top of page
  • Jérôme Chenal

Artificial Intelligence Based Methods for Smart and Sustainable Urban Planning: A Systematic Survey

Tekouabou, Stephane Cedric Koumetio; Diop, El Bachir; Azmi, Rida; Chenal, Jerome

2022


Artificial Intelligence Based Methods for Smart and Sustainable Urban Planning: A Systematic Survey

Publié dansArchives Of Computational Methods In Engineering

Date2022-11-01

EditeurDordrecht, SPRINGER

ISSN1134-3060


1886-1784

Mots-clés (libres)

machine learning-methods; form; regression; network; metrics; orange; impact

Le document apparaît dansPublications validées par des pairs


Travail produit à l'EPFL


Articles de journaux


Publié

Date de création de la notice2022-11-21


The world's cities are facing various challenges such as rapid urbanization, poverty, climate change, pollution, sustainable and inclusive development. Building futuristic smart and sustainable cities is proving to be a response to these challenges. Recent publications show that urban planning decision support is increasingly using artificial intelligence through machine learning methods to address these challenges. 172 papers were published in 2020 compared to 8 before 2010 with a forecast of at least 200 papers in 2021. Despite the explosive number of scientific publications on applications artificial intelligence for urban planning decision support, few studies have made a systematic assessment of the state of the art to inform future research. This would help focus on the approaches used, the planning problems most commonly addressed, the data used and even the study areas. We found that the top 5 most addressed urban planning issues include: land use/cover, urban growth, urban buildings, urban mobility and urban environment. Furthermore, a large amount of data was used from sensors and simple or ensemble machine learning methods were more used in this case. Deep learning methods are more used for land use/cover, buildings and climate issues which are mostly based on satellite image data. On the other hand, China and the United States are the most studied territories while Africa is almost not. A high intensity of collaboration between researchers affiliated with Chinese, American and English institutions was observed. Thus, urban planning researchers should benefit from this synthesis work by understanding the general idea of the application of machine learning methods in urban planning, its trends, issues, current challenges and future research directions.

2 vues

Posts récents

Voir tout
bottom of page