Speed prediction in large and dynamic traffic sensor networks

作者：

Highlights：

• Dynamic traffic sensor networks bring challenges in the context of urban mobility

• We evaluate three approaches for speed prediction over large/dynamic sensor networks

• The global and cluster-based approaches provide accurate and robust prediction models

• The global approach solves the cold start problem

• We provide a large dataset and assess the effectiveness of the three approaches

摘要

•Dynamic traffic sensor networks bring challenges in the context of urban mobility•We evaluate three approaches for speed prediction over large/dynamic sensor networks•The global and cluster-based approaches provide accurate and robust prediction models•The global approach solves the cold start problem•We provide a large dataset and assess the effectiveness of the three approaches

论文关键词：Smart cities,Intelligent transportation systems,Short-term traffic prediction,Dynamic sensor networks,Machine learning,Urban mobility

论文评审过程：Received 20 January 2019, Revised 16 September 2019, Accepted 25 September 2019, Available online 11 October 2019, Version of Record 15 February 2021.

论文官网地址：https://doi.org/10.1016/j.is.2019.101444