Driver Behavior at Highway-Rail Grade Crossings Using Naturalistic Driving Study Data and Driving Simulators

摘要

While it has been long recognized that a full understanding of driver behavior at highway-rail grade crossings (grade crossings) is one of the key elements to improve safety, collection of data to analyze the behavior has been challenging. Various approaches such as statistical, observational, and driver survey studies, have been tried over time with limited success. This paper discusses two promising new approaches for such analysis: the use of a Naturalistic Driving Study (NDS) database, and analysis conducted with driving simulators. The NDS data contains video and vehicle sensor data of drivers in their own vehicles driving routes they use in their normal daily routines. This data can be mined to look at driver behavior at grade crossings, including the behavior of drivers who pass through grade crossings without an incident. The current work at Michigan Tech has concentrated on developing tools for reduction and organization of the NDS data, so it's more appropriate for mining efforts. The second approach uses driving simulators to study driver behavior at grade crossings, as simulators are quickly becoming a leading tool for researching various aspects of driver behavior. Current simulator research at Michigan Tech has investigated driver behavior in the grade crossing environment. Although driving simulators are an excellent tool for analyzing driver behavior, additional work is needed to validate the results in the grade crossing environment. A combination of NDS and simulator data may provide the link needed for this validation and is highlighted as one of the key objectives for future research.