Intelligent vertiport traffic flow management for scalable advanced air mobility operations

Advanced air mobility (AAM) operations will pose new challenges that require innovative air traffic management (ATM) and uncrewed aircraft system (UAS) traffic management (UTM) solutions. Notably, emerging vertiports must support vertical take-off and landing (VTOL) vehicles, on-demand AAM services, denser airspace volumes, and dynamic airspace structures. Additionally, traffic flow management systems must cater for stricter flight envelopes, micro-weather variations, small uncooperative aerial objects, limited vertiport occupancy, and battery restrictions of electric vehicles. This requires large volumes of unlabelled data that conventional algorithms cannot effectively process in a timely manner. This work thereby proposes a data model for vertiport traffic management, and investigates intelligent solutions to leverage this vast data infrastructure. It considers on-demand vertiport flight authorisation as a demonstrative use-case of emerging AAM requirements, and proposes a data model aligned with safety-layers and corridor-based airspace proposals in several global AAM concept of operations (ConOps). On-demand scheduling of electric VTOL (eVTOL) aircraft is first formulated as a constrained optimisation problem, and solved using mixed-integer linear programming techniques. The limitations of this approach are subsequently addressed through a deep reinforcement learning (DRL) solution that is quicker and more robust to system uncertainty. This investigation thereby proposes a pathway towards scalable, intelligent and multi-agent systems for AAM resource management and optimisation.