The closure of signalized intersections and the rerouting of traffic into midblock U-turn median openings (U-turn slots) paved the way for free-flow traffic conditions (uninterrupted flow) in an intersection. However, direct left turn (DLT) vehicles will not have to take a series of right-turn plus U-turn (RTUT) maneuvers in order to duplicate the same path taken prior to the closure. As a consequence, RTUT vehicles must weave cross (lane change) into the innermost lane towards the U-turn slot. Currently, methods for analyzing the speeds in weaving segment were derived exclusively on freeway weaving sections. Thus, there is a need to develop a method to estimate these speeds specifically for urban streets.
The weaving maneuver that closely compares to the weaving of a RTUT vehicle is the two-sided Type C freeway weave. The vehicle makes a right turn into the major road, weaves across two or more through lanes, and exits by making a U-turn maneuver at the adjacent median opening.
Data collection was accomplished by using video recording equipment at the two study sites. Data extraction was carried out in a computer by manual counting after the video footage has been converted to digital format. Separate datasets were gathered for the two models evaluated based on their requirements.
In this study, the HCM Weaving Model for Type C was first used to predict weaving and nonweaving speeds. No calibration was carried out for the HCM model in this study. The second model was the NCHRP Ramp weave Model. The weaving maneuver where this model was derived from is the off-ramp right- turn movement that weaves across the arterial to make a left-turn at the downstream intersection. Two study sites were selected for data collection. Weaving and nonweaving speed model were calibrated for each site.
Results of the HCM model show that weaving speeds were predicted satisfactory well while nonweaving speeds were significantly different. For the NCHRP models, correlation analysis show that weaving and nonweaving speeds decreases with an increase in weaving flow rate and arterial flow rate. After the models were calibrated, the model validations were performed next using the excess field date to check the accuracy of the model predictions. Paired samples t-test was performed for this purpose. The calibrated models accurately predicted the observed speeds with the 95% level of confidence. The resulting NCHRP models can be used to estimate weaving and nonweaving speeds in other sites having the same configuration.
