BEER-SHEVA, Israel, Jan. 26, 2018 — Researchers have demonstrated a method for detecting whether a drone camera is illicitly recording a targeted subject or private home.
Previous studies have shown methods for detecting whether a drone is nearby; however, according to the team, none of these studies has shown how to distinguish between the legitimate and illegitimate use of a drone.
Researchers from Ben-Gurion University of the Negev (BGU) and Weizmann Institute of Science investigated different methods for detecting whether a specific point of interest (POI) was being tracked by a drone, even if the first person view (FPV) channel was encrypted. The methods they proposed trigger a physical stimulus that influences the encrypted FPV channel.
Researchers investigated the influence of changing pixels on the FPV channel. Based on their observations, they showed that an interceptor can perform a side-channel attack to detect if a target is being streamed by analyzing the encrypted FPV channel that is transmitted from a drone. Researchers demonstrated their technique in two use cases: in one case the target was a private house and in another the target was a subject.
To show how a privacy invasion against a house could be detected, researchers used smart film placed on a window and entered software commands on a laptop to access the encrypted FPV channel seen by the drone operator. Using this method, they demonstrated a way to detect that a neighbor using a DJI Mavic drone to capture images of his own home could illicitly stream video of his neighbor’s house as well.
In a second test, researchers demonstrated how an LED strip attached to a person wearing a white shirt could be used to detect targeted drone activity. When researchers flickered the LED lights on the shirt, this action caused the FPV channel to send an “SOS” by modulating changes in data sent by the flickering lights.
“This research shatters the commonly held belief that using encryption to secure the FPV channel prevents someone from knowing they are being tracked,” researcher Ben Nassi said. “The secret behind our method is to force controlled physical changes to the captured target that influence the bit-rate (data) transmitted on the FPV channel.”
The method can be used on any laptop that runs Linux OS, and does not require any sophisticated hacking or cryptographic breaking skills.
“The beauty of this research is that someone using only a laptop and an object that flickers can detect if someone is using a drone to spy on them,” said Nassi. “While it has been possible to detect a drone, now someone can also tell if it is recording a video of your location or something else.
“Our findings may help thwart privacy invasion attacks that are becoming more common with increasing drone use,” Nassi said. “This could have significant impact for the military and for consumers because a victim can now legally prove that a neighbor was invading their privacy.”
The research was published in ArXiV (ArXiv:1801:03074v1).
Game of Drones — Detecting Streamed point of interest from encrypted FPV (first person view) channel. Courtesy of Cyber Security Labs at Ben Gurion University and Weizmann Institute.