Empirical Study of Lane-Changing Maneuvers in a Weaving Area Based on Reconstructed Trajectories of Floating Car Data

Abstract

The concentration of lane-changing maneuvers is the main contributor to speed drop and congestion in weaving areas. Despite their importance, lateral maneuvers have received less attention in research than longitudinal driver behavior. The main reason for this is the lack of appropriate data for comprehensive analysis. In this paper, we have used reconstructed lane-level accurate trajectories collected during 12 days in a weaving section with more than 3.3 km length. Examining the classification of maneuvers into mandatory and discretionary versus their classification based on the direction of the maneuver showed that the latter could be better described by traffic variables such as changes in density and speed of the target lane compared with the source lane. We have also shown that the origin–destination pattern of drivers and the time of day (traffic conditions at different hours) significantly affect the number and location of lane-change maneuvers. In addition, the location of performing the weaving maneuvers at different times of the day appears to affect the travel time experienced by drivers. This data source and in-depth analysis open many opportunities for improving empirical traffic flow theory, better design of weaving sections, and active management through vehicle-to-everything (V2X) communication and online drivers’ guidance toward cooperative behavior.