An AI-based Holistic Dynamic Framework for a safe Drone’s Operations in restricted and urban areas

Projektinformationen

AI4HyDrop

ID Finanzhilfevereinbarung: 101114805

DOI

10.3030/101114805

Projekt abgeschlossen

EK-Unterschriftsdatum 8 Juni 2023

Startdatum 1 September 2023

Enddatum 28 Februar 2026

Finanziert unter

• Climate, Energy and Mobility

Gesamtkosten

€ 1 993 112,50

EU-Beitrag

€ 1 993 112,50

Koordiniert durch

UNIVERSITETET I SOROST-NORGE
Norway

Periodic Reporting for period 1 - AI4HyDrop (An AI-based Holistic Dynamic Framework for a safe Drone’s Operations in restricted and urban areas)

Berichtszeitraum: 2023-09-01 bis 2024-08-31

Zusammenfassung vom Kontext und den Gesamtzielen des Projekts

With an increasing number and diversity of potential drone operations, managing the airspace to accommodate these drones will become an increasingly sophisticated task, especially in densely populated urban areas encompassing restricted zones with dynamic environmental and operational influences. Due to the associated higher probability of conflicts, and ultimately collisions, such areas require management of dedicated structured airspace, operations, and services to help mitigate these potential hazards. Several projects are currently working on defining ConOps for U-space services. Corus and Corus-XUAM have defined a possible capabilities model, such as airspace organization and services. However, a holistic framework is necessary to create an effective and efficient flow of information between the various capabilities in order to systematically organize the airspace usage. Such an automated Air Traffic Management System will be essential for the introduction drone operations at scale. AI4HyDrop evaluates the various stakeholder needs, delivering validated concepts, defining a methodology for an airspace structure organization and associated U-space services. The framework considers the information from other services such as meteorological and separation provision, which can then be used for flight planning approval process, prioritization. In addition, essential elements such as surveillance and contingency planning can be addressed. The framework incorporates various AI based tools and associated information flows necessary to addresses the complexity, safety and scalability required for implementing such U-space services. The proposed framework represents a digital step change in ATM, using AI as a means to overcome many critical barriers foreseen in the introduction of automated U-space services. The findings could later be expanded to support general airspace management.

Arbeit, die ab Beginn des Projekts bis zum Ende des durch den Bericht erfassten Berichtszeitraums geleistet wurde, und die wichtigsten bis dahin erzielten Ergebnisse

We conducted online and physical stakeholder workshops. The outcomes of these workshops are pivotal for the AI4HyDrop project, which capitalizes on the collective expertise of stakeholders and expert advisory boards to tailor a framework addressing current challenges and future needs of U-Space systems. Notably, these findings will be distilled into targeted project requirements, ensuring that the decentralized intelligence advocated by AI4HyDrop remains attuned to real-world complexities, prioritizing safety, and efficacy in urban airspace management.

We presented one of the initial result in conference of Control, Automation, Robotics, Optimization, Decision, Cybernetics, Computer Science and Information Technologies (CODIT) on 01-04 July 2024 at Valetta, Malta. The title of the paper is "A Survey of AI-based Models for UAVs’ Intelligent Control for Deconfliction". The main topic is about deconfliction or separation management problem. In the research, we explore the use of machine learning to develop an intelligent control system to resolve drone deconfliction.

A stand-alone micro-scale AI wind model is being developed for one of the cases (Prague City). The micro-scale AI wind model is parameterized on meso-scale wind speed and meso-scale wind direction within the city. The AI model development involved selecting the main business area around Prague City where a vertiport may come up and generating the training database to train the AI model. This training database is generated using 32 hight-fidelity computational fluid dynamics (CFD) at different wind speeds and direction. It is subjected to unsupervised machine learning to identify dominant spatial patterns for the AI model. The GitHub for the wind and turbulence model is available. The other model for Oslo city is in preparation.

Fortschritte, die über den aktuellen Stand der Technik hinausgehen und voraussichtliche potenzielle Auswirkungen (einschließlich der bis dato erzielten sozioökonomischen Auswirkungen und weiter gefassten gesellschaftlichen Auswirkungen des Projekts)

N/A. The production of results is in progress and we will provide the complete results at the end of the project.

Drone detection test