Lane Choice Model for Signalized Intersections with an Auxiliary Through Lane

An auxiliary through lane (ATL) is a commonly used congestion relief measure at signalized intersections in which fiscal and right-of-way constraints do not allow for full widening between intersections. Previously developed models have predicted the flow in the ATL as a function of macroscopic elements, such as through-movement demand and the ratio of average green time to cycle length. However, these models explain neither driver behavior nor motivation to use the ATL, nor can such an approach be used in microscopic simulation models. This paper presents empirically developed models for driver lane choice at signalized intersection approaches with one ATL and one continuous through lane (CTL). These models were developed from a calibration data set that covers eight ATL approaches in three U.S. states, for a total of 12 h of observation. The results suggest that the utility of the ATL is a function of each through-movement driver’s arrival time (during either the effective red or green phase), and the queue lengths in either lane at the time of arrival. After calibration, the models were evaluated by applying them to data from a validation site from outside the calibration data set. The paper concludes with a discussion of how the models should be implemented within the lane-change algorithms of microsimulation systems.