A traffic queueing model for exit lanes for left-turn intersections

Abstract

The unconventional design of exit-lanes for left-turn (EFL) has been confirmed to be an effective way to increase intersection capacity when left-turn demand is high. However, because of the unique feature of EFL control, there would be a strong oscillation in the left-turn queues at signalized intersections when the EFL strategy is used, which makes the analytical evaluation of the EFL strategy very difficult. In particular, the random nature of traffic demand would make it even harder to estimate the queues. To this end, in this paper, we propose a signal state-dependent queueing model to quantify the dynamic behavior of left-turn queues at signalized intersections under EFL control. The obtained results will contribute to establishing an analytical numerical framework that gives us the possibility to study the effect of various factors on the left-turn queues at the EFL intersections and provide insight into the unique queueing behavior that is most likely to manifest this strategy. We validated the proposed model using field data, and it shows a good agreement between the estimated results and the field results. In addition, the numerical experimental results show that the obtained performance measures are practically beneficial to provide a numerical tool to quantify the uncertainty in the queues of left-turn movements with EFL control and thus enrich the guidelines for further improving the use of EFL design.