This Is AuburnElectronic Theses and Dissertations

Development of Crash Modification Factor for Low-cost Median Opening Treatments at Unsignalized Intersections of Rural Divided Highways




Priyanka, Humayra Kabir

Type of Degree

Master's Thesis


Civil and Environmental Engineering


An effective access management technique is such that it has the lowest level of access control to balance traffic operation and safety. Increasing the level of control can increase the overall delay. On the other hand, no access control can be hazardous to some vehicle movements. Research shows that minor street crossing movements and left turns on the major street are the most hazardous movements. The drivers making these movements need to select a gap from two directions of oncoming traffic. Understanding driver behavior during these movements can provide better insights which will help to decide on the implementation of roadway treatments. Alabama Department of Transportation implemented low-cost median opening treatments (i.e., stop bars/yield lines, stop/yield signs, and double yellow lines on median opening) to improve the safety of unsignalized intersections on rural divided highways. But the safety benefits of these treatments are unknown. At first, the study aims to analyze their impact on driver behaviors using Naturalistic Driving Study (NDS) and field video data. A traffic conflict analysis is also conducted to evaluate the safety effectiveness as an indirect or surrogate method for a safety study. Finally, crash data analysis is conducted as the direct method of the safety study. Crash modification factor (CMF) is one of the tools to quantify the effectiveness of treatment. There is no CMF for these median opening treatments in the CMF clearing house. Therefore, the study aims to develop crash modification factors (CMFs) for these median treatments and compare the results of the conflict study with the developed CMF.iii The study collected NDS data, which includes video data of a total of 428 trips taken by 65 study participants. The NDS data analysis showed that major road AADT and speed play an important role in drivers' behavior at unsignalized intersections. The study also collected traffic conflict data (48 hours of video data for each location) and five years of crash data (2016 to 2020) at six groups of unsignalized intersections (treated vs. non-treated) in Alabama. The conflict analysis found that the treated intersections have 8% to 40% fewer conflict rates than the nontreated intersections. The cross-sectional Empirical Bayes (EB) method was conducted to develop the CMF. The predicted crash frequency was estimated using the Safety Performance Functions (SPF) in the Highway Safety Manual (HSM) and a local calibration factor developed by ALDOT. The EB method was applied to calculate the expected average crash frequency for each group of intersections (treated vs. non-treated) using the predicted and observed crash frequency and the CMFs for each group of intersections. The combined CMF for these treatments is 0.70, with a standard deviation of 0.22. The result suggests that these treatments reduce the expected crash frequency by 30%. This CMF and the traffic conflict study results can be applied to project-level decision-making.