Ukraine is using dozens of domestic systems enhanced by artificial intelligence to enable its drones to reach battlefield targets without being piloted, a senior official said, revealing new details about the race against Russia to harness automation.
Systems using artificial intelligence allow cheap drones carrying explosives to locate or fly to their targets in areas protected by extensive signal jamming, which has reduced the effectiveness of manually piloted drones.
In March, a drone strike carried out its mission by targeting an oil refinery in Russia. The Russian government said its defense systems intercepted more than 150 drones.
The shift toward the use of AI, particularly in locating drone targets and controlling their flight, is a major emerging front in the technology race that has unfolded since Russia launched its full-scale invasion in February 2022.
“Currently, there are several dozen solutions on the market from Ukrainian manufacturers… they are being purchased and delivered to the armed forces and other defense forces,” Ukraine’s Deputy Defense Minister Kateryna Chernohorenko said of drone AI systems.
She said they were currently being used in a targeted manner in special operations.
Automated drone systems are in high demand among soldiers looking for ways to overcome the increasing use of electronic warfare on the battlefield.
Electronic warfare systems create a protective dome around their location, sending powerful signals that disrupt communication between drones and their pilots, causing them to lose control of the aircraft and miss their target.
These systems, once used only to protect higher-value equipment, have become a common feature in trenches and on common vehicles used by soldiers seeking to protect themselves from the threat of first-person view (FPV) drones.
These small, inexpensive drones, originally built for civilian enthusiasts to race, have become the most commonly used attack drones on the battlefield, with both countries increasing their production to millions per year.
A Ukrainian official told Reuters in July that the target hit rate of most first-person view units had fallen to 30%-50%, while for new pilots that rate could be as low as 10%, and that signal jamming was the main problem.
The official predicted that AI-operated first-person view drones could achieve accuracy rates of around 80%.
Computer vision
Dmytro Vovchuk, chief operating officer of NORDA Dynamics, a Ukrainian company that makes software for drones, told Reuters they were developing a product that used computer vision, a type of AI technology, to guide attack drones toward their target.
The software allows the pilot to select a target using the drone's camera, at which point the aircraft completes the rest of the flight autonomously.
The company has sold more than 15.000 units of its automated targeting software to drone manufacturers, with more than 10.000 of those already delivered.
While in raw that is a large number, it is still a small proportion of the 4 million drones Ukraine says it is capable of producing annually.
Vovchuk said attacks cannot always be confirmed visually due to the heavy presence of electronic warfare systems around high-value targets.
“From what we saw, three tanks were definitely destroyed with our systems, as well as many (strikes) on logistical targets,” he said, adding that the system was also used to hit field headquarters.
information via Reuters
Read also
- Total Linhas Aéreas confirms purchase of Chinese aircraft COMAC
- ANAC publishes EVE eVTOL airworthiness criteria
- Salgado Filho Airport: Fraport announces results of the second week of reopening
