Data Sources for Urban Traffic Prediction: A Review on Classification, Comparison and Technologies
B P Ashwini, R. Sumathi
Abstract
Traffic prediction plays a vital role in the process of urban traffic management. Various traffic prediction applications will include different traffic analysis parameters, such as traffic signal control, navigation systems, incident detection, etc. Generally, traffic prediction is considered as a data-intensive task and it includes challenges in capturing, pre-processing, integrating, and analyzing the huge volume of spatial - temporal data. The performance of the traffic prediction is mainly dependent on the right, reliable, and compatible data sources. In this paper, an attempt has been made to deploy a classification of data sources for traffic prediction including innovative sources such as social media and cellular network data. Contributions towards traffic prediction under each data source are presented. Also, a comparative study has been carried out on various data sources based on the quality parameters such as data accuracy and reliability, pre-processing complexity, data adequacy, infrastructure cost, and maintenance overhead.