In a recent Forbes report, technology correspondent David Hambling examines how Ukrainian forces are deploying AI-enabled drones to counter Russian jamming. These small drones, especially FPV models, have become decisive weapons “destroying roughly twice as much enemy equipment as every other weapon put together.” However, jamming reportedly brings down “at least half of FPVs before they can reach the target.”
‘Michael,’ Commander of the Typhoon drone unit of the National Guard of Ukraine, shared mixed results from field testing.
“We’ve tested this solution a couple of times, and the results were quite inconclusive. While it was able to lock on target, the target needed to be highly distinguishable from the background,” he explained.
The Russians claim they have recovered a Ukrainian FPV with AI target acquisition and terminal guidance.
— Roy🇨🇦 (@GrandpaRoy2) November 11, 2024
Long awaited, these jam-resistant drones will hopefully soon appear frequently. pic.twitter.com/wARWRySdrL
Developers claim their AI-equipped drones can achieve an impressive 80% hit rate against targets, supposedly surpassing human operators. However, Michael disputes this assertion.
“A skilled pilot with a good technical setup and properly configured stack settings can achieve this level of success,” he noted, suggesting experienced human operators still match or exceed AI performance.
Ukrainian developers are keeping costs down by using affordable hardware like a cheap Raspberry Pi Zero—costing $100—or a Google Coral AI dev board for their autonomous drones. This stands in contrast to Russia’s Lancet drones, which utilize more powerful US-made NVIDIA processors for their targeting systems.
The technology faces several practical challenges. Current AI systems struggle with tracking soldiers in woodland areas and lack the tactical sophistication to target vulnerable points like “the thinly-armored turret rear of Russian tanks.” Most FPVs use low-resolution cameras that limit AI effectiveness, and environmental factors create additional complications, with Michael noting that “target acquisition is especially difficult in summer conditions.”
Despite limitations, Michael identifies promising use cases.
“This feature could be particularly useful for engaging unarmored vehicles such as cars and motorcycles, as well as small infantry groups on roads,” he said.
Hambling suggests these drones could excel in ambush roles, potentially “turning the drone into a long-range anti-tank mine.”
Russia’s Lancet drones faced similar issues with their lock-on function, though these problems “now seem to have been solved,” with the feature used in about a third of attacks.
Ukrainian manufacturers acknowledge current limitations, with Oleksii Babenko, CEO of Vyriy Drone, stating, “Machine vision isn’t developing as fast as we’d expected.” Nevertheless, the Ukrainian government ordered “the first batch of 3,000 drones with machine vision” in November, with Deputy Prime Minister Mykhailo Fedorov planning to “significantly increase” the proportion with autonomous targeting.
Hambling concludes that while AI-enabled drones “have not yet swept the battlefield,” they have evolved from “laboratory experiments to viable if limited weapons” in just 18 months and will be “definitely a space to watch in 2025.
Read more:
- Ukraine’s 228 years of combat footage powers next-generation AI drones
- Ukraine develops new AI drones to counter Iranian-made Shaheds
- Japan to supply AI drones to Ukraine for humanitarian demining
