're there.

2. People watch movies and expect the freaking Terminator. I'm sorry, folks. We're just not there yet. Accept that making a versatile, independent, real-world agent is an extremely challenging task that will require the very best from chemistry, materials science, mechanical engineering, statistics, cognitive science, psychology, neuroscience, chemistry, computer engineering, &c. &c. &c.

wake me up when they demonstrate multi-purpose mechnical manipulators that can do as many different things, or more, than the human hand.

Spoken like someone who has never marveled at their own hand. Have you any idea just how many things it can do? But rest easy, it's being worked on.
posted by tss at 9:53 PM on May 12, 2005

Dunno why I mentioned chemistry twice, save to say that battery technology is IMHO the #1 thing holding back truly widespread application of robotics technology today.

Fortunately, effective robots will not require the very best from English composition.
posted by tss at 10:31 PM on May 12, 2005

cool post, stacyhall l. thanks.
posted by dhruva at 12:58 AM on May 13, 2005

Watching the movie where the robot folds a hat... at the end, when the guy picks up the hat and plays with it, I couldn't help cringing when the guy's arm looked like it was right under that machine part that creases paper. I guess I'm a safety freak for thinking that the guy's arm could be chopped off by a paper folder. (I also don't like those paper cutters with the chopping arm...)

Anyways, It'll be cool when the robot can also do stuff like this.

I also look forward to when robots can work with even more flexible stuff -- when clothes and shoes can be made all by robot.
posted by mhh5 at 1:00 AM on May 13, 2005

Man, I don't want my clothes and shoes to be made by robots, I want them to be washed and folded by robots. (OK, I don't actually want my shoes folded.)

Having built a few little robots and seen their limitations, this is pretty impressive.
posted by surlycat at 2:21 AM on May 13, 2005