SafeTeam progresses on the human factors aspects on the use of digital assistants to aviation, including a deeper understanding on the technology and processes that will facilitate the adoption of AI tools and integration into operations, enhancing human cognitive abilities and potentially automation. The project will also look into approval and certification issues, concretely on aspects related to the human ability to operate sophisticated AI tools and explainability of AI operations.
Facilitate a human-centric approach to automation and its integration into a wide spectrum of air traffic operations.
Propose methodologies for the assessment and monitoring of the system performance, with special focus on safety and resilience, to enable seamless Human-Machine cooperation.
Progress in the development of Digital Assistants for aviation operations in support of human performance for all development, testing, validation and verification phases.
Support the definition of regulatory and certification requirements for automation tools to address market needs and societal acceptance.
Novel approach to human factors, safety and resilience
Organisational and regulatory preparedness
Case Study 1 – Digital assistance for area ATC (enroute)
Digital assistants have the innovation potential to transform the operational room by anticipating complex traffic scenarios, forecasting airspace sector workload and allowing more efficient and environmentally-friendly routing—all while increasing safety.
SafeTeam will incorporate human factors expertise according to the targeted level of autonomy and corresponding controller-machine tasks distribution. The safety-critical level will be considered in the definition of the use case and its implementation. The validation exercises will monitor how the information provided is ingested by the controller as well as its impact on the workload, situational awareness, and other safety indicators.
Case Study 2 – Unstable Approaches prediction for aircraft cockpit (UA)
This case study aims at providing predictions to the crew in the cockpit when there is high probability that the flight becomes unstable.
The design principles, distribution of automation roles and validation exercises proposed in SafeTeam will make this use case achieve TRL6. Meanwhile, the performance during these exercises will be monitored and assessed by human factors experts. Aspects like the concrete threshold, or how much heads up needs to be given to crew in order for predictions to be useful, will be defined in collaboration with operational experts (pilots and safety managers) and human factors professionals, as well as the appropriate interface defining how the information needs to be showed.
Case Study 3 – Digital Assistance for evidence-based training (EBT)
An intelligent agent will be developed to observe sensor data from real and virtual aircraft alongside other environmental data sources (cockpit microphones, eye tracking, video feeds).as the actuator of the digital assistant in this case study. The agent will automatically interrogate these data sources with the aim of supporting instructors in their assessment of pilot competencies.
An agent that can analyse human data and pilot competencies can be considered to be highly innovative in the field of pilot training. This agent will help bring concordance across instructors by providing objective data for them to engage with crew and identify and resolve root causes to issues.
D2.1 Design principles for digital assistants and HF assessment methodology
D4.1 Human-machine collaboration in en-route operations
D2.2 Human performance and automation integration report
D4.2 Human-machine collaboration in destabilised approaches
D3.1 Human factors design principles for an en-route digital assistant
D4.3 Human-machine collaboration in Evidence Based Training
D3.2 Human factors design principles for a stabilised approach digital assistant
D5.1 Bow-Tie analysis for AI case studies
D3.3 Human factors design principles for Evidence Based Training
D5.2 Regulatory cohesion for digital assistants implementation
This project has been funded by the European Union under Grant Agreement 101069877.