Real-time validation of an AI-based solution for tactical air traffic complexity prediction and resolution

Abstract

This paper presents the real-time human-in-the-loop validation of ASTRA, an AI-based solution designed to predict and resolve 4D Areas of Relatively High ATC Complexity (4DARHACs) in congested, en-route airspace. Developed within the SESAR framework, ASTRA forecasts complex traffic events up to one hour in advance and proposes resolution strategies using flight level, speed and lateral clearances. A series of real-time simulations with Flow Management Position (FMP) operators, Air Traffic Control Officers (ATCOs) and ATCO supervisors evaluated the solution’s operational feasibility, human performance impact, and effects on capacity, efficiency, environmental performance, and safety. The results demonstrate ASTRA’s ability to predict and resolve 4DARHACs, and show that its recommended solutions reduce ATCO workload, increase en-route capacity and safety and, in most cases, lower fuel burn and CO2 emissions. The participants positively assessed the HMI and concept, while identifying integration, coordination, and scenario-realism improvements as key areas for future development.