For those of us who like to maintain a simmering, low-grade paranoia about robots — it’s entirely healthy! — here’s some news from the front lines.
Researchers at Harvard have created a swarm of precisely 1,024 tiny robots that can organize itself into complex shapes, keying off the behavior of insects like bees and ants. The aptly named Kilobots use vibration motors to scoot across flat surfaces and infrared lights to communicate with one another.
Previous robot swarms have been limited to dozens or hundreds of bots, but the Kilobot project applies some new techniques. The robots operate via an algorithm that determines three collective behaviors: edge-following, gradient formation and localization. These behaviors, when combined, allow the bots to assemble themselves into shapes — a starfish pattern, say, or even a letter of the alphabet.
If an individual bot breaks down, the swarm uses a technique called cooperative monitoring to work around the problem and recover. Even if an individual bot can’t directly sense the failure of a nearby companion, the swarm’s “hivemind” programming allows for the group to respond as a whole.
More complicated algorithms may be possible down the road, researchers say, allowing for swarms that could recover from large-scale damage to the group. The swarm could potentially use “wounded” individual bots to assist the mobility of functional bots, similar to the way ants will spontaneously form bridges with their bodies.
That’s just one of the many evocative ideas in the report, and by “evocative” I mean “terrifying.” Here’s another, taken from the project’s press materials:
“Although these thousand-bot swarms do make occasional mistakes, the researchers say that their self-assembly process never stops.”
Credit: Michael Rubenstein, Harvard Unversity