Authors: Yue Liu, Gang-Len Chang and Ruihua Tao
Conference: 2007′ IEEE Conference on Intelligent Transportation Systems, Seattle, WA
This paper presents a video-based system for measuring dynamic dilemma zones at signalized intersections. The proposed system is cost-effective and sufficiently reliable for measuring various critical information associated driver behavior, including the speed before-and-after the yellow phase, the acceleration/deceleration rates and the approximate reaction time to an encountered yellow phase, and differences in driving patterns. Such information is critical for understanding the spatial distribution of dynamic dilemma zones and the design of strategies to improve intersection safety. This paper details the key components of the proposed system and the systematic procedures for field operations, data extraction, and system validation. The estimation of dilemma zones based on 1123 field observations of drivers’ responses via the proposed system at six intersections of high accident frequency are also reported in this paper.