Observations of macroscopic non-mandatory lane-changing behaviors on the Capital Beltway

This paper explores the interrelations between macroscopic non-mandatory lane-changing behavior and traffic flow characteristics on a freeway segment during the peak and off-peak periods. The field data were collected from a main segment of the Capital Beltway. Based on the results of exploratory analyses, this study has developed two groups of models to capture the impact of traffic conditions on the macroscopic lane-changing patterns, defined as the frequency of lane changes and the percentage of vehicles changing lanes. Both Poisson regression and logistic regression have been employed for estimation of various model specifications.