| The safety problem has become more and more severe with the increase of travel demands. It is difficult to get sufficient accident data to do effective research on localized safety problems. There are also many problems surrounding accident data as a measure for safety. The main problems are the reliability of the reports, the human errors and there could be a long wait time until an adequate size sample is obtained. On the other hand, traffic conflict studies have been accepted because the data and the effectiveness of a treatment can be collected and evaluated in a short period of time. Finally, traffic conflict data provides information about traffic volumes, routine conflicts, erratic maneuvers, and severe conflicts.; Left turn movement from a driveway is the source of many accidents.{09}In order to overcome this problem, reduce conflicts, and improve safety along the arterial road, transportation engineers and planners have used access management as an efficient tool to improve the safety characteristics of a road system. Median treatment is one of the access management techniques where making a right turn plus U-turns are alternatives to direct left turns. The safety impact of that treatment needs to be clarified.; More than three hundred hours of field data was collected in the field from ten sites in Tampa, Florida area including driveway volume, through traffic volume, number of conflicts, speed and geometric data. A safety effects database was set up to perform the statistical analysis. The objective of this study is to develop models that can predict conflict rates for the two different types of left turn maneuvers. Based on those rates, a correct decision may be made and a suitable access technique may be applied. From the data analysis for both maneuvers, it was clear that making a U-turn is safer than a direct left turn. According to conflict data with the conflicting volumes, a median treatment chart was another output from this study. Finally, two models were performing in the form of linear equations to predict the conflict rates of direct left turn and right turn plus U-turn at unsignalized intersections. |
