Citation

Tian J, Treiber M, Ma S, Jia B, Zhang W. Transp. Res. B Methodol. 2015; 71: 138-157.

Copyright

(Copyright © 2015, Elsevier Publishing)

DOI

10.1016/j.trb.2014.11.003

PMID

unavailable

Abstract

The essential distinction between the Fundamental Diagram Approach (FDA) and Kerner's three-phase theory (KTPT) is the existence of a unique gap-speed (or flow-density) relationship in the former class. In order to verify this relationship, empirical data are analyzed with the following findings: (1) linear relationship between the actual space gap and speed can be identified when the speed difference between vehicles approximates zero; (2) vehicles accelerate or decelerate around the desired space gap most of the time. To explain these phenomena, we propose that, in congested traffic flow, the space gap between two vehicles will oscillate around the desired space gap in the deterministic limit. This assumption is formulated in terms of a cellular automaton. In contrast to FDA and KTPT, the new model does not have any congested steady-state solution. Simulations under periodic and open boundary conditions reproduce the empirical findings of KTPT. Calibrating and validating the model to detector data produces results that are better than that of previous studies.