ACHIEVING HUMAN-MACHINE COLLABORATION WITH ARTIFICIAL SITUATIONAL AWARENESS

Project Information

AWARE

Grant agreement ID: 101167442

DOI

10.3030/101167442

EC signature date 28 June 2024

Start date 1 June 2024

End date 30 November 2026

Funded under

• Climate, Energy and Mobility

Total cost

€ 1 940 625,00

EU contribution

€ 1 940 625,00

Coordinated by

SVEUCILISTE U ZAGREBU FAKULTET PROMETNIH ZNANOSTI
Croatia

Periodic Reporting for period 1 - AWARE (ACHIEVING HUMAN-MACHINE COLLABORATION WITH ARTIFICIAL SITUATIONAL AWARENESS)

Reporting period: 2024-06-01 to 2025-05-31

Summary of the context and overall objectives of the project

The necessary introduction of artificial intelligence into air traffic management brings many advantages, but also challenges. AI agents are joining air traffic control teams as separate members, capable of processing vast amounts of data and providing insights and support to their human team members. In these interactions, human air traffic controllers understand the machines - how they operate, what their mission is, how best to use their strenghts, and how to mitigate their limitations. Machines, for now, lack the understanding of what the humans are doing, thus not allowing them to fulfill their role and full potential in human-machine teams.

Building on the results of a previous SESAR project, which demonstrated that an AI system is capable of using air traffic information to build an understanding of the traffic situation (i.e. artificial situational awareness), project AWARE aims to broaden that awareness by including human controllers' goals and intent. It would do so by tracking visual attention, but improved by including other inputs such as how the controller interacts with their work station. By also forming a complete picture of the traffic situation, the system would also be able to detect if the human controller is experiencing loss of situational awareness and assist them in bringing them back "into the loop". This would allow the system to provide better, human-centric support to ATCOs. By modifiying its outputs to their needs, it could improve their performance and reduce workload in complex traffic situations. ATCOs would be able to modify how support from the AI system manifests, or more precisely which actions the asisstant completes. It is also of interest to see if the benefits of using an AI assistant may be transferred to other roles within air traffic management, beyond en-route air traffic control.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

All developmental work in the first period was based in theoretical research. Experimental scenarios were developed in cooperation with ATCOs and SMEs. Some will serve to provide human-in-the-loop simulation data for modules and ATCO baseline values, while others were developed to assess and validate the system. System modules' functions have been thoroughly examined to best support ATCOs while utilising the new intent information. Plans have been developed for the remainder of the project, highlighting the most important research questions and keeping any changes within the research scope of the project.

System-wise, it was necessary to introduce changes into the artificial situational awareness (ASA) system architecture. After the architecure was updated, it was crucial to connect the system to its primary data source (the Polaris ATM System) and to create space for data not utilised in project AISA - eye-tracking, HMI interactions, etc. Simultaneously, work was done on the modules which will help the system to create artificial situational awareness, assess ATCO intent and workload, provide support, and perform chosen ATCO tasks.

AWARE ConOps is completed and submitted, describing how the project aims to build on previous work and findings by improving the ASA system. Planned improvements focus on determining ATCO intent, building a human-machine interface, and introducing adaptable automation. New data sources (intent/HMI/prioritization modules) will introduce meta-data which was not available in project AISA, so there are changes which will be introduced into the Data Management Plan when those modules are finished. All completed and upcoming research is laid out in two versions of the Exploratory Research Plan. Most other deliverables have been submitted for comments, while others are under consideration of the consortium and should be available soon.

Current work is divided into developing:

1) Modes of human-machine collaboration;
2) A method to track attention and intent;
3) ATCO supporting tools.

A necessity for further work is access to air traffic situation data. Data will be exported from an ATC simulator, so necessary adjustments - for the chosen airspace and module needs - are being introduced. Data samples are continuously provided to all partners who are working on software development to get feedback and improve information flows. Data preparations will culminate in data collection experiments, providing the final piece for module development.

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

General objective of project AWARE is the development of an AI assistant application, created to support ATC operations. The support comes in the form of a platform which employs enhanced artificial situational awareness - developed beyond the scope of project AISA to include ATCO intent and goals. The platform should include a method for adaptable selection and implementation of (supporting or direct) actions.

These results would directly affect safety of ATC operations by providing an additional safety net. An AI application capable of assessing intent could help to keep air traffic controllers in the loop or provide support to return them to the loop in case of loss of situational awareness. This would decrease activations of other safety net alerts.

By supporting ATCOs in completing tasks or by autonomously completing some tasks, the AI assistant application would decrease the human controllers' workload, allowing them to more easily handle higher amounts of traffic. This would result in higher en-route capacity.

Improving ATC processes would positively affect both environmental and cost concerns. Better traffic handling is directly connected to lower emissions via a more efficient aircraft utilization - either by enabling more direct routes or higher flight levels. Better handling naturally increases the overall efficiency of the system.

By using an actual ATM system (compared to a research-focused one), this project provides a more direct path towards higher TRL levels. Future avenues should consider more closely integrating the artificial situational awareness system with the underlying platform. Further development could include other parts of ATC systems (e.g. voice communication, other ATCO positions). Providing support in a manner more familiar to the ATCOs would be beneficial for system acceptance, a point highlighted within the results of project AISA.