Evaluation of the operational and the safety performance of unconventional intersections

This research evaluated and compared the operational and the safety performance of five types of unconventional intersections, including the conventional median U-turn (MUT), unconventional MUT, upstream signalized crossover (USC), double crossover intersection (DXI) and crossover displaced left-turn (XDL). The traffic simulation model VISSIM was used to model the intersections and assess the operational performance based on delay and capacity. Vehicle trajectory data were extracted from VISSIM and analyzed in the Surrogate Safety Assessment Model (SSAM) to produce conflict results based on conflict indicators TTC and PET. For the operational performance, the results were generated based on default driving behavior parameters. For safety performance, the results were generated based on both default and safety-oriented calibrated driving behavior parameters. Three methods were applied to determine the signal phases and cycle length, including trial-and-error, simple progression and using the signal optimization software Synchro. Overall, the unconventional MUT has the lowest capacity among the five designs, around 950 veh/hr. Compared to this value, the capacity of the conventional MUT is 47% higher; the capacities of the USC and DXI are 111% and 89% higher; the capacity of the XDL is about 174% higher. By comparing the delay and the conflict results, it can be found that, in general, higher delays will cause a higher number of conflicts. In balanced volume conditions, the unconventional MUT is most favored in terms of both delays and conflicts when the approach volumes are low. For intermediate approach volumes, the best choice is the XDL, followed by the USC. For high approach volumes, the XDL is the most recommended design. However, the DXI and conventional MUT are not recommended in balanced volume conditions. In unbalanced volume conditions, the unconventional MUT has both the lowest delays and the lowest number of conflicts when the minor-to-major road volume ratio is low. For an intermediate volume ratio, the DXI is the most recommended design. When the ratio is relatively high but still in unbalanced conditions, the XDL is most recommended. In addition, the conflict results generated by calibrated models are more consistent and accurate than those generated by default parameter models.Applied Science, Faculty ofCivil Engineering, Department ofGraduat