Minimization of road network travel time by prohibiting left turns at signalized intersections

Abstract

Left turns have potential efficiency problems. The prohibition of left turns is explored as means of shortening total travel time and improving overall efficiency performance. The objective of this thesis is to explore whether prohibiting left turns can improve the efficiency of urban road networks using existing infrastructure. Improving the efficiency refers to reducing the travel time of all vehicles in the network. A model is proposed to determine the effects of left turn prohibition with the objective to minimize the total travel time. The first task is to forecast the distribution of demands as vehicles are redistributed in the network after left turns are prohibited. Prohibiting left turns not only affects the route choices of the affected vehicles, but also influences the vehicles' other movements because the prohibited turns may increase traffic flows on some links and cause delays. A stochastic user equilibrium model is applied to forecast the distribution of demands. Optimizing signal settings is another important task in the absence of left turns. The whole signal timing plan of the affected intersection has to be changed because the prohibited left turn is removed from the signal group. The corresponding signal timing is adjusted according to the redistributed traffic flow. Further, the lanes for prohibited left turn should be reassigned to make use of their capacities at intersections. This thesis presents two methods of signal setting optimization: the stage-based method and the lane-based method. Both methods consider the influences of left turn phasing types and left turn prohibition. Using the proposed method, it is determined that prohibiting left turns may reduce the total travel time in the network, though this reduction has not been observed for every origin-destination path. The proposed method can handle various traffic demands. Protected left turns with small flows, left turns with large opposing flows, and permitted left turns at intersections with high saturations have a higher probability of being prohibited. This research provides insight into network design and congestion management in urban road networks. Using the proposed model, the left turn prohibition problem can be solved analytically. Signal setting optimization methods are improved, and can handle the absence of left turns. The findings from the numerical solution could contribute to the usage of left turn prohibitions in practice.