This European Defence Fund (EDF) topic supports the participation of multiple consortia in a common technological challenge focused on **AI-based tactical situational awareness** using **swarms of small unmanned ground vehicles (UGVs) and unmanned aerial vehicles (UAVs/drones)**. The general objective is to help protect soldiers by enabling robots and drones to provide an accurate, real-time understanding of the tactical environment. Building on and complementing previous work on hidden threat detection, the topic seeks integrated solutions that combine intelligent unmanned platforms with advanced sensing technologies and AI to detect, classify, and prioritise threats, so that threats can be avoided or neutralised.

A central feature of this topic is that funded consortia must test their solutions under realistic conditions in a **quantitative, objective, and comparable** way. To enable this, each selected consortium will benefit from a **common test environment** and must participate in the evaluation campaigns organised by the challenge organisers. The specific objective is to assess how well AI can support tactical situational awareness in complex, dynamic environments, by integrating cutting-edge technologies into unmanned collaborative platforms and demonstrating their performance in realistic scenarios with real-time insights for human decision-making.

Proposed systems are expected to cover: autonomous deployment of heterogeneous swarms to gather data efficiently; real-time data fusion across multiple sensor modalities (e.g., visual, radar, acoustic, environmental) into a unified situational picture; AI-driven analysis using advanced machine learning to detect patterns, identify and prioritise threats or areas of interest; dynamic adaptation to changing conditions such as weather, lighting, or obstacles; and a human-machine interface that is intuitive and user-friendly, enabling operators to receive alerts, interact with the system, and potentially receive automatically suggested actions/plans to be validated by the operator.

Within the scope, proposals must address technological solutions for automatic detection and characterisation of threats in complex ground-application environments, combining advanced sensors, information fusion, and unmanned ground and aerial systems to extend detection capabilities. The solution must be measurable and evaluable in the challenge’s test environment. Proposals should include clear criteria to assess work package completion, explicitly including: (i) participation in the test campaigns; (ii) delivery of sensor data collected during field tests; and (iii) delivery of descriptions of the systems submitted for testing. Consortia should also explain how they will harness synergies and complementarity with other activities planned, ongoing or completed in autonomous drone navigation.

Funding eligibility is limited to research-oriented activities that generate knowledge (mandatory) and, optionally, integrating knowledge, studies, and design activities. Activities such as system prototyping, testing, qualification, certification, and lifecycle-efficiency developments are ineligible under this topic. Mandatory tasks include research on new approaches for advanced threat detection and classification (combining AI and traditional techniques), research on swarming and collaborative coverage planning for heterogeneous UAV/UGV swarms in dynamic and contested scenarios, and full participation in the evaluation campaigns (including planning discussions, experimental field and data-based campaigns with navigation performance measurements, sensor-data collection and sharing, and debriefing workshops).

Functional requirements emphasise accurate threat detection and recognition for known threat classes, with the option to detect unexpected new threats or conditions; robust navigation/localisation with and without GNSS; measurement using protocols and metrics aligned with the preliminary evaluation plan; sensor data recording for full replay and experiment reproduction, including describing which data classes can be shared; and relevance for future operational integration with strong attention to SWaP-C (size, weight, power–cost) and distributed computing across swarm members, allowing a mix of low-cost and more powerful platforms. Additional desirable features include jamming-resistant intra-swarm communications, the ability to compare detected information with known domain data to resolve conflicts or raise alerts, and algorithms for optimised task allocation among swarm members to maximise autonomy and endurance.

Opening: 22.01.2026
Deadline(s): 29.09.2026