AI-controlled drone swarm: Airbus, Quantum Systems and Spleenlab demonstrate autonomous swarm technology

Airbus Defence & Space, the German drone company Quantum Systems and the software company Spleenlab have achieved a decisive breakthrough in autonomous swarm technology in close cooperation. At its centre is a drone swarm coordinated by artificial intelligence (AI).

The results were presented at the Airbus Drone Center near Manching. The demonstration marked the high point and conclusion of an intensive research phase within the Bundeswehr (German Armed Forces) study KITU 2 “Artificial Intelligence for Tactical UAS” – meaning unmanned aircraft systems (UAS), explained in English as “Unmanned Aerical Vehicle”. The Bundeswehr commissioned the study in 2023. It was to produce innovative solutions for the AI-supported control of drone swarms and explore how far mixed UAS swarms with different capabilities can be coordinated by AI.

For the first time in the tests, a purpose-built mission AI took over the control and coordination of the UAS. The aim was for missions to run reliably even when radio interference occurs or individual drones fail completely.

With the new technologies, a few operators can effectively direct large swarms of autonomous UAS – even in highly dynamic, interference-prone environments.

The AI was trained in a highly specialised simulation environment using advanced “deep reinforcement learning” methods. Through continuous self-optimisation, it can refine its tactics and make more efficient and precise decisions in tactical operations.

Drone swarm remains operational even under interference

In the successful tests, the Vector and Scorpion UAS from Quantum Systems and two further multi-purpose drones from Airbus were trialled in swarm flight. Their reconnaissance data was merged in real time into a common operational picture that fed into Airbus's battle management system “Fortion Joint C2”. The Vector drones also showed that, with GPS jammed – “GNSS-denied”, i.e. without a global navigation satellite system (GNSS), as in Ukraine – they can carry out missions such as joint reconnaissance and target acquisition autonomously. The AI thus demonstrates that it can increase the resilience of the UAS to interference and ensure autonomous operation even under difficult conditions.

“We are not only expanding the technical capabilities of our drones. We are giving customers and users a real advantage in real-world scenarios. It is about protecting soldiers and increasing security. There will be no way around software-based and AI-supported systems for drone technology in the future.”
Sven Kruck, CRO & Managing Director at Quantum Systems

Foundation for future major programmes

The findings will be used to assess how behaviours learned in simulations can be transferred to real UAS. This will also examine whether, and to what extent, AI-controlled behaviour is superior to conventional manual control approaches.

The research results of the KITU2 study are to benefit, among others, autonomous systems for major Bundeswehr programmes such as the Main Ground Combat System (MGCS) and the Future Combat Air System (FCAS).