AI-driven controller boosts insect-like robot's speed and agility
MIT's insect-scale flying robot, using a new AI control system, achieved a 450% speed increase and performed 10 flips in 11 seconds. The technology aims to enable navigation in tight spaces like earthquake rubble where larger drones can't go.
The robot's two-part control system balances computational efficiency with flight performance, a departure from earlier versions where human operators manually tuned the controller. The research team, led by Kevin Chen, has spent over five years refining the robotic insect design, with the latest iteration featuring larger flapping wings driven by soft artificial muscles that beat rapidly enough to sustain agile maneuvers.
Published in Science Advances, the work involved collaboration across MIT's electrical engineering and aeronautics departments. The robot's size—comparable to a microcassette and lighter than a paperclip—positions it for applications in confined environments, though the team's immediate focus remains matching insect-level agility under real-world disturbances like wind.
Search-and-rescue operations could benefit significantly if this technology matures, as the robot's small size may allow it to access collapsed structures where larger equipment cannot reach. Emergency responders could potentially deploy such devices to locate survivors in earthquake rubble, though real-world conditions—including dust, moisture, and unpredictable debris—would likely require further refinement. The technology may also influence other micro-robotics applications, from industrial inspection to environmental monitoring, but widespread deployment remains dependent on continued reliability improvements.