Bivariate modeling of time headways in mixed traffic streams: a copula approach

ABSTRACT

Vehicle headway distributions provide an efficient representation of the uncertainties in traffic flows that have diverse applications in capacity estimation, safety evaluation, planning, design of roadway systems and in traffic operation studies. The current study is aimed at investigating the applicability of the copula approach to accommodate the dependence structure between vehicle headway and the centerline separation between the interacting vehicles in vehicle-following conditions for different vehicle-pair combinations, using trajectory data from video recordings. The results of the study indicated that Frank copula could model the joint distribution for almost all leader–follower vehicle pairs. As importantly, the joint distribution and the conditional non-exceedance probabilities of time headways demonstrated the importance of explicitly considering centerline separation in time headway modeling and the need to develop bivariate vehicle type-specific behavioral models for mixed traffic streams, in order to augment the reliability of existing models, better understanding and modeling of non-lane-based traffic streams.

KEYWORDS:

• Time headway
• centerline separation
• copula
• mixed traffic
• heterogeneous

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

The term “vehicle-following” is similar as car-following that is used to describe the following behaviour of cars with the leading cars. To account for vehicle heterogeneity in mixed traffic streams, vehicle-following is a more suitable terminology and is therefore used in this paper.