Borrowing a page out of nature’s handbook, a team of researchers at Harvard University have developed a fleet of miniscule robots that can arrange themselves into complex configurations without the use of a core artificial intelligence. Equal parts automaton and swarming hive, each of the Kilobots the team used to demonstrate their research are a little larger than a penny.
In February, computer science professor Radhika Nagpal led a team that engineered a smaller, proto-horde of machines based on termites. Algorithmically the TERMES robot groups were limited to about 100 units and building a larger group was cost-prohibitive. Nagpal’s advance, however, detailed in this month’s issue of Science, has created a swarm of more than 1,000.
Every tiny member of “the Borg” — to borrow a Star Trek term for a collective body — is composed of three parts, designed to facilitate the only three pieces of information it needs to know. An infrared sensor, motor and microprocessor allow the Kilobots to follow a swarm’s edge, understand its current position versus its original position and to recognize their distance from one another.
The programming developed by the university’s Wyss Institute for Biologically Inspired Engineering is modeled after the organizational mechanisms that allow social organisms to function as a single unit, despite their being discrete. Similar to the ways that schools of fish maneuver and reorganize themselves in response to obstacles or danger, a Kilobot swarm is able to configure itself nearly instantaneously simply by paying attention to one another. Unlike bees or ants, which rely on pheromonal direction from a queen to act, the team’s Kilobots would be able to continue to carry out their prime directive even if the mass was attacked.
The applications for automated swarms of robots are manifold, according to the researchers. With the proper programming, similar machines could form barriers in response to danger, coordinate to repair structures in dangerous areas or assist in the transportation of goods.