Harvard faculty members recently unveiled a swarm of tiny robots, dubbed “kilobots.” The swarm consists of 1,024 members, each unit about the size of an after-dinner mint with three miniature legs. These self-organizing robots hint at the future of collective intelligence and an Internet of cooperative things.
Scientists can submit a command for the swarm to collectively assume a shape, such as the letter “K” or a star, and the robots reassemble to create the shape. There is no preset position defined for any of the kilobots; units at the outer edge of the group use spatial relationships to move around the swarm and find an optimal place. Relative positioning is communicated via LEDs and infrared (IR) receivers and transmitters. As each unit finds its position, the remaining kilobots communicate and fall into place until the shape is complete.
The Harvard team has experience in self-organizing robots: its TERMES project consisted of several robots reacting cooperatively to each other and their surroundings to build pre-determined structures.
Kilobots: dumb while apart, together they’re smart
Kilobots consist of little more than a circuit board, LED light, IR transmitters and receivers, and tiny motors. Individually, their intelligence is highly limited, which is the challenge and the purpose behind the project. To understand why researchers are so interested in collective intelligence, consider their inspiration: nature.
From schools of fish to bee hives, collective intelligence is instrumental to group behavior and survival across the animal kingdom. Self-organizing and self-directing species are capable of complex social structures. Ant colonies wage war, harvest swaths of foliage, and form structures to cross barriers. Even “smarter” species—such as dolphins, elephants and humans—live and work together in groups.
According to research associate Michael Rubenstein, “biological collectives involve enormous numbers of cooperating entities—whether you think of cells or insects or animals—that together accomplish a single task that is a magnitude beyond the scale of any individual.”
Applications for self-organizing cooperative intelligence
As the Internet of Things shapes discussions about near-term innovations, collective AI can shed light on potential applications for IoT 2.0. These cooperative capabilities will come into focus with the advancement of complementary technologies, including nanotechnology and wireless power transfer.
The current vision for smart, connected products is tethered to monitoring information and directing actions. Some of these applications (especially on the consumer side) communicate from machine to person (M2P), like Nest thermostat systems. In commercial and industrial uses, communication occurs machine to machine (M2M). Even in M2M applications, today’s interactions are primarily informational. Kilobot and TERMES experiments demonstrate the potential of the next step: smart, connected, and cooperative devices.
Hive-driven construction and repair
Smart swarms could be used to assemble and repair structures that are too costly, dangerous, or impossible to deploy human resources. Skyscraper construction, underwater operations, Superfund site cleanups, and even satellite repair could benefit from self-organizing, cooperative robot swarms.
Self-organizing polymorphous tools
As smart swarms reduce in size and improve their capacity for physical cohesion, there is potential to replace a large collection of tools with a polymorphous smart hive capable of forming into a desired structure on request. Reassembling from wrench to hammer is a simple example, but as technology advances, there will be far-reaching capabilities to benefit specialists who need to travel light in the field. Polymorphous tools, machines and even structures could redefine how humans live and work.
Additional applications, such as defense, intelligence, emergency response, and health sciences clearly have potential uses for hives of smart, connected, cooperative products as well. Kilobots aren’t just a passing curiosity, they represent a future convergence of artificial intelligence and the Internet of Things.
Photo Credit: Radhika Nagpal