A solution to track malicious drone operators

A solution to track malicious drone operators Cybersecurity

The disorders attributable to the theft of untimely drones could well become ancient history. Researchers have in fact found a way to locate drone operators likely to cause damage or disruption in protected airspace.

As a reminder, the damage caused by these impromptu flights near sensitive areas – such as military bases or airports – had been highlighted by the Gatwick drone incident, following which the main British airport had to close for 33 hours between December 19 and December 21. London’s second airport, inundated with passengers this Christmas, had to cancel flight after flight because of a malicious drone. In total, 140,000 passengers were impacted by this drone flight.

Academics from the Ben-Gurion University of the Negev (BGU) have demonstrated a potential way to end these acts. They examined how the analysis of flight paths can be useful in finding malicious operators. “Currently, drone operators are located using RF techniques and need sensors around the flight area, which can then be triangulated,” said one member of the research team.

A machine learning solution

If solving this problem was “a challenge due to the quantity of other Wi-Fi, bluetooth and IoT signals in the air, which obstruct the signals of drones”, the Israeli researchers finally succeeded, relying on the use of a neutral network. Rather than focusing on attempts to unravel a variety of signals, the network has been formed to predict the position of an operator using only flight paths – even in motion.

AirSim, an open source multi-platform drone simulator was used to perform the tests, using 10 km of roads and realistic obstacles such as buildings. In total, the algorithms were able to predict the position of a drone operator with an accuracy of 78% during the simulations. Although the experience is small, the BGU indicates that possible areas for improvement include improving the machine learning system, or even attempts to find out the level of skill or training of an operator.

The research team now intends to repeat the experiments with drones in real time. BGU’s research was presented at the fourth international symposium on cybersecurity, cryptography and machine learning (CSCML 2020) on July 3.

Source: ZDNet.com

Source: www.zdnet.fr

Rate article