Variations in Driver Behavior: An Analysis of Car-Following Behavior Heterogeneity as a Function of Road Type and Traffic Condition

Abstract

Microsimulation modeling is a tool used by practitioners and researchers to predict and evaluate the flow of traffic on real transportation networks. These models are used in practice to inform decisions and thus must reflect a high level of accuracy. Microsimulation models are comprised of sub-models, which control individual vehicle movements throughout the simulated network. These sub-models must be calibrated to accurately capture realistic driving behavior. This research utilizes data collected by the FHWA Living Laboratory instrumented research vehicle to produce evidence of global trends in car-following behavior. Unlike similar studies, this analysis focuses on the physical action taken by the driver—the acceleration—rather than the outcome of that action—speed selection or temporal/spatial gap. This approach enables better interpretation and comparison between car-following behavior in varying “driving environments,” that is, on different roadway functional classifications (freeway v. interstate), operational conditions (work zone v. non-work zone), and traffic conditions (congested v. uncongested). This analysis produced conclusive evidence that intra-driver car-following behavior is heterogeneous and is a function of the driving environment. Trends in acceleration behavior were examined on an aggregated psychophysical plane, which accounted for inter-driver heterogeneity, and a statistical analysis identified regions of significantly different acceleration behavior. Lastly, heterogeneity in car-following acceleration behavior in work zones and non-work zones was also verified.