| Título: | Data-driven modeling of urban airspace availability for emerging air mobility operations |
| Autor: | João Vitor Turchetti Ribeiro |
| Programa: | Engenharia de Infraestrutura Aeronáutica |
| Área de Concentração: | Transporte Aéreo e Aeroportos |
| Orientador : | Mayara Condé Rocha Murça |
| Ano de Publicação : | 2023 |
| Curso : | Mestrado Acadêmico |
| Assuntos : | Controle de tráfego aéreo |
| t | Aeronaves de decolagem vertical |
| t | Mobilidade urbana |
| t | Espaço aéreo |
| t | Aprendizagem (inteligência artificial) |
| t | Método de Monte Carlo |
| t | São Paulo (cidade) |
| t | Transportes |
| Resumo : | Urban Air Mobility (UAM) is an emerging aviation ecosystem that will use novel highly automated aircraft such as Electric Vertical Takeoff and Landing (eVTOL) vehicles to meet the rising societal needs for more efficient and rapid urban mobility and cargo transportation. UAM is expected to introduce novel flight networks into already busy and complex airspace surrounding major cities and metropolitan regions, representing a major challenge for air traffic management. This work provides a data-driven approach to modeling the urban airspace availability for emerging UAM operations toward supporting their safe and efficient integration. Using historical flight tracking data, clustering analysis is first performed to learn the current patterns of urban airspace use by conventional traffic and identify the airspace volumes that are least constrained and best accessible for UAM flights. Meteorological data is then incorporated into the machine learning framework to create a probabilistic model of the spatiotemporal distribution of conventional traffic flows. This model enables the prediction of active airport arrival/departure patterns and the resulting airspace availability for UAM given dynamic operational conditions. The data-based approach is demonstrated for the Sao Paulo metropolitan area, which is the largest in Brazil and a promising market for UAM. It allowed for a detailed characterization of current traffic patterns in the Sao Paulo terminal airspace as well as for predictions of their occurrence with an accuracy higher than 95\%. Moreover, it enabled a high-fidelity simulation of the Sao Paulo urban airspace operations under different scenarios of UAM integration, allowing for the quantification of performance trade-offs for each scenario. The simulation outputs emphasize the importance of dynamically managing the urban airspace through the allocation of geofences for protecting existing flight procedures when they are active without over-restricting the UAM accessibility. With an appropriate spatial resolution, we find that dynamic geofencing is able to provide a significant reduction of more than 80\% in the number of conflicts at a reasonable efficiency loss of less than 10\% for UAM flights. The results of this work provide new insights regarding the value of dynamic urban airspace management as well as novel predictive capabilities for its implementation in support of UAM. |
| Data de Defesa : | 12/07/2023 |
| Texto na íntegra : | [Visualizar] |