Traffic Analysis as a Service via a Unified Model

This paper describes a traffic analysis platform that is implemented as a service, so different government authorities or geographic locations can utilize its functionalities more efficiently and effectively. In particular, most of the research in traffic analysis is solely focused on numerical analysis, while the conceptual relationships among moving objects are rarely captured. In this paper, we investigate into traffic relationships by a unified conceptual model that captures traffic dynamics under changing contexts to facilitate service implementation. By using the unified model, we can capture domain knowledge in a robust and consumable format for our proposed services for different users and on different geographic databases. The knowledge can also be mapped into a service computational model for other value-added traffic analysis. Instead of using the previously proposed spatio-temporal ER model, we define our model by augmenting the traditional ER model with symbolic contexts to better capture spatio-temporal context variations and their relationships with moving objects. We demonstrate the expressive power of our model by using Beijing taxi trajectory data to show that our approach is effective in representing traffic relationships for analysis.