Abstract
This paper tries to apply seasonal adjustment methods based on Hierarchical Bayesian estimation to clarifying a mechanism of capacity bottleneck phenomena in basic section (i.e., access-controlled section). For the purpose, first, homogenous traffic flow data in a sag section in terms of characteristics of car-following behavior is collected by so-called ‘self-iteratively following-the-leader method’ on driving simulator. Second, the obtained traffic flow data is converted to time headway data. Third, prior distributions are defined satisfying following three prior knowledge: the smoothness of time headways of successively running cars, a slight change of time headways of an identical car in chronological order and a small change of the mean of time headways in platoon in chronological order. The result of an implementation trial shows that the seasonal component reflects the effect of longitudinal alignment on traffic flow.
