Researchers at MIT have created a significant advancement in aerial microrobotics by developing an artificial intelligence-based control system for tiny flying robots. These insect-scale machines, which weigh less than a paperclip and are roughly the size of a microcassette, are designed with potential applications in search and rescue operations following natural disasters. Their small dimensions would allow them to navigate through confined spaces and debris that larger drones cannot access.
The team’s new controller represents a major breakthrough in robotic flight performance. By implementing a two-part control system, the researchers achieved a 450 percent increase in flight speed and a 250 percent improvement in acceleration compared to previous iterations. The robot demonstrated sufficient agility to execute 10 consecutive somersaults within an 11-second timeframe while maintaining its intended flight path despite wind disturbances.
The control system combines two complementary approaches. The first component uses a model-predictive controller that employs mathematical models to plan complex maneuvers such as aerial flips and sharp turns while accounting for the physical limitations of the robot. This planning system then trains a deep-learning model through imitation learning, creating a real-time decision-making policy that operates with sufficient speed for live flight without excessive computational demands.
The improved flight capabilities bring the robotic insects closer to matching the natural performance of biological insects in terms of speed, acceleration, and body angle control. The research represents the culmination of over five years of development in the field of microrobotics and was published in Science Advances. The findings underscore how advances in hardware design and software control systems can work synergistically to unlock new capabilities in robotic flight.
Article Attribution | Read More at Article Source
Article summary produced by Claude AI