AI drone swarm: Thales tests COHESION demonstrator in flight trials

Nobody today disputes that drones bring operational value on the battlefield. Two factors, however, still limit their effectiveness: each drone usually needs its own operator, and a secure, resilient data link must be maintained throughout the mission. Thales is therefore researching artificial intelligence for the AI drone swarm. Flight tests took place on 16 October, in which the Thales COHESION demonstrator showed how a high degree of autonomy in drone swarms could be achieved with AI and intelligent agents.

The demonstrator's system architecture is designed so that operators can adapt the level of autonomy of their swarms to the requirements of each mission. Thales expects this to deliver “an unprecedented level of flexibility”. If individual drones and/or entire swarms operate autonomously, a permanent data link to the control station is no longer required – so signal jamming cannot harm these drones.

Because the drones perceive and analyse their local environment, they can share target information, assess hostile intentions and set mission priorities. Collaborative tactics and optimised flight paths also accelerate the OODA loop (observe, orient, decide, act).

The human remains decisive in the drone swarm

What is to remain unchanged, however, is that the most important decisions rest with the human operators. Human oversight at all times is guaranteed by trustworthy, cyber-secure human-in-the-loop AI.

Hervé Dammann, Executive Vice President of the Thales Land and Air Systems division, stressed: “We are proud to develop innovative solutions combined with strong ethical values. Our solutions are demonstrable, applicable, incremental and deployable, and act as a force multiplier without increasing the cognitive load on operators, guaranteeing that they retain their central role in the decision-making process.”