ATOS ASSOCIATION / 21 July, 2025 / Denmark
The study systematically confronts critical gaps in the existing operational concepts by addressing the following research questions :
- What are the essential constraints and information requirements for autonomous UAV operations in the final U4 stage of U-space, and what are the gaps when compared to EUROCONTROL’s Concept of Operations?
- How can a safe separation distance be determined for an uncooperative drone of an unknown model, and how can an autonomous system plan a maneuver to avoid it?
- How can conflict-free, 4-D trajectories be planned for multiple UAVs to minimize flight time for each aircraft?
Summary of Research Contributions
The research delivers three core contributions that form a progressive framework for autonomy. The first pillar establishes the foundational information needs for full autonomy (TRL 2) by systematically classifying operational constraints and identifying critical information gaps between the needs of an autonomous system and the services defined in the EUROCONTROL Concept of Operations. The second pillar introduces a novel method for ensuring safe separation (TRL 3) from one of the most severe threats: uncooperative drones of unknown models (UDUMs). This is achieved through the innovative MATHryoshka method, which calculates a dynamic, type-based safe separation distance. The third pillar presents a validated algorithm for operational efficiency (TRL 4), the Gen4jectory 2.0 algorithm, which plans conflict-free, time-minimized 4-D trajectories for multiple rotary-wing UAVs operating simultaneously.
Contribution I: Establishing the Foundational Information Needs for Full Autonomy (TRL 2)
The first contribution addresses the undefined information architecture for the final, fully autonomous U4 stage of U-space. To resolve this, a systematic review was conducted to identify and classify the essential constraints that autonomous guidance systems will face. This analysis spanned six critical domains: flight physics, trajectory computation, collision avoidance, communication, navigation, and surveillance (CNS), institutional rules, and mission types. By meticulously comparing the information needed to address these constraints against the services outlined in the latest U-space Concept of Operations, numerous high-priority information gaps were exposed.
Key examples of this missing information include the need for a dynamic, natural turbulence map for very-low-level flight; data on UAV wake vortex categories to prevent upsets from larger aircraft; a method to detect and classify on-surface dynamic obstacles (e.g., vehicles, pedestrians); and, most critically, a framework for identifying, classifying, and safely separating from non-communicating airspace intruders, such as uncooperative drones of unknown models (UDUMs).
The significance of this foundational work, which provides the first comprehensive blueprint for the information architecture required for U4, was validated by the U.S. Federal Aviation Administration (FAA). The FAA requested a formal presentation of these findings to inform and support its own UTM research and development efforts, underscoring the global relevance and applicability of the research beyond the European context.
Contribution II: A Novel Method for Safe Separation from Uncooperative Drones (TRL 3)
The second contribution confronts the severe mid-air collision risk posed by uncooperative drones of unknown models (UDUMs). Because the performance, maneuverability, and intent of a UDUM are unknown, traditional conflict detection and resolution methods are rendered unreliable, creating an unpredictable and unacceptable safety risk.
This research introduces the novel MATHryoshka method as a solution. The core innovation of this method is that it circumvents the need to identify the specific model of the rogue drone, which may be a custom-built or uncatalogued airframe. Instead, it leverages modern computer vision systems to identify the drone’s general type (e.g., fixed-wing, rotary-wing, parafoil). By analyzing a global UAV database, the method establishes the maximum performance envelope for each UAV type, including its maximum possible speed and maneuverability constraints.
The MATHryoshka method then applies a deterministic, physics-based model that calculates a dynamic safe separation distance under a “worst-case” assumption: that the UDUM is the fastest and most agile aircraft of its identified type and is actively maneuvering to cause a collision. This approach transforms an unknown, unpredictable threat into a bounded, manageable risk. It provides a safety philosophy grounded in verifiable physical limits rather than probabilistic models. By assuming the most dangerous scenario, the method generates a guaranteed safety bubble that is geometrically and physically sound, offering a clear and certifiable safety case that is far more likely to be accepted by regulatory bodies such as the FAA and EASA.
Contribution III: The Gen4jectory 2.0 Algorithm for Efficient, Conflict-Free Operations (TRL 4)
The third contribution addresses the challenge of planning time-minimized, conflict-free 4-D trajectories for multiple, simultaneous UAV missions in complex urban environments.
The Gen4jectory 2.0 algorithm was developed to solve this multi-agent planning problem. Its novelty lies in the unique fusion of three key components:
- A Simplified Physical Model: The algorithm employs a mass-point model that accounts for the most essential aerodynamic forces and UAV-specific performance data. This allows for computationally efficient yet physically realistic estimations of flight time between waypoints.
- Advanced Pathfinding: It utilizes the Theta* algorithm, a state-of-the-art, any-angle path planner. Unlike traditional grid-based methods, Theta* generates smoother, more direct routes by performing line-of-sight checks, resulting in shorter and more efficient flight paths.
Guaranteed Conflict Avoidance: To ensure safety, the algorithm reserves a 4-D spacetime volume for each UAV’s trajectory segment. It then uses the Separating Axis Theorem (SAT)—an exact geometric test—to perform an OBB-vs-OBB collision check against all other reserved volumes. This mathematically guarantees that no Loss of Separation (LoS) can occur between planned trajectories.
The algorithm’s performance was validated through hundreds of simulations in environments ranging from open airspace to dense urban settings. The results demonstrate that Gen4jectory 2.0 can successfully plan conflict-free trajectories for fleets of up to 100 UAVs, with computation time scaling in a predictable (near-quadratic in obstacle-free scenarios and superlinear in dense environments) manner. This confirms its viability for managing moderately dense air traffic.
Conclusion: Scientific Significance and Future Directions
In summary, this doctoral dissertation delivers a comprehensive and progressive framework for enabling autonomous UAV operations. It advances the field by moving from foundational theory (TRL 2), through analytical safety methods (TRL 3), to a practical, validated algorithm for multi-agent planning (TRL 4). The research provides tangible, physics-based solutions to safety-critical and efficiency-limiting problems that have, until now, hindered the path toward fully autonomous U-space operations.
The international significance of this work is evidenced by invitations from both the U.S. Federal Aviation Administration and Odense Robotics, Denmark’s national cluster for automation, to share these research findings.
This research should be viewed not as an end-point, but as a cornerstone for the next phase of development. The key future research directions are clear: building the global UAV performance database, potentially using computational fluid dynamics (CFD) as a cost-effective method, and integrating real-world complexities such as wind, weather, and sensor noise into the planning algorithms. This work lays the scientific groundwork necessary to create a truly safe, efficient, and autonomous drone ecosystem.
Authored by Dr. Ivan Panov, a seasoned aerial robotics expert with a Ph.D. in Aerial Robotics from Aarhus University and a Master’s degree (cum laude) in Flight Aerodynamics. Dr. Panov’s expertise spans autonomous guidance, 4-D trajectory planning, and flight safety, with a career that includes research at Aarhus University and project leadership roles in the aviation sector. This doctoral research provides a new framework for enabling autonomous Unmanned Aerial Vehicle (UAV) operations within the European U-space environment.
Having completed his Ph.D., Dr. Panov is now seeking a new position to apply his expertise in aerial robotics and autonomous systems.
Credit to: Dr. Ivan Panov and ATOS