MIT engineers have created an insect-sized flying robot that can match the speed and agility of actual bugs for the first time. The 4-centimeter device, weighing less than a paperclip, flies 447% faster than previous microrobots and pulled off 10 consecutive somersaults in just 11 seconds while battling wind gusts.
Why Bigger Drones Can't Do This Job
Here's the problem with traditional search-and-rescue drones: they're too big. When an earthquake collapses a building, survivors often end up trapped in spaces too tight for conventional quadcopters. Real insects navigate these environments effortlessly, but until now, robotic versions flew too slowly and clumsily to be practical.
"We want to be able to use these robots in scenarios that more traditional quadcopter robots would have trouble flying into, but that insects could navigate," says Kevin Chen, associate professor at MIT's Department of Electrical Engineering and Computer Science and head of the Soft and Micro Robotics Laboratory.
The Two-Part AI Brain Making It Possible
The breakthrough comes from a clever AI control system that works in two stages. First, a model-predictive controller acts as an expert planner, using complex math to map out aggressive maneuvers like aerial somersaults and sharp turns. Think of it as a flight instructor who knows exactly how to execute every trick.
But running that sophisticated planner in real time would require too much computing power. So the team trained a lightweight AI "policy" through imitation learning, essentially compressing the expert's knowledge into a fast, efficient decision-making engine.
The result? The robot accelerates 255% faster than its predecessors and stays within 4 to 5 centimeters of its planned flight path, even when 160-centimeter-per-second wind gusts try to knock it off course.
Bio-Inspired Moves That Matter
The microrobot can now perform saccade movements, where it pitches aggressively, zips to a new position, and pitches backward to stop. Real insects do this constantly to stabilize their vision and navigate complex environments.
"This bio-mimicking flight behavior could help us in the future when we start putting cameras and sensors on board the robot," Chen explains.
The device uses soft artificial muscles that flap its wings at extremely high rates, mimicking how actual insect wings work. Chen's team has spent over five years developing these robotic insects, recently creating a more durable version with larger wings for improved agility.
What Comes Next
The immediate roadmap includes integrating onboard cameras and sensors so these microrobots can fly outdoors without relying on external motion-capture systems. The team also wants to explore how swarms of these tiny flyers could coordinate navigation and avoid mid-air collisions.
For disaster response, the implications are significant. Autonomous swarms could penetrate collapsed structures, locate survivors in spaces larger robots simply cannot reach, and relay information back to rescue teams.
"For the micro-robotics community, I hope this paper signals a paradigm shift," says Chen. The research, published in Science Advances, was funded by the National Science Foundation, Office of Naval Research, and Air Force Office of Scientific Research.
Sarah Bergbreiter, a mechanical engineering professor at Carnegie Mellon who wasn't involved in the study, calls the work "especially impressive" given the challenges of manufacturing tolerances at small scales and environmental disturbances. She notes the findings point toward "future insect-scale robots with agility approaching that of their biological counterparts."
