Modeling multi-class disordered traffic flow subject to varying vehicle composition using the concept of traversable distance

Abstract

Disordered traffic stream at the microscopic level can be described as a permeable medium. Each vehicle is considered to traverse through a series of lateral gaps created by other vehicles. We develop a multi-class traffic flow model that considers such viable and accessible gaps for individual vehicle classes to traverse downstream. The model accounts for the varying shares of different vehicle classes. The concept of traversable distance and modified equilibrium speed functions are used to model the interplay among multiple vehicle classes. Using a higher order finite volume approximation method, evolution for a two-class traffic stream is shown. This model replicates prominent empirical characteristics exhibited by multi-class disordered traffic such as overtaking and creeping. Varying shares of smaller vehicles affect queue formation and discharge characteristics, and have significant impacts on roadway capacity. The model also could compute dynamic class-specific travel times at different vehicular compositions more realistically.