Predicting Speed-Related Traffic Violations on Rural Highways

Abstract

Speeding is considered as the most frequent violation type among the other traffic offences. This fact may occur due to the wide prevalence of speed enforcement among the road network. This paper aims to investigate criteria for selecting the best locations of speed cameras on rural highways. Data were collected at 76 sites of existing fixed speed cameras on rural highways of the Emirate of Abu Dhabi during a period of three months. The historical speeding violations, traffic information, aspects of the speed camera, and road characteristics were collected at each site. The statistics show that about 8,110,985 speeding violation tickets were issued between the years from 2008 till 2015 which represents around 80% of total violations. A model to predict the frequency of speeding violations was developed by using negative binomial regression approach. About fifteen independent variables/predictors were examined and are seen to significantly affect the occurrence of speeding violations frequency at a confidence level of 95%. These variables include trafficrelated variables (i.e. traffic volume, average speed, and percentage of trucks); site and camera's characteristics related variables (i.e. posted speed limit, speed margin, direction of enforcement camera, straight road segment, existence of speed change zone); and types of day. These findings can be used as selection criteria to find the best locations for installing speed cameras in the future enforcement programs.