US scientists' unique umbrella can deceive enemy's strike drones

Researchers at the University of California, Irvine have developed a novel method to deceive autonomous strike drones that use camera-based target-tracking technology. Their approach, called FlyTrap, involves using an ordinary umbrella covered with a specially designed graphic pattern that tricks the drone’s neural network tracking system. The drone’s AI interprets the umbrella’s pattern as a person moving away, prompting the drone to move closer to maintain tracking distance. This “distance-pulling” attack enables the drone to be physically captured with a net or forced to crash, representing a significant advancement over previous attacks that only caused drones to lose tracking. The FlyTrap technique works locally without external signals or wireless connectivity and is effective under various weather and lighting conditions. The researchers successfully demonstrated the attack on three commercial drones—the DJI Mini 4 Pro, DJI Neo, and HoverAir X1—highlighting vulnerabilities in widely deployed autonomous tracking systems used in security, law enforcement, and border patrol. The team has responsibly disclosed these security flaws to the drone manufacturers