Validation of a Macroscopic Lane-Changing Model

Abstract

Systematic lane changes inside a weaving area could lead to accidents and disrupt the entire traffic flow. Recently, a simple macroscopic lane-changing model that determines the number of lane changes from on-ramp flows and the number of lanes was proposed to describe traffic dynamics inside the lane-changing area at a merging section. However, this macroscopic lane-changing model had not been validated with vehicle trajectory data. This study attempted to fill that validation gap. First, lane changes were categorized into two types: left-lane changes and right-lane changes. Then the following two hypotheses derived from the macroscopic lane-changing model were validated: (a) the number of left-lane changes within two adjacent lanes linearly increases from the leftmost lane to the shoulder lane and (b) the number of left-lane changes linearly increases with the flow rate of on-ramp vehicles inside the lane-changing area. Traffic balance among lanes was analyzed, and the two proposed hypotheses were validated for different periods and road geometries by using NGSIM vehicle trajectory data. The validation results confirm that the two linear relationships in the macroscopic lane-changing model exist.