Impact of Advertising Signs on Freeway Crashes within a Certain Distance in Michigan

In spite of some improvements, more treatments are needed for transportation safety. Advertising signs are among factors that adversely affect drivers’ vision, and increase the number of crashes at some locations. This paper discusses 42 advertising signs located on different freeway sections in Michigan that could contribute to crashes at certain distances from the signs. The total number of crashes was collected from the state crash database for the period from 2009 to 2013. A buffer area of 1,500 ft was defined for before and after each sign using the ArcGIS 10.0 software package. The distribution of the crash distance consisted of two gamma distributions, and parameters of mixed distribution were investigated by employing the expectation maximization algorithm (EM algorithm) as an iterative statistical method. The R software package was utilized for this algorithm to define the parameters of the mixed distribution and to minimize misclassification. Welch’s formula was used to define the border, and Newton’s procedure was employed to find an X value (i.e., identified distance) to satisfy Welch’s formula. Results show the selected advertising signs could increase the number of crashes 552 ft from the signs. A regression model also indicated that among other factors such as land use characteristics, median type, ramp presence, number of lanes, and other type of signs on the freeway sections, these 42 signs have significantly caused an increase in crashes, and it is recommended to relocate them to other parts of freeways in order to improve traffic safety.