Cooperative threat engagement with heterogeneous drone swarms

Important results from the project

To investigate methods that we believe are essential for the development of unmanned and autonomous drones, able to act in swarms, cooperate and interact with each other, in order to deal with an emerging threat. I.a. to develop technologies for distributed decision-making, task allocation and threat localization in real-time, in a dynamic environment. We also develop descriptive models for how a threat engagement mission should be designed for a drone swarm, as well as how discrete tasks in such a mission should be defined and realized using machine learning and behavior trees.

Expected long term effects

The project´s results demonstrate how to develop and demonstrate a drone swarm-based threat-engagement capability. The result harnesses the potential of small, inexpensive drones in combination with model-based development, machine learning and open-source software to develop an effective defense capability. This capability has been evaluated in a simulator where relevant KPI´s related to the design of the mission have been used. The results show that the capability is able to defeat a threat in the form of a drone threspassing a protected area in 94% of the simulated missions.

Approach and implementation

There is a Swedish part and a Brazilian part. The work has followed a project plan (statement-of-work) which was signed by both parties and which roughly stipulates how the work is to be carried out through seven work packages carried out in parallel by each party. A schedule stipulates when each work package must be completed and when each work package´s deliverables must be presented and then exchanged in the form of a written report with the other party. There is an industrial and an academic part in each country. On the Swedish side, SAAB AB and Mälardalen University are found.