These Tiny Little Robots Are Helping Us Understand the Science of Swarms

Have you met the Kilobot? It’s small, simple, inexpensive—not the most impressive robot, on its own. But the Kilobot wasn’t made to work alone, or even in twos or threes. No, where the Kilobot really thrives is in groups. Since 2010, the size of the group in which it cooperates has steadily grown, and with it the range of complex behaviors that researchers can use the Kilobots to study.

The complex behaviors I’m talking bout are the kind that arise from the interactions of countless simple, individual actors. Researchers call this phenomenon “emergent behavior,” and if you’ve ever seen a colony of ants, studied microscopic footage of cell division, or watched a flock of starlings become a roiling, shape-shifting cloud, you’ve witnessed it for yourself. Kilobots were designed by Harvard research scientist Michael Rubenstein to simulate and study these behaviors in a programmable robotic analog. Last year, his team documented that a Kilobot horde more than 1000 members strong could self-organize into a range of complex shapes, making it the largest, most technically impressive robotic swarm ever created.

In the latest episode of KQED’s Deep Look, Joshua Cassidy and Johanna Varner report on how Rubenstein and his colleagues are using Kilobots to study some of nature’s most complex emergent systems, and how they might one day teach us to reprogram our own software:


Contact the author at


Share This Story