ALMA Demonstration Shows How Autonomous Drones Can Identify and Engage Targets

Scaleout Systems has demonstrated a functional concept for a drone-based surveillance and reconnaissance system that uses onboard artificial intelligence to detect, identify, prioritize, and locate potential threats.

“We have built a functional concept of how to do this for a drone-based surveillance and reconnaissance system,” Helander told Ars. “This is what we did at the public demonstration in Sweden.”

ALMA combines autonomous drones with onboard AI

Scaleout Systems is participating in the Affordable Loitering Modular Ammunition (ALMA) Project, led by BAE Systems Bofors. The project aims to develop low-cost autonomous one-way attack drones. The ALMA concept was first publicly demonstrated at the Winter Demo 2026 event in Sweden in January.

According to a source describing the demonstration, the drone autonomously “uses AI to detect, identify, and locate all potential threats it discovers.” The Scaleout Systems presentation stated that “all data is processed by dedicated onboard computing, allowing the system to run in real time without any external processing.”

Using its onboard AI, the drone automatically prioritized the highest-value target defined by the mission: an armored engineering vehicle. It then flew to the target and dropped explosives. A human operator could still control and direct the drone, but the system was also able to complete its mission without direct human commands.

Swedish Air Force test demonstrated resilient edge AI

In June, Scaleout Systems tested the technology at a Swedish Air Force base in Uppsala. The Swedish military has already obtained a license to use Scaleout Systems’ primary software platform.

The demonstration showed how forward-deployed computing nodes at military bases can continue running AI inference and active-learning processes after losing connectivity with a central computing node at Scaleout Systems’ research laboratory. When connectivity was restored, local AI model updates were shared with the central computing node.

Federated learning could support distributed military networks

This federated-learning strategy allows a distributed network of AI models to learn from aggregated data. Helander said the approach could eventually be scaled across geographic regions or countries. “In principle, cooperation between NATO members could be terminated,” he said.