Transportation professionals and engineers have identified red light running as a major traffic safety hazard. However, there are currently only limited tools available to assist transportation professionals in selecting intersections for remediation or law enforcement based upon red light running violation risk. The goal of this research was to develop statistical models to predict red light running violation frequency (violations/hour) based upon traffic operational and intersection geometric characteristics for four-approach intersections. Red light running violation data over the time period of 2-8 PM was gathered from 19 intersection approaches in four states (Alabama, Texas, Iowa, and California) and was compiled and analyzed. A total of 1,775 violations were observed over 554 hours (for an observed rate of 3.2 violations/hour). Additionally, fourteen geometric and traffic operational characteristics were recorded for each intersection. The results of this research present several regression equations (linear, curvilinear, and multiple linear) which can be used to predict the red light running frequency that can be expected at an intersection approach based upon several geometric and traffic operational characteristics. Variables showing predictive power included average daily traffic (ADT), number of approach lanes, speed limit, number of lanes crossed by the approach, and distance to preceding and following intersections.
Abstract