Date of Award


Publication Type

Master Thesis

Degree Name



Civil and Environmental Engineering


Applied sciences


Chris Lee




To improve traffic safety on freeways, many traffic researchers have used real time data to predict the likelihood of crashes, using number of crashes as the measure of safety. The parameters of speed, volume or density have been used extensively in previous research to calculate the crash likelihood. This research studied the combined effects of volume and density to predict crash likelihood using real time data a short time before crash occurrence. The volume-density relationship provided a measure of growth and dissipation of queue on the freeway, known as the shock wave speed. Using this shock wave speed and quantifying various types of shock waves, analysis was done to predict crash likelihood. The results of logistic regression analysis indicated that increasing the speed of forward shock wave decrease crash likelihood. Using a log-linear relationship and including exposure measures, it was found that diverging sections, normal weather conditions, low shock wave speeds and forward moving shock waves indicated increased likelihood of crashes. Finally, using an odds ratio to compare the combined effects of shock wave speed and shock wave type, it was determined that forward moving shock waves yield a greater likelihood of crash for both low and high shock wave speeds.