Use of Vehicle Feature Information for Estimating Section Travel Times on Freeway

Abstract

Section-based traffic information is known to better reflect traffic dynamics and changes compared to the point traffic measurements. Furthermore, travel time, one of the section-related traffic measurements, plays a significant role in traffic operation and control and is directly available by tracing individual vehicles. Therefore, many studies have been conducted to obtain accurate travel time based on available infrastructure and traffic sensors. This paper proposes a new approach to estimate section travel times using vehicle features extracted from different sensors. Inductive vehicle signatures and video vehicle images are collected and further analyzed for deriving useful vehicle features to represent physical characteristics of individual vehicles. A new method, which estimates section travel times using both vehicle signatures and vehicle images, is presented. The promising results present new insights in utilizing various detector outputs toward better travel time estimation.