The process of urbanization on a global scale has generated a significant increase in metropolitan populations, which in turn brings with it a series of challenges for the management of transport infrastructure. In this context, the importance of accurate traffic flow prediction in urban traffic management systems stands out, with the city of Catania being used as the focus of this study. However, predicting traffic flow faces challenges, such as the high costs associated with acquiring extensive real traffic data, particularly when installing sensors on every road is not feasible. To address the issue of missing traffic flow data on roads without sensors, this study proposes an innovative approach based on machine learning, using the Long Short-Term Memory architecture and incorporating both sensor and floating car data. The results obtained are thoroughly analyzed and discussed, highlighting the ability of the LSTM model to provide accurate predictions of urban traffic, even with a limited amount of data available. These findings highlight the relevance of employing advanced machine learning techniques in the efficient management of urban mobility, aiming to improve the quality of life in cities and deal more effectively with the challenges of contemporary urban traffic.

Enhancing Urban Traffic Management through Machine Learning Prediction Models for Sensor-less Roads

Thamires De Souza Oliveira
;
Sarah Di Grande;Mariaelena Berlotti;Salvatore Cavalieri;Vincenza Torrisi;Giovanni Calabro';Giuseppe Inturri
2024-01-01

Abstract

The process of urbanization on a global scale has generated a significant increase in metropolitan populations, which in turn brings with it a series of challenges for the management of transport infrastructure. In this context, the importance of accurate traffic flow prediction in urban traffic management systems stands out, with the city of Catania being used as the focus of this study. However, predicting traffic flow faces challenges, such as the high costs associated with acquiring extensive real traffic data, particularly when installing sensors on every road is not feasible. To address the issue of missing traffic flow data on roads without sensors, this study proposes an innovative approach based on machine learning, using the Long Short-Term Memory architecture and incorporating both sensor and floating car data. The results obtained are thoroughly analyzed and discussed, highlighting the ability of the LSTM model to provide accurate predictions of urban traffic, even with a limited amount of data available. These findings highlight the relevance of employing advanced machine learning techniques in the efficient management of urban mobility, aiming to improve the quality of life in cities and deal more effectively with the challenges of contemporary urban traffic.
2024
Traffic flow prediction, machine learning, LSTM, urban mobility
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.11769/642393
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 0
  • ???jsp.display-item.citation.isi??? ND
social impact