Vision-Based Approach to Study Driver Behavior in Work Zone Areas

Work zone safety is a pressing concern, and it has been intensified by the increasing number of roadway maintenance and rehabilitation projects. However, the majority of work zone studies are based on crash data analysis, and there is limited research on monitoring and studying actual driver behavior in work zone areas. In this study, a vision-based sensing system developed through a research project sponsored by the National Academy of Science National Cooperative Highway Research Program Innovations Deserving Exploratory Analysis (NCHRP IDEA) program was used to monitor driver behavior in work zone areas. The purpose of this study is to investigate driver behavior at work zones. Work zone transition and advance warning areas were divided into four zones so the merge timing and behavior of vehicles (e.g. trucks and others) under different roadway geometry could be evaluated quantitatively. Two right lane one-lane closure work zones, one on a straight road segment and the other on a curved road segment, were studied by reviewing two sets of 30-minute video collected on I-95. Preliminary results show that vehicles tend to merge close to the work zone taper on a curve, and truck drivers tend to merge earlier than other drivers. Future research needs are also discussed.