The impact of the leading vehicle type on car-following behaviours

Modelling car-following in an effective and accurate way is of great importance for several areas of application, such as microscopic traffic simulation and ADAS (Advanced Driving Assistance Systems). Heterogeneity can be observed in driving behaviors if car-following data are analyzed. Part of this dispersion depends on the inherent heterogeneity across drivers, that could react in different ways to very similar stimuli. However, another source of dispersion could be related to a misleading identification of the stimuli or to an improper identification of the context in which the stimuli are evaluated. This work is oriented to analyze if a non-negligible part of the observed heterogeneity can be explained by considering the type of the leading vehicle. Observation of car-following trajectories has been carried out in a large experiment involving one hundred drivers, with different leading vehicles. These observation have been interpreted by means of behavioral models and the parameters of these models have been separately identified for different types of leading vehicles. Comparison of modelling parameters, as well as directly observed variables as speed and spacing, allows for testing if the type of leading vehicle actually influences the car-following behavior.