If you could put a reaction control system into them to allow them to rendezvous with each other, you might have some interesting fractionated satellites.
This is a good example of Arthur C. Clarke’s rule that when a distinguished elder scientist says that something can’t be done, he’s very likely wrong.
Still waiting for someone with sufficient funding to build more than single digit units to do anything with configurable robotics 🙂
In this case, Trent, I can easily see mass production bringing the per unit cost way down. Each of the circuit boards is relatively small and square, so the whole robot could be built on a single panel, and an SMT run to populate a few hundred panels would cost much less per panel than the few prototype units they have going. A quick back of the napkin calculation gives a cost less than $100 per unit for a run of 1000.I expect MIT will do a few more prototype revisions before getting to that point, but you’re not going to be waiting very long.
I think this idea is brilliant. Imagine landing 1000 on Mars as a half-meter cube and having them take off in ten different directions, and then 100 different directions. Who cares if you lose half of them, or even 80%? And after a few design iterations they could be the size of sugar cubes…
The Hackaday crowd is going to be all over this like white on rice.
They’ve been saying the same thing for decades.
No-one ever has the funding to build more than a few, and the only people who build anything useful are the people who leave MIT.
Self assembling robots? Hmmm… That may not be such a good idea.
http://www.youtube.com/watch?v=7yBnl_krN_U