| Recently, many state and local transportation agencies have begun considering using U-turns as alternatives to direct left turns through installing nontraversable medians on multilane highways. Some of the concerns are that this proposed change may transfer operational problems to the downstream median opening and increase delay or travel time of vehicles from driveways. However, the impact of this treatment on roadway operations is not clear. Past studies that have attributed operations gains to this treatment have not focused on the specific situation at the isolated intersection or at the median opening where U-turns occur. A comprehensive study of operational impacts of this treatment would be beneficial, both in setting design policy and in project-level design.; Full-scale field studies were conducted to obtain the data necessary to quantify operational effects of this treatment. Traffic flow data were collected at ten sites in the Tampa and Clearwater areas, Florida. About 300 hours of field data were collected, including delay, travel time, traffic volume, speed, traffic control, and geometric data. An operational effects database was developed to perform a statistical analysis.; In this study, delay models of Direct Left Turn (DLT) and Right Turn plus U-turn (RTUT) were developed as a function of major and minor-road traffic flow rates. Analysis indicates that the average running time of RTUT in the weaving section has a linear relationship with the length of the weaving segment. This was incorporated into a travel time model that was used to compare the average travel time for the two movements. The curves developed based on delay and travel time models present a good picture of the operational impacts of the two types of movement.; Speed analysis indicates that the reduction in major-road through-traffic speed due to both movements was, on average, very small. Moreover, a ratio model of RTUT was developed to estimate how many drivers would like to make a RTUT rather than a DLT. Additionally, a before-and-after analysis shows the weighted average total delay at the full median opening was 15%–22% longer than that at the directional median opening. Finally, the operations models were used to develop guidelines for selecting a median treatment. |
