A Dynamic Real-Time Incident Detection System for Urban Arterials—System Architecture and Preliminary Results

See related content

by Chao-Hua Chen, Univ of Maryland, College Park, United States,
Gang-Len Chang, Univ of Maryland, College Park, United States,

Document Type: Proceeding Paper

Part of: Pacific Rim TransTech Conference: Volume I: Advanced Technologies

Abstract: This study presents a system which is capable of detecting incidents in real-time and adjusting itself to traffic dynamics through the embedded learning mechanism. The entire system consists of three major components: a dynamic traffic flow prediction model, an incident identification model, and the incident monitoring process. The proposed traffic flow prediction model employs the event-based simulation concept to capture the traffic dynamics, but its parameters are time-varying through a learning mechanism. The predicted traffic conditions in comparison with the real-time detected traffic constitutes the core of incident detection. The incident identification model serves to provide the severity information of a detected accident and to assist in determining the occurrence of accidents. The function of the proposed monitoring process is to integrate all individual models as an effective system.

Subject Headings: Traffic models | Traffic accidents | Traffic flow | Dynamic models | Flow simulation | Traffic analysis |

Services: Buy this book/Buy this article


Return to search