The slender carbon-fiber body of the robo-insect is just half an inch long, the team reports today in Science. Its wingspan is a bit more than an inch. The transparent wings flap 120 times per second.
"The way the wing moves on our robot is very similar to how real houseflies move their wings," said Harvard graduate student Kevin Y. Ma.
The Harvard fly is not the first tiny flying robot, but it may be the tiniest. A team at KMel Robotics has built drones that are about six inches across, have rotors instead of wings, and can fly autonomously in swarms. Aerovironment's Nano Hummingbird has flapping wings, as its name suggests, that span just 6.5 inches.
But until now, flapping-wing robots as small as the Harvard one have achieved stable flight only by being attached to guide rails. The latest Harvard drone dispenses with those. "It's flying without any training wheels," Ma said.
Houseflies are annoying pests and carriers of disease, but they are also wonders of aeronautic engineering. "It's a very interesting research question to ask how flies move their wings and how they generate the lift forces they need to fly," said Ma.
A housefly maintains its course and altitude by flapping its wings independently—and mimicking that was the major challenge for Ma and his colleagues. "In order to stabilize a flapping-wing robot in the air, you have to have control of each wing separately," Ma said.
Larger robots can run on electromagnetic motors, but on this scale a different technology is needed. The "flight muscles" of the robo-fly's wings are piezoelectric actuators—strips of ceramic that expand and contract in an electric field. Thin hinges of plastic attached to the body frame serve as joints. A delicately balanced control system separately commands the flapping of both wings.
The control system is a computer in the lab, not on the fly. The drone has no onboard power supply either: There is no battery tiny and powerful enough. At the moment, the robo-fly gets both power and control signals through a hair-thin umbilical cord that trails behind it—so though it flies freely, it doesn't fly far. Within three to five years, Ma said, the team hopes to have a wireless version that can buzz around a room.
The potential applications are legion. "Because they're so small and manueverable," Ma said, "they could be used to search for human survivors under collapsed buildings or in other hazardous environments. They could be used for environmental monitoring—dispersed into the atmosphere to sense trace chemicals. And they could be used to assist with pollination of crops"—an application that sounds particularly intriguing at a time when a mysterious disease, colony collapse disorder, is decimating honeybee populations.
Before robo-bees can replace real bees, however, the Harvard researchers have a lot of work to do to improve the drone's flying capabilities. "It was very difficult to get it to be stable," Ma said. "It is not yet ready for a blast of wind."