In the forthcoming years, the role of highly automated and autonomous systems, including AI powered technologies, is going to increase at a fast pace in complex socio-technical systems. The ATM domain is not an exception. A recent study carried out upon request of the SESAR Joint Undertaking (JU) highlighted that higher levels of automation supporting air traffic controllers’ workload and reducing their stress are key for a future-proofed ATM system. Nevertheless, although it is generally agreed that the future of the ATM system will evolve towards higher levels of automation, a shared vision emerged as needed in order to develop a research roadmap with a breakdown of specific research actions. Based on this consideration, a long-term vision and research roadmap for automation in ATM was then detailed, as a basis for the definition and coordination of future research activities.
This vision is coherent with the one expressed by EASA in its AI Roadmap. In the document the agency traces the way for a human centric-approach to aviation and proposes a classification of AI/ML applications in three levels, considering the degree of oversight of a human on the machine: assistance to human, human-machine collaboration, and more autonomous machine. While discussing the implication of AI for the aviation sector, the roadmap identifies some key challenges that the aviation community shall address to profit by the disruptive power of highly automated systems.
The Strategic Research and Innovation Agenda (SRIA) for Digital European Sky highlights certification and standards as key tools to accelerate innovation. It emphasizes developing new validation and certification methods for advanced automation that ensure transparency, legal compliance, robustness, and stability in an AI-driven, human-centric ATM environment. SRIA identifies three priority areas for a new certification approach: “capacity on demand and dynamic airspace,” “U-Space and urban air mobility,” and “Artificial intelligence for aviation.”
Against this background, the HUCAN project addresses existing gaps by developing a novel certification approach for ATM-related airborne and ground systems with advanced, human-centric automation, including AI-based solutions. This approach considers not only technical reliability but also the impact on human performance, human-system interaction, and societal, ethical, and value-based factors. Additionally, it aims for a unified framework that accommodates both deterministic and non-deterministic algorithms and adapts to varying levels of software automation and autonomy.
The novel approach will support both the approval/certification process and the design phase of highly automated technologies, proposing to this end the development of two interconnected and tangible products: (1) a new holistic and unified certification method for systems based on advanced automation; (2) a set of suitable guidelines and associated toolkit for streamlining the development of highly automated and AI-powered technologies, targeted to manufacturers and developers of concepts based on advanced human-centred automation. This high-level aim will be reached by means of measurable objectives: (1) Review the most prominent trends and challenges in advanced automation applications in different domains, with particular focus on ATM. The different solutions (including those based on AI) will be considered in light of the degree of autonomy they are able to introduce in the procedures. (2) Analysis of legal and regulatory features and critical issues of automated/AI technologies, including an assessment of the current certification systems, highlighting their potential and limits. (3) Design novel methods and procedures of certification of ATM-related systems based on high levels of automation, which can encompass all the elements affected by the systems' operativity, and thus including the technical aspects as well as legal and regulatory requirements and safety provisions. (4) Develop a set of suitable guidelines and associated toolkit for the development of automation and AI technologies, with the aim of streamlining the certification/approval thereof.
