Placeholder Image

Subtitles section Play video

  • So, imagine you had a bag of programmable matter,

  • by which I mean, engineered modules that cooperate with

  • each other to form shapes.

  • You could reach in and pull out some object you needed,

  • like a wrench or a coffee cup,

  • And then when you were done with it, you could put it

  • back into the bag, and it would break apart and its modules

  • would be available to form the next object that you needed.

  • So, biology provides the inspiration that this is, in fact, possible.

  • This is a simulation by my colleague Jonathan Bachrach,

  • showing a mechanical protein folding itself into several objects.

  • This is a chain of engineered modules, and each module

  • has a motor that can exert a force on its joint.

  • By setting the angle on each joint, it's possible to fold the representation of any shape.

  • Although, in practice I think you'd want to have lots of these chains

  • working together, just like biology does it.

  • We had a grant from DARPA to build a prototype of this,

  • and the grant had a requirement that the modules be smaller

  • than one cubic centimeter. And, because we wanted to

  • make a lot of these modules, we wanted the modules to be

  • cheap and simple, and so we decided to not use gearing,

  • just to have the motor directly drive each axis.

  • We looked, but we couldn't find any off-the-shelf motors

  • that were small enough, and could exert enough continuous force,

  • without burning themselves out.

  • So we invented a new type of motor, which we call

  • an electropermanent motor.

  • This motor works by using coils to remagnetize its permanent magnets

  • all the way around their hysteresis loops on every step.

  • What this means is, when you remove power, the device holds its position,

  • and it can exert its maximum force with very low average power input.

  • Once we had the motor working, we designed the rest of

  • the mechanical protein. It's basically a chain of motor rotors and stators,

  • the rotating part and the stationary partsinterlocked together,

  • with a flexible circuit wrapped around it for power and control.

  • We had to learn watchmaking techniques to build the prototypes, which was fun.

  • So, here's a video, probably taken at about 2am, of the

  • first module turning in a pair of scissors on my desk.

  • You'll see the coiled portion of the flexible circuit coiling and uncoiling

  • to connect the module to its neighbor; it took us a while to get that right.

  • And then once we had one module working we made four more;

  • here is a chain forming shapes.  It starts out as a line, and then it forms:

  • a left handed helix

  • a right handed helix

  • a periscope

  • and an L-shape.

  • Every module gets its instructions to turn left, right, or straight;

  • that's like the DNA code for the shape,

  • and then the motors fold the chain up into the shape.

  • The motors are strong enough to lift one other segment, which is OK, it works,

  • but to have performance on par with geared systems,

  • we'd like to be able to lift 2-3 other modules,

  • which we think we can get to with a lighter structure and with better materials.

  • Still, as far as we know this is the highest resolution

  • chain-type programmable matter system built to date.

  • We started this project with a vision of programmable matter,

  • but we ended up inventing a motor that can hold its position without power.

  • We're still working on programmable matter, of course,

  • but working with our industrial partners, we've discovered

  • that there are a lot of applications for such a small, low-power motor,

  • in aerospace and medical applications, and so we're working with them now

  • to push this technology out into the world.

So, imagine you had a bag of programmable matter,

Subtitles and vocabulary

Operation of videos Adjust the video here to display the subtitles

B1 motor module programmable chain working shape

(Tiny) Reconfigurable Robots at MIT

  • 232 8
    Josh Chen posted on 2013/02/01
Video vocabulary