Ukraine is rushing to develop AI-powered war drones

Signal interference

The need for AI-enabled drones has become more pressing as both sides roll out electronic warfare (EW) systems that can jam signals between pilots and drones.

In particular, small, cheap FPV (first-person view) drones became the primary way for both sides to strike enemy vehicles in 2023, but their hit rate is declining as interference increases.

“We are already working on a concept where in the near future there will be no connection between pilots and drones on the front lines,” said Max Makarchuk, head of artificial intelligence at Brave1, a defense technology accelerator set up by the Ukrainian government.

Makarchuk said FPV hit rates have been decreasing. Currently, most FPV devices have a hit rate of 30% to 50%, while for novice pilots, the hit rate may be as low as 10%.

He predicts that the hit rate of an AI-operated FPV drone could reach around 80 percent.

To counter electronic warfare threats, manufacturers including Swarmer have begun developing features that allow drones to lock onto targets via their cameras.

Electronic warfare systems create an invisible signal jamming bubble over the equipment and soldiers they protect.

If the pilot loses contact with the drone, they can no longer control it and it will either fall to the ground or continue flying in a straight line.

Automating the final stages of a drone’s flight to a target means it no longer requires a pilot, eliminating the effects of electronic warfare jamming.

AI drones have been in development for years but have until now been seen as expensive and experimental.

Bendett said Russia had been developing AI-enabled aerial and ground drones before the 2022 invasion, with some success.

In Ukraine, the main task for manufacturers is to produce a cheap AI targeting system for drones. This would allow for large-scale deployment of the system along the entire 1,000 km front line, using thousands of FPV drones per week.

Costs can be kept down by running AI programs on a Raspberry Pi, a small, inexpensive computer that has gained worldwide popularity in addition to the educational uses for which it was designed.

Makarchuk said he estimated it would cost only about $150 per drone to install a simple targeting system that could lock onto shapes visible to the drone’s camera.