Ukraine used AI drones against Russia for the first time

Ukraine used AI drones against Russia for the first time

Ukrainian startup Strategy Force Solutions (StratForce) claims that their mother drones have successfully conducted the first autonomous missions with FPV strike drones during tests against Russian targets.

As reported by Forbes, this is a significant step in the application of artificial intelligence in combat.

"A mission costing $10,000 replaces what previously required missile systems costing $3-5 million," Forbes quotes StratForce’s CTO Andriy, whose last name is undisclosed.

StratForce’s approach involves using the reusable mother drone GOGOL-M, which delivers two FPV-type strike drones. These drones can hit precise targets at distances of up to 300 km, allowing them to have a significant impact on vulnerable targets such as parked aircraft, air defence sites, or infrastructure.

Although StratForce is currently unable to release images of the attacks, this work continues the previously noted use of Ukrainian drones and FPV with AI-based targeting. In fact, it can be considered as an operational version of a long-range autonomous system being developed by the Pentagon’s Intelligence Directorate or a mother system recently showcased by China.

"Combining them [small FPV-type drones] with AI-driven mother drones, we can ensure precise strikes," notes Andriy.

Andriy explains that the core idea came to him many years ago when he was a boy and saw a woman inspecting railway tracks. He thought at the time that this exhausting job should be done by a robot. Years later, Andriy developed drones for automated infrastructure inspection that can detect problems and provide detailed images.

The full-scale invasion of Russia in 2022 interrupted his work on infrastructure inspection, and he refocused his AI project on defending Ukraine.

The key product of StratForce is the SmartPilot system, which uses a combination of advanced sensors and artificial intelligence. This system, rooted in infrastructure inspection, requires high precision.

"In a sense, it’s like an unmanned car," explains Andriy. "For autopilot to work properly, it needs many cameras. There are fewer obstacles in the air, but we make the system lightweight. This is how we developed our camera, LIDAR, and communication system, enabling the AI to navigate, coordinate actions, and maneuver around obstacles."

LIDAR, or laser radar, creates a 3D map of the environment and works under any lighting and weather conditions. SmartPilot uses a multi-sensor data fusion approach to perceive the environment and recognize targets. This world view allows the AI to make decisions, plot the flight path, and carry out the mission like a human pilot.

Andriy claims that SmartPilot can conduct missions autonomously, finding its way to the specified location and hitting targets there. "It provides autonomous flight, navigation, and interaction without GPS and constant operator control," he says.

In typical use, the mother drone GOGOL-M with a 20-foot (6-meter) wingspan flies to the target zone and releases two FPV-type drones. Each of them has its own lightweight version of SmartPilot. The mother drone returns, leaving the FPV drones to find, identify, and attack predefined targets such as airbases, missile launch sites, or air defence assets.

The system is also effective against vulnerable infrastructure such as oil storage, distribution networks, and railways. Unlike drones like the "Shahed," which strike only one target, the mother drone can hit multiple targets and then be reused.

SmartPilot is not limited to static targets. "It supports ambush missions, landing and waiting for targets, as well as autonomous real-time search," says Andriy. Drones can land and wait at an airbase for aircraft to arrive or emerge from shelters. They can also be left on a convoy’s path, ready for an autonomous attack when it appears.

StratForce can currently produce 50 GOLGOL-M mother drones per month and 400 FPV-type strike drones, but this will depend on receiving military contracts.

Feedback from trial missions has been positive. "It’s like a video game. I just set waypoints, choose targets, and watch it work," Andriy quotes one user. Another user added, "It’s fascinating to watch from the coordination center. I just wish we could increase the warhead and range to 500 km."

An important element of StratForce’s solution is the software that allows the integration of drones of any size, shape, or configuration. The first crucial step—delivering several small autonomous strike drones over long distances—has already been achieved. Forbes notes that this is likely to cause serious problems for Russia in the coming months.