Search

• Airspace availability for Urban Air Mobility is learned and predicted from data. • A probabilistic model predicts traffic patterns and their spatial confidence regions. • A lateral buffer is used for procedural separation from conventional traffic. • Dynamic airspace allocation increases accessibility by more than 5% from 3000 ft. • Airspace availability is more sensitive to the lateral separation buffer. Emerging Urban Air Mobility (UAM) operations are expected to introduce novel air traffic networks in metropolitan areas in order to provide on-demand air transportation services and alleviate ground congestion. Yet, metropolitan regions are typically characterized by complex and dense terminal airspace structure that accommodates arrival and departure traffic from large metroplex airports. Therefore, UAM operations are expected to be initially integrated into urban airspace without interfering with conventional operations and compromising current safety and efficiency levels. This paper presents a data-driven approach to identify and predict available urban airspace that is procedurally separated from conventional air traffic towards supporting UAM integration. We use historical aircraft tracking and meteorological data to learn the spatial distribution of air traffic in the terminal airspace and create a probabilistic traffic model to predict active traffic patterns and their spatial confidence regions given current operational conditions. We demonstrate the approach for the city of Sao Paulo and its closest commercial airport, Congonhas (CGH), in Brazil. The results show that leveraging the traffic flow dynamics to allocate the urban airspace dynamically is beneficial to increase UAM accessibility by more than 5% from 3000 ft. Moreover, airspace availability is found to be highly sensitive to the applied separation requirements, emphasizing the importance of leveraging advanced technologies to progressively make such requirements less stringent. [ABSTRACT FROM AUTHOR]