Using Reinforcement Learning and ADS-B Data for the Development of a Sense-and-Avoid System to the Integration of Autonomous eVTOL in the São Paulo Urban Airspace

Ulisses Ghidoni; Lucio Vismari; Joæo Batista Camargo J£Nior

doi:10.29327/xxi-sitraer-2024.887985

Anais do SITRAER · XXI SITRAER 2024

Anais do SITRAER – XXI SITRAER (2024)

Using Reinforcement Learning and ADS-B Data for the Development of a Sense-and-Avoid System to the Integration of Autonomous eVTOL in the São Paulo Urban Airspace

Ulisses Ghidoni
Lucio Vismari
Joæo Batista Camargo J£Nior

University of São Paulo

Baixar PDF DOI: 10.29327/xxi-sitraer-2024.887985 Também publicado na Even3

Evento: XXI SITRAER – Fortaleza, CE, Brasil, 2024
Volume: Anais do XXI Simpósio de Transporte Aéreo – SITRAER 2024 · ISBN 978-65-272-1008-5 · DOI 10.29327/xxi-sitraer-2024
Páginas: 515–522
Editora: Even3 Publicações
Editores: Francisco Heber Lacerda de Oliveira, Fabiano Rocha, Giovanna Miceli Ronzani Borille, Evandro José da Silva, Mauro Caetano

Abstract

The growing demand for air transportation in São Paulo requires advances in the air traffic system for the integration of autonomous electric vertical take-off and landing aircraft (eVTOLs). Based on the success shown using Artificial Intelligence (AI) in autonomous air combat simulation scenarios, there is great potential for developing an AI-based sense-and-avoid systems using Automatic Dependent Surveillance - Broadcast (ADS-B). This paper introduces an initial discussion about the authors ongoing research project that are using simulation and Reinforcement Learning (RL) algorithms to train an AI-based agent to detect and avoiding (resolving) air traffic conflicts based on situational awareness messages obtained from ADS-B system. In the first attempt to develop a scenario to evaluate the tools, X-Plane's characteristic of only allowing real-time simulations made it difficult to study the reward functions and the variables that could be observed. This work aims to contribute to autonomous urban air mobility solutions in Brazil, especially in highly demanded and non-segregated airspaces like São Paulo Helicontrol region.

Keywords: eVTOL; Reinforcement Learning; ADS-B; Sense-and-avoid; Urban Air Mobility

Como citar

ABNT

GHIDONI, Ulisses; VISMARI, Lucio; J£NIOR, Joæo Batista Camargo. Using Reinforcement Learning and ADS-B Data for the Development of a Sense-and-Avoid System to the Integration of Autonomous eVTOL in the São Paulo Urban Airspace. In: SIMPÓSIO DE TRANSPORTE AÉREO, 21., 2024, Fortaleza. Anais [...]. Recife: Even3, 2024. p. 515-522. ISBN 978-65-272-1008-5. DOI: 10.29327/xxi-sitraer-2024.887985. Disponível em: https://sitraer-sbta.org/anais-2024-887985.html.

APA

Ghidoni, U., Vismari, L., & J£Nior, J. B. C. (2024). Using Reinforcement Learning and ADS-B Data for the Development of a Sense-and-Avoid System to the Integration of Autonomous eVTOL in the São Paulo Urban Airspace. In F. H. L. Oliveira, F. Rocha, G. M. R. Borille, E. J. Silva, & M. Caetano (Eds.), Anais do XXI Simpósio de Transporte Aéreo – SITRAER 2024 (pp. 515–522). Even3. https://doi.org/10.29327/xxi-sitraer-2024.887985

BibTeX

@inproceedings{ghidoni2024using,
  author = {Ghidoni, Ulisses and Vismari, Lucio and J£Nior, Joæo Batista Camargo},
  title = {{Using Reinforcement Learning and ADS-B Data for the Development of a Sense-and-Avoid System to the Integration of Autonomous eVTOL in the São Paulo Urban Airspace}},
  booktitle = {Anais do XXI Simpósio de Transporte Aéreo – SITRAER 2024},
  editor = {Oliveira, Francisco Heber Lacerda de and Rocha, Fabiano and Borille, Giovanna Miceli Ronzani and Silva, Evandro José da and Caetano, Mauro},
  pages = {515--522},
  year = {2024},
  address = {Fortaleza, CE, Brasil},
  publisher = {Even3 Publicações},
  isbn = {978-65-272-1008-5},
  doi = {10.29327/xxi-sitraer-2024.887985},
  url = {https://sitraer-sbta.org/anais-2024-887985.html}
}