Improving the performance of headway control tools by using individual driving speed data

Abstract

Fleet control operation tools for high frequency bus service seek to maintain headway regularity between consecutive vehicles. Almost always, these tools suggest a series of control actions that must be executed by drivers that are assumed to be identical in their driving behavior and that this behavior is identical in all driving contexts. However, not all bus drivers drive in the same way and their behavior can be affected by different factors related to the environment in which they carry out their work. In this paper the behavioral difference is characterized by the average speed at which each of them drives along a given route. This work focuses on determining the impact of this heterogeneity on headway variability, and on how (erroneously) considering drivers to be homogeneous impacts the performance of headway regularity control tools based on holding decisions. The results show that the reduction in waiting times when the holding control strategy is applied compared to the case without control increases when speed variability across drivers also increases. We show that an easy way of improving headway regularity is to stratify drivers by line according to historical driving speed, regardless of whether or not a control action is applied.