JARVIS

Main Goals of the Project

The rapid spread of automation and Artificial Intelligence (AI) is accelerating the adoption of autonomous systems across multiple industries. This trend, known as digitalisation, is also profoundly transforming aviation. In a context marked by the increasing complexity of the entire aviation ecosystem — including aircraft, air traffic control (ATC), and airports — digitalisation introduces innovative solutions in the form of Digital Assistants (DAs). These systems work in synergy with human operators (pilots, air traffic controllers, and airport operators), supporting operational tasks and helping ensure both safety and economic efficiency in complex scenarios.

The JARVIS Consortium, led by Collins Aerospace, aims to develop and validate three Air Traffic Management (ATM) solutions:

• an Airborne Digital Assistant (AIR-DA, Technology Readiness Level TRL4),
• an ATC Digital Assistant (ATC-DA, TRL4),
• an Airport Digital Assistant (AP-DA, TRL6).

The AIR-DA is designed to increase the level of automation in the flight deck and, through AI-based functionalities, acts as an enabler for reduced crew operations and, in the longer term, single-pilot operations. Its adoption will allow pilots to handle complex scenarios without compromising safety and security, while reducing workload.

The ATC-DA will enhance automation in control towers, with benefits for both environmental Key Performance Indicators and airspace capacity management, enabled by AI-based technologies.

Finally, the AP-DA will increase automation in airports, improving safety and security, particularly in intrusion detection scenarios.

The adoption of AI-driven technologies in the aviation ecosystem represents a compelling opportunity but also introduces significant challenges. JARVIS addresses three key aspects common to all three Digital Assistants: (i) assured AI design, to ensure trustworthy, explainable, safe, and ethical decision-making algorithms; (ii) Human-AI Teaming, to develop human-centric solutions that maximise collaboration between humans and autonomous systems; and (iii) big data and cloud infrastructures, required to support the transition from centralised architectures to distributed edge-to-cloud models.

CIRA Activities in the Project

Within the JARVIS project, CIRA played a key role in the development and validation of the ATC Digital Assistant (ATC-DA), structured around the following main activities:

• Leadership of Solution 0365, with responsibility for the development and TRL4 validation of the ATC-DA, the digital assistant supporting air traffic control operations.
• Design and implementation of the Conflict Resolution Advisor functionality within the ATC-DA, based on Artificial Intelligence algorithms, in particular Reinforcement Learning techniques, for generating operational solutions for aircraft conflict resolution in the en-route phase. The activity included the development of the corresponding software prototype and its integration with air traffic control support systems provided by Leonardo and ENAV.
• Validation of the solution through real-time simulations with a human-in-the-loop approach, involving air traffic controllers at the ENAV laboratories in Ciampino (Rome).
• Contribution to cross-cutting research activities within the project in the areas of AI Assurance and Human-AI Teaming.
• Execution of a research activity within the AI Assurance domain, in collaboration with EASA, focusing on generalization techniques in Reinforcement Learning models, aimed at improving the robustness and reliability of performance in complex and variable operational scenarios.

General Project Information

Funding Programme: HORIZON-JU-RIA - HORIZON JU Research and Innovation Actions

Start Date: 01/06/2023; End Date: 31/05/2026

Coordinating Organization: Collins Aerospace Ireland, Limited

Partners: DeepBlue, NAIS, CIRA, NLR, Indra, ENAIRE, CRIDA, Athens International Airport, AIRBUS, DLR, ENAV, Leonardo, Boeing, ENAC, EUROCONTROL, Swedavia Airports

Website: https://www.sesarju.eu/projects/JARVIS

CIRA Contact: Gabriella Gigante, g.gigante@cira.it