Citation

Zhang C, Yang X, Yan X. J. Wuhan Univ. Technol. Transp. Sci. Eng. 2006; 30(6): 973-975.

Affiliation

School of Traffic and Transportation, Tongji University, Shanghai 200092, China

Copyright

(Copyright © 2006, Wuhan University of Technology)

DOI

unavailable

PMID

unavailable

Abstract

Traffic incident detection is an important part of freeway emergency management system. It is reviewed researches of automatic incident detection methodology, analyzed the shortcoming of existing algorithms based on fixed detector, such as dissatisfied detection efficiency, restricted transplant capability, etc. A new incident detection methodology using floating cars' data is proposed according to the fact that incidents affect car speed obviously. The result shows that the new incident detection algorithm has high identification rate (92.5%), low false identification rate (1.2%) and short mean time to identification (1.6 min).