Nine tiny robots, each half the height of a toothpick, idle in a makeshift arena. The light atop every Lego-sized brick of exposed silicon and circuitry turns from red to green, and immediately the mini-droids are in motion, all waddling in the same general direction. The squadron has been instructed to cluster together and huddle close. When a robot completes its task, its bulb blinks blue.
The man in control is 25-year-old PhD student Justin Kim. Perched on a high stool in a basement lab at the University of Toronto, workbench to his left and robots to his right, Kim is helping to advance the nascent — but potentially lifesaving — field of swarm robotics.
Kim’s is a type of mechanical engineering modelled on the collective intelligence exhibited by small insects like honeybees, ants, and termites. An individual termite isn’t particularly strong or savvy — but in groups, termites band together to build eight-foot-tall fortress-like mounds. Same deal with swarm robotics. “It’s why I love it,” Kim says: “simple robots, using simple rules, can do such complex stuff.” At 32mm tall and with a 16mm x 16mm footprint, each of his droids — called mROBerTO, short for “millirobot Toronto” — is equipped with infrared sensors it uses to navigate its environment, plus Bluetooth-like technology that lets it communicate with other bots. They cost $60 apiece in parts.
Our journalism depends on you.
You can count on TVO to cover the stories others don’t—to fill the gaps in the ever-changing media landscape. But we can’t do this without you.
“The practical applications are exciting,” Kim says. Imagine a team of human-sized bots, working in conjunction with emergency workers, on a search-and-rescue mission. Picture them working in the wake of a natural disaster, gathering data and delivering relief. Conjure them crisscrossing the ocean floor, revolutionizing environmental monitoring. Or replace wading through tangled seaweed with exploring the rusty surface of Mars. (In a less peaceable vision, the U.S. military has already considered combat applications.)
- Studying the human eye to develop robots — and surveillance systems — that may keep you safer
- How automation threatens Ontario workers
- The Agenda: Here come the robots
Childhood toys were Kim’s introduction to robotics. Growing up in South Korea, he would play with remote-controlled cars, modifying them by adding cameras and IR sensors. After he moved to Ontario at eight years old, his inner robophile flourished. His Grade 9 physics teacher showed him the basics of circuitry. In Grade 12, that same teacher brought in a box of parts — junk from his basement — and told his students to build something cool. Kim ginned up a wire-controlled car equipped with two motors, which he thought was pretty good. That is, until he saw that one student had made a RC vehicle that could stop on a dime upon sensing an obstruction. Was Kim jealous? “Yes, definitely.”
He went on to study mechanical engineering at U of T, first giving serious thought to the swarm concept in 2011, during his time in the school’s extracurricular robotics club. “It was so simple, and the potential was so great,” he says. “The field was so new.” Kim believes swarm teams are still 10 to 15 years away from surveying sinkholes, but he’s using his PhD studies to carve out a place as an authority in the field. His robots are the world’s smallest that can “think on their own,” he says. They’re also the most modular, which makes them excellent fodder for researchers looking to test multiple algorithms on the same swarm.
As the robo-team executes Kim’s second command, “separate,” the greatest force holding swarm robotics back reveals itself. Two bots scamper away from the crowd and tumble over all tuckered out, their light bulbs extinguished. Kim explains that mROBerTO units last only an hour or two on a single lithium-polymer battery — a long time relative to the competition, but not much use in an emergency, let alone on Mars. The tech falls short in other ways, too: the field needs more-advanced algorithms, plus improved sensory and communication tools. And while Kim sources only widely available parts — the kind you could get at Home Depot — others have seen promising advancements dashed by manufacturers who stopped producing vital components.
But none of this deters Kim, who believes the solutions are attainable. He hopes to complete his PhD in the next three years. That means creating a team of more than 20 millirobots that perform to as-yet-undetermined specs. Afterwards, he wants to work in research and development, and to design his own hardware (better, more-specialized hardware = better robots). If all goes well, Kim says, mROBerTO will follow him wherever he goes: “One day, they’ll be fast enough to run with me.”
Nathaniel Basen is a Toronto-based freelance journalist.