Asymmetric stochastic Tau Theory in car-following

In this paper, we examine the distribution of headways collected from Next Generation SIMulation (NGSIM) datasets. Statistics show that the observed headways follow a certain log-normal type distribution within each preselected velocity range. To explain this phenomenon, an asymmetric stochastic extension of the Tau Theory is proposed. This novel approach assumes that the observed headway distributions are due to drivers’ continuous headway adjustment behaviours, and that the intensity of headway change is proportional to headway magnitude. The agreement between the model predictions and the empirical observations indicates that driving behaviours may be implicitly governed by the physiological Tau characteristics of drivers.

► We propose an asymmetric stochastic extension of Tau Theory.
► Physiological Tau characteristics govern driving behaviors in an implicit and stochastic way.
► We verify the above conclusion based on NGSIM datasets.