Application of distributed traffic simulation for passing behavior study

The suitability of a prototype distributed traffic simulation for passing behavior studies was examined. The prototype was based on the High Level Architecture and was designed to combine the microscopic traffic simulation program VISSIM with the driving simulator DriveSafety. Specifically developed for the passing behavior application, the prototype was capable of modeling traffic on two-lane, two-way rural roads. Vehicle information was transferred from the VISSIM federate to the DriveSafety federate to control the creation of vehicles and their speeds. The passing experiment had a repeated-measure, factorial design with two levels of impeding vehicle length and two levels of impeding vehicle speed. Twenty-four test drivers, each with at least 5 years of driving experience, successfully completed a pass under each experimental condition. The passing time and passing distance were observed to be significantly greater when the impeding vehicle was a truck than when it was a passenger car. The mean distance traveled in the left lane and the mean end gap distance were found to be approximately double those reported previously. Likewise, the mean start gap distance was found to be triple that reported previously. These differences were likely due to drivers' propensities to underestimate distances in simulated environments. This study methodology had the advantage of capturing detailed driver information, but concerns about the validity of the methodology and the transferability of the results to the real world warrant further examination.