Modeling Probabilistic Capacity Expressed with Longitudinal Alignment of Upstream Sag

Abstract

This paper points out that the ratio of longitudinal gradient change at upstream sag is a significant factor for breakdown in flow. In this study, traffic detector data collected on Tohoku Expressway in 2016 and longitudinal gradient data are used to identify positions of bottlenecks and to prepare data set containing with or without the breakdown, traffic flow-rate and ratio of longitudinal gradient change at a sag of bottleneck and their neighboring sags. First, as a result of calibration of logistic regression models using the data set, the ratio of the longitudinal gradient change at the upstream sag of the bottleneck as well as that at the bottleneck are found as significant explanative variables. On the other hand, the ratio of longitudinal gradient change of the downstream sag is not significant. Second, probabilistic capacity curves considering the ratio of the longitudinal gradient change of the bottleneck sag and the right upstream sag are estimated by applying a parametric survival model. This paper concludes that it is essential to revisit mechanisms of flow breakdown at bottleneck in access-controlled section and countermeasures to mitigate congestion.