Periodic Reporting for period 1 - CODA (COntroller adaptive Digital Assistant)
Reporting period: 2023-09-01 to 2024-08-31
Air traffic control is reputed as one of the five most stressful professions there is. Addressing the mental workload of controllers is therefore an important area of SESAR research and innovation and one that can be addressed using artificial intelligence. The project aims to develop a digital assistant capable of predicting future traffic, and assessing controllers’ stress levels and attention span, and whether they would be capable of handling the anticipated workload. The assistant would decide how to act, following an adaptation strategy: it may, for instance, increase the level of automation, enable additional AI-based tools, or request changes to the airspace (sector splitting).
Delivery of the state-of-the-art review, documented in D2.2 - OSED.
Delivery of the CODA concept and definition of use cases, documented in D2.2 - OSED.
Delivery of the CODA functional requirements and future roles, documented in D2.1 - FRD.
The first version of the task prediction model (v1), including its architecture, requirements, implementation, and validation has been completed.
Creation of the mental state prediction model with its architecture and requirements. The outcome is being reported in D3.2.
Regarding Models validation, Scenarios and data gathering methods have been defined, exercise scenarios have been modelled within the ESCAPE platform, initial data gathering methods (questionnaires, devices, CWP) have been prepared, data outputs analised, and software developed.
Delivery of the CODA Indexes identification, documented in D4.1 - Indexes description and integration with prediction models.
Wearable and reliable devices for the Validation activities have been identified and the CODA sensor selection and development documented in D4.1 - Indexes description and integration with prediction models.
The HMPE based on mental states, human-machine teaming, and ATCO’s performance have been defined ands documented in the D4.1 - Indexes description and integration with prediction models
Adaptation strategy guidelines have been integrated into D5.1: Adaptation and Human-AI Interaction Strategy and Teaming Playbook.
The content of the guidelines and HMI have been integrated in Deliverable 5.1—Adaptation and Human-AI Interaction Strategy and Teaming Playbook.
Initial data sources, including user activity, situation complexity, and aeronautical information, have been identified. The team has begun scenario building and conducted initial workshops, applying user-centric principles to iterate on low-fidelity prototypes
Initial drafts of the Human-AI Teaming Playbook have been produced, outlining different LOAs and defining how authority and responsibilities can be shared between human operators and AI systems. Exploring pre-set behaviours and goal-based control methodologies has begun, and preliminary feedback from stakeholders has been incorporated.
Data aggregation has started, and efforts are in progress to define communication protocols and consistency checks. Cognitive models are being integrated into the ESCAPE platform (WP3).
The validation plan has been generated. A first exercise (a workshop with relevant stakeholders) has been carried out to validate the CODA concept
Delivery of the CODA concept and definition of use cases, documented in D2.2 - OSED.
Delivery of the CODA functional requirements and future roles, documented in D2.1 - FRD.
The first version of the task prediction model (v1), including its architecture, requirements, implementation, and validation has been completed.
Creation of the mental state prediction model with its architecture and requirements. The outcome is being reported in D3.2.
Regarding Models validation, Scenarios and data gathering methods have been defined, exercise scenarios have been modelled within the ESCAPE platform, initial data gathering methods (questionnaires, devices, CWP) have been prepared, data outputs analised, and software developed.
Delivery of the CODA Indexes identification, documented in D4.1 - Indexes description and integration with prediction models.
Wearable and reliable devices for the Validation activities have been identified and the CODA sensor selection and development documented in D4.1 - Indexes description and integration with prediction models.
The HMPE based on mental states, human-machine teaming, and ATCO’s performance have been defined ands documented in the D4.1 - Indexes description and integration with prediction models
Adaptation strategy guidelines have been integrated into D5.1: Adaptation and Human-AI Interaction Strategy and Teaming Playbook.
The content of the guidelines and HMI have been integrated in Deliverable 5.1—Adaptation and Human-AI Interaction Strategy and Teaming Playbook.
Initial data sources, including user activity, situation complexity, and aeronautical information, have been identified. The team has begun scenario building and conducted initial workshops, applying user-centric principles to iterate on low-fidelity prototypes
Initial drafts of the Human-AI Teaming Playbook have been produced, outlining different LOAs and defining how authority and responsibilities can be shared between human operators and AI systems. Exploring pre-set behaviours and goal-based control methodologies has begun, and preliminary feedback from stakeholders has been incorporated.
Data aggregation has started, and efforts are in progress to define communication protocols and consistency checks. Cognitive models are being integrated into the ESCAPE platform (WP3).
The validation plan has been generated. A first exercise (a workshop with relevant stakeholders) has been carried out to validate the CODA concept
egarding the broader expected impact of the project, CODA results are expected to make a difference in impact beyond the immediate scope and duration of the project. In particular, the CODA project will produce the preliminary conceptual design of the ATCO Digital Assistance Tool, providing an understanding of how the tool’s basic principles will be used (Please refer to section 1.2 Methodology). In the long term, the CODA project will impact many of the research and innovation needs addressed in the SRIA to achieve the Digital European Sky programme. In detail, the CODA project will have an impact on:
- the Connected and Automated ATM. The Digital Assistance Tool will boost the level of automation in the ATM. This will contribute to achieving the European ATM Master Plan vision to reach at least level 2 (task execution support) for all ATC tasks and up to level 4 (high automation) for some of the tasks.
- Capacity-on-demand and dynamic airspace. The CODA system will allow a dynamic reconfiguration of resources (HUMAN-AI TEAMING) and new capacity-on-demand (ADAPTIVE AUTOMATION FOR TASK ALLOCATION AND EXECUTION) services to maintain safe, resilient, smooth and efficient air transport operations while allowing for the optimisation of trajectories, even at busy periods.
Artificial intelligence (AI) for aviation. The system's predictivity and prescriptibility will optimise its ability to identify potentially problematic solutions and correct them before an event occurs.
The CODA Project demonstrates a significant contribution to the realisation of the Digital European Sky vision (SESAR Phase D) about achieving:
fully scalable services supported by a digital eco-system, providing an enabler for adaptable systems able to respond and anticipate disruptions and problematic situations effectively
high and full automation (level 4/5), providing a concrete example of a system adapting the level of automation to the contextual condition and the states of operators, ensuring the best possible level of automation in the different conditions
We believe that the CODA project has a clear breakthrough potential in the long term (2028–2029) as it paves the way for adaptable systems that can improve efficiency and safety. We want to highlight that the CODA approach is implemented in this proposal to a specific use case focused on the ATCO. Still, it could also be applied to support other actors of the ATM system (e.g. network managers) and the aviation sector in general (e.g. airlines). Several target groups have been identified and reported in section 2.2.1 Stakeholders engagement.
The outcome and impact of the project will be:
Scientific. The new predictive model equipped with adaptive automation will contribute to scientific advances across and within different disciplines (Artificial Intelligence, neuroscience), creating new knowledge and reinforcing scientific equipment, instruments, and computing systems.
Economic/technological. The project will contribute to technological and economic development by bringing a new product into the market: the CODA Digital Assistance tool. The tool's characteristics and functionalities will increase ATM efficiency and profits and decrease costs by foreseeing future situations that, if not detected, could lead to inconveniences.
- the Connected and Automated ATM. The Digital Assistance Tool will boost the level of automation in the ATM. This will contribute to achieving the European ATM Master Plan vision to reach at least level 2 (task execution support) for all ATC tasks and up to level 4 (high automation) for some of the tasks.
- Capacity-on-demand and dynamic airspace. The CODA system will allow a dynamic reconfiguration of resources (HUMAN-AI TEAMING) and new capacity-on-demand (ADAPTIVE AUTOMATION FOR TASK ALLOCATION AND EXECUTION) services to maintain safe, resilient, smooth and efficient air transport operations while allowing for the optimisation of trajectories, even at busy periods.
Artificial intelligence (AI) for aviation. The system's predictivity and prescriptibility will optimise its ability to identify potentially problematic solutions and correct them before an event occurs.
The CODA Project demonstrates a significant contribution to the realisation of the Digital European Sky vision (SESAR Phase D) about achieving:
fully scalable services supported by a digital eco-system, providing an enabler for adaptable systems able to respond and anticipate disruptions and problematic situations effectively
high and full automation (level 4/5), providing a concrete example of a system adapting the level of automation to the contextual condition and the states of operators, ensuring the best possible level of automation in the different conditions
We believe that the CODA project has a clear breakthrough potential in the long term (2028–2029) as it paves the way for adaptable systems that can improve efficiency and safety. We want to highlight that the CODA approach is implemented in this proposal to a specific use case focused on the ATCO. Still, it could also be applied to support other actors of the ATM system (e.g. network managers) and the aviation sector in general (e.g. airlines). Several target groups have been identified and reported in section 2.2.1 Stakeholders engagement.
The outcome and impact of the project will be:
Scientific. The new predictive model equipped with adaptive automation will contribute to scientific advances across and within different disciplines (Artificial Intelligence, neuroscience), creating new knowledge and reinforcing scientific equipment, instruments, and computing systems.
Economic/technological. The project will contribute to technological and economic development by bringing a new product into the market: the CODA Digital Assistance tool. The tool's characteristics and functionalities will increase ATM efficiency and profits and decrease costs by foreseeing future situations that, if not detected, could lead to inconveniences.