During the last week of May, ASTRA – an EU exploratory research project under SESAR – reached an important milestone with the successful execution of its third and final validation exercise. This exercise consisted of a human-in-the-loop real-time validation of an AI-based solution that was developed by the project to improve Air Traffic Management, particularly to predict and resolve complex traffic events (also known as hotspots) in busy enroute airspace. This solution has the potential to reduce Air Traffic Controller workload, increase airspace capacity, and improve fuel efficiency and safety.
The exercise was carried out at the facilities of Skyguide – the Swiss Air Navigation ¸£ÀûÔÚÏßÃâ·Ñ Provider – with the participation of 10 experienced operational staff members – including Air Traffic Controllers, Flow Management Position personnel, and Supervisors – who made use of the ASTRA solution in various simulated scenarios with complex air traffic. The aim of this exercise was to demonstrate the complete solution in a representative environment; obtain feedback regarding operational feasibility, human-machine interaction, and other aspects; and measure the impact of ASTRA on end user workload and other performance areas.
This exercise has been the culmination of around 21 months of research and development by the project’s five partners: the Institute of Aerospace Technologies (IAT) at the University of Malta; (Spain); (Italy); (Switzerland); and (Switzerland). The primary role of the IAT in ASTRA has been to develop the part of the solution that provides end users with suggestions – in the form of air traffic clearances – to resolve complex traffic events. This has been achieved through the application of deep reinforcement learning techniques.
The project – which is being coordinated by Dr Inġ. Jason Gauci from the IAT – will now focus on the compilation and analysis of the results of the final validation exercise; and the communication and dissemination of these results through publications, presentations and demonstrations at various aviation-related events and conferences.
For further information about ASTRA, please visit the and follow .
The exercise was carried out at the facilities of Skyguide – the Swiss Air Navigation ¸£ÀûÔÚÏßÃâ·Ñ Provider – with the participation of 10 experienced operational staff members – including Air Traffic Controllers, Flow Management Position personnel, and Supervisors – who made use of the ASTRA solution in various simulated scenarios with complex air traffic. The aim of this exercise was to demonstrate the complete solution in a representative environment; obtain feedback regarding operational feasibility, human-machine interaction, and other aspects; and measure the impact of ASTRA on end user workload and other performance areas.
This exercise has been the culmination of around 21 months of research and development by the project’s five partners: the Institute of Aerospace Technologies (IAT) at the University of Malta; (Spain); (Italy); (Switzerland); and (Switzerland). The primary role of the IAT in ASTRA has been to develop the part of the solution that provides end users with suggestions – in the form of air traffic clearances – to resolve complex traffic events. This has been achieved through the application of deep reinforcement learning techniques.
The project – which is being coordinated by Dr Inġ. Jason Gauci from the IAT – will now focus on the compilation and analysis of the results of the final validation exercise; and the communication and dissemination of these results through publications, presentations and demonstrations at various aviation-related events and conferences.
For further information about ASTRA, please visit the and follow .
