Modeling Freeway Lane Changing Behavior

Drivers continuously evaluate the surrounding traffic and the roadway environment, and make decisions about lanes and travel speed. The objective of this thesis is to develop a lane changing model that can be used in microscopic traffic simulation models to capture drivers' lane changing behavior. Lane change is modeled as a sequence of four steps: decision to consider a lane change, choice of left or right lane, search for an acceptable gap to execute the decision, and performing the lane change maneuver. First, a decision is made whether a driver will consider changing lanes. If a decision to consider changing lanes is made, a lane is chosen from the alternatives. Finally, the gap acceptance model determines whether the available gap in the target lane is sufficient for a safe merging and lane change can be completed. A discrete choice framework is used to model the lane changing behavior. The framework allows for modeling the impact of different elements of the traffic and roadway environment on driver behavior. The model is applied in the special case of merging from an on-ramp. Results from the estimation of the parameters show that in addition to the gap length, other important factors that affect drivers gap acceptance behavior are relative speed, distance remaining to the point at which lane change must be complete, and delay in completing merging. Finally, the estimated model is tested in a micro-simulation environment. Thesis Supervisors: Moshe E. Ben-Akiva, Title: Professor of Civil and Environmental Engineering, MIT Haris N. Koutsopoulos, Title: Associate Professor of Civil and Environmental Engineering, Carnegie Mellon University Rabi G. Mishalani, Title: Research Associate of Center for Transporting Studies, MIT