Northwestern University Researchers Develop 'Invisible' Drone
Researchers at Northwestern University in Evanston, Illinois, have developed a drone named Phantom Twist that is significantly less visible to the human eye. Presented at RSS 2026 in Sydney, this drone achieves its near-invisibility by spinning rapidly between 15 and 25 Hz. This high-speed rotation exploits the 'persistence of vision' phenomenon, where the human eye integrates visual information over a short period, causing the rapidly moving drone to appear as a transparent blur against its background. The design was computationally optimized to minimize visual overlap of its components as it spins. This optimization process used a metric called Learned Perceptual Image Patch Similarity (LPIPS) to quantify how well the drone blends with its background. The resulting Phantom Twist design has a LPIPS score significantly lower than human-designed versions or conventional quadrotors, making it an order of magnitude harder to see. The drone's flight is controlled by pulsing the motor speed during each rotation, with altitude adjusted by overall thrust, and its spinning nature provides passive stability. Components like the motor, propeller, batteries, controller, and counterweights are connected by carbon fiber rods. While currently controlled via an optical tracking system in a lab setting, the researchers are optimistic about enabling outdoor flight and potential sensing applications, such as 360-degree imaging. Potential uses range from covert surveillance to wildlife observation, aiming for a less intrusive presence.
This research demonstrates how computational design can optimize for perceptual qualities, such as low visibility, which are difficult for humans to intuit. The Phantom Twist's design leverages fundamental aspects of human vision and physics to achieve its effect. While the immediate application may seem to lean towards surveillance, the researchers also highlight non-intrusive wildlife observation, suggesting a dual-use potential. Future development will likely focus on mitigating the drone's acoustic signature and enhancing its autonomy for real-world deployment. The underlying principle of computationally optimizing for human perception could extend to other fields, influencing how we design interfaces, vehicles, and even urban environments to better integrate with human sensory systems in the coming decades.
AI-generated to prompt reflection — not editorial opinion, not advice, not a statement of fact. How this works.