To design invisible drones, researchers have tried camouflage, transparent materials and light-bending optical systems. But engineers at Northwestern University used "motion blur," which an announcement from the school notes is the effect that makes fast-spinning fans seem to disappear. "The drone spins up to 25 times per second, which is too fast for the human eye to see clearly. While it isn't completely invisible, it morphs into a ghostly smudge that seamlessly blends into the background... " To design the drone, the Northwestern team first used a computational model to generate roughly 20,000 drone configurations capable of stable flight. Then, they used artificial intelligence (AI) and optimization algorithms to repeatedly rearrange the drones' major components, including a motor, propeller, circuit board, counterweight and batteries. After sifting through many different configurations, the algorithms determined the ideal placement of the drone's components to minimize its visibility from virtually every viewing angle while allowing for stable flight. After selecting promising candidates, the engineers simulated each drone spinning in flight and overlaid those images a hundred real-world backgrounds. Then, they used a perception model that approximates human vision to determine how noticeable each design appeared. Designs that blended into their surroundings received lower visibility scores. The team selected the 500 lowest-scoring designs and applied the optimization algorithm, which repeatedly adjusted the positions of components to further minimize those scores. "The design process was fully automated," [said Northwestern's associate CS professor Michael Rubenstein, who led the work]. "Then, when we were confident that a drone met all our criteria, we built it." They've named the drone "Phantom Twist"," and plan to design future iterations with more transparent materials or quieter propulsion to make it even less noticeable. The Northwestern team presented the work on July 16 at Robotics: Science and Systems 2026 in Sydney, Australia... Thanks to long-time Slashdot reader fahrbot-bot for sharing the news.Read more of this story at Slashdot.
A Tutorial on GeoAI: Designing Footprint Extraction from NAIP Imagery Using U-Net, Grounding DINO, SAM, and Mask R-CNN