University of California, Irvine computer scientists have discovered a critical security vulnerability in autonomous target-tracking drones that could have far-reaching implications for public safety, border security and personal privacy. The discovery exposes a fundamental weakness in how drones perceive and track moving objects, one that operates at the physical level rather than through electronic hacking.
Ordinary umbrellas with AI-generated designs can trick the aircraft into moving steadily closer to the umbrella holder, who can then capture them with nets or cause them to crash. The UC Irvine researchers' tests successfully demonstrated FlyTrap attacks on three commercial drones, the DJI Mini 4 Pro, the DJI Neo and the HoverAir X1.
The mechanism exploits how autonomous drones operate. The aircraft's computer logic interprets images on the umbrella as a person moving farther away, even though they're stationary. To maintain its tracking distance, the drone moves steadily closer to the umbrella holder, until the aircraft can be caught with a net or crashed. This represents a new category of attack entirely. A completely passive, low-cost physical object can take consistent, deliberate control of a drone's flight path.
What makes this research particularly significant is its practicality. The system functions locally without the need for external signaling or wireless data connectivity. It can work in a variety of weather and lighting conditions, and it employs a progressive distance-pulling strategy and manipulates drone-tracking algorithms. There is no requirement for technical knowledge or specialised equipment; someone holding a printed umbrella pattern can deploy it.
The implications reach far beyond consumer drones. These AI-powered functions are increasingly deployed in applications including border control, security surveillance and law enforcement operations. The paper points to instances in which criminals could use a distance-pulling attack to evade detection by law enforcement drones. Unpiloted aircraft patrolling border zones could be similarly hampered by a FlyTrap-like attack. At the same time, people being stalked could use the UC Irvine researchers' technique to eliminate a harassing drone.
The research presents a genuine complexity: the same vulnerability that threatens law enforcement capabilities could potentially protect individuals from unlawful surveillance. Lead author Shaoyuan Xie, a UC Irvine graduate student researcher in computer science, said "Our findings highlight urgent needs for security improvements in [autonomous target-tracking] systems before wider deployment in critical infrastructure."
The team has responsibly disclosed these vulnerabilities to manufacturers DJI and HoverAir. The team is sharing its findings and specifications for the FlyTrap attack platform in a paper presentation this week at the Network and Distributed System Security Symposium in San Diego. Chen's research group discovered what it calls a distance-pulling attack that physically draws victim drones closer to an attacker. An ordinary umbrella covered with a specifically designed visual pattern can deceive neural network tracking systems used by autonomous drones.
As autonomous systems play an expanding role in public safety and infrastructure protection, the FlyTrap research reveals that security cannot be assumed simply because a system is technologically advanced. The vulnerability sits at the intersection of artificial intelligence and physical reality, a boundary where conventional cybersecurity measures do not apply. Until manufacturers and policymakers address this gap, the widespread deployment of autonomous tracking technology in sensitive applications carries genuine risks.