Performance Evaluation of Median U-Turn Intersection for Alleviating Traffic Congestion: An Agent-Based Simulation Study

Abstract

**Please note that the full text is embargoed** ABSTRACT: In recent years, urban mobility has become a big concern for Dhaka city, the capital of Bangladesh, due to severe traffic congestion. At-grade signalized intersections are one of the bottlenecks that significantly affect the capacity of the urban road network resulting in long waiting time and disrupted traffic flow. In this article, we studied the performance of Median U-Turn (MUT) intersection, a nonconventional low-cost design, as an alternative to the traditional four-way intersection. A high-fidelity simulation model was developed using the Agent-Based Modeling (ABM) approach to mimic the traffic flow of one of the busiest intersections in Dhaka city. Data were collected from a field survey during morning peak hours. The simulation results show that the average travel time to cross the intersection can be decreased by 16.8% by implementing MUT intersection, although the average number of stops increases by 5.5% in comparison to the existing traditional four-way intersection. Overall, the composite performance metric shows an improvement of 14.3%. Additional simulation experiments reveal that the performance of MUT intersection is comparatively better in case of moderate to high traffic volumes than low traffic volumes. The findings of this study provide valuable evidence to the transport planners to consider MUT at busy intersections for alleviating traffic congestion in Dhaka city.