WC Archive

An ensemble of material that contains sufficient

local computation

actuation

storage

energy

sensing & communication
Which can be programmed to form interesting dynamic shapes and configurations.

Claytronics is their way of bringing this concept into reality.

With claytronics, millions of tiny individual devices -- "claytronic atoms" or "catoms" -- would assemble into macro-scale objects, connecting and disconnecting as they move. The current large proof-of-concept catoms (measuring 4.4 centimeters) connect and move via magnets, much like the "replicating" robots devised by Cornell last month; later versions, operating at scales where electromagnetic or electrostatic connections would be hard (or impossible) to control, will likely use something akin to the non-sticky adhesive found on geckos' feet. The catoms could have LCD or LED surfaces able to produce a faintly glowing image, so that what appeared to be a person (say) made of millions of tiny microbots would actually look like the person. That's important to Mowry and Goldstein -- they see claytronics as a way of being multiple places at once.

Claytronics research arose out of a combination of work on microscale computing devices and work on telepresence, so it's not surprising that the researchers emphasize the utility of the claytronics system as a means of doing virtual meetings with apparent physical presence. They also suggest that the system could have a role in telemedicine, allowing a patient and doctor to be on different continents, but each able to the physical presence of the other with claytronic emulations. As New Scientist puts it (subscription required), it's not teleportation, but it's the next best thing.

Videos of prototype catoms and animations of future designs in action can be found on the CMU website. As they make clear, the challenges remaining to work out for this vision to come about are profound. The hardware aspect of the task is obvious and significant: taking crude, clumsy massive devices and make them work at a size on the order of a cell. But the software challenge is, if anything, more staggering. Coordinating the work of a few dozen robots in a swarm is an incredibly hard problem, let alone being able to program millions to work in concert. How does each catom indicate its position, and know where to go? How do they indicate when a catom is missing or non-functional, and shift to take its place? How do they say, "now change?"

Goldstein and Mowry refer to the larger context of transmissable synthetic reality as "pario," from the Latin for "create, make, produce, bring forth," and put together a simple presentation (PDF) describing how the system works. To me, the vision they describe seems both possible (eventually) and remarkably timid. Programmable matter is as much of a leap over fabbing as fabbing is over current rapid prototyping. It's beyond "design for disassembly" -- it's "design for infinite reassembly."

(Via We Make Money Not Art)