Airports are protected from drones by artificial intelligence

Airport security can be enhanced by a newly developed artificial intelligence-based system capable of locating the controllers of drones flying into the protected airspace unauthorized. Scientists have developed a tool based on artificial intelligence to reduce the danger posed by drones to aircraft. There can be several such risk factors: on the one hand, drones can collide with aircraft, and on the other hand, they can interfere with sensitive radio signals, which can even result in the pilot losing control of the aircraft.

And these are not just problems that exist in theory, as drones have caused disruptions at several airports in the past. It has been more than once in the world that runways had to be closed due to illegal drone activity. But perhaps the biggest press coverage was triggered by the December 2018 incident at London Gatwick Airport, when drones spotted near the runway had to stop the launching and receiving of planes altogether. The incident affected roughly 1,000 flights and 140,000 passengers. And the identities of the perpetrators remained unknown in almost all cases. According to Eliyahu Mashhadi, a senior researcher at Ben-Gurion University in Israel, the biggest problem is that drone operators use RF (radio frequency) technology, and it is almost impossible to measure the position of controllers using the triangulation method. This is because, in addition to the many other RF signals around the airport, there are also Wi-Fi, Bluetooth, and IoT signals that make it difficult to identify those connected to the drone.

Complicating matters further is the fact that each drone brand may have a different set of symbols. Moreover, to measure the guide of the drone, the radio signals would have to be recorded near the drone and then could only be measured with cryptographic and electronic military equipment. The tool, developed by the university, seeks to overcome these difficulties by replacing radio frequency with artificial intelligence. The method analyzes the flight path of the drone from various perspectives, using the neural network. And the detected samples make it possible to determine the exact position of the drone pilot. When tested on the simulated drone routes, the new model was able to determine the control position with 78% accuracy, and researchers are now further developing the artificial intelligence-based system with data from real drones. The more reliable and accurate this solution will be, the safer airports can become in the future.