Identifying Traffic Flow Patterns of Urban Roads from Geographical Context Information

A vast amount of traffic flow information can be collected and stored with the large-scale deployment of traffic detectors in urban road networks. Identifying a limited number of common traffic flow patterns from a large amount of road section traffic flow information can enable traffic managers to classify and manage urban roads and greatly improve their management efficiency. Traditional classification of traffic flow patterns is based on temporal context factors. This paper will present a method for recognizing large-scale urban road traffic patterns that considers both temporal and spatial context factors. A traffic flow similarity measuring approach is constructed based on the pre-classification of historical traffic flow data, and unsupervised classification is used to recognize traffic flow patterns. This study’s findings provide a better understanding of the traffic flow patterns of large-scale urban roads.