Presentation of the Accident Predictive Model Based on Speed Hump Characteristics in Urban Intersections

The high statistics of accidents compared to the population and vehicle ownership index in Iran has put reducing the number of accidents more than ever before on the agenda of road engineers. Speed is one of the most effective factors in causing accidents. One of the physical tools for controlling the speed of vehicles is the use of speed bumps and humps in the streets. This study examines the effect of proper implementation of speed humps on reducing the speed of vehicles, as well as reducing the number of accidents. To this end, the speed of vehicles before and after installing 64 speed humps in areas 2 and 10 of Mashhad was recorded through field research. In addition, the physical specifications of speed bumps, such as width, height, distance from intersections, number of lanes on the streets, and distance from warning signs to speed bumps, were measured. Then, using statistical analyses, the effectiveness of each of these parameters in reducing vehicle speed was evaluated. Furthermore, based on the accident statistics of the targeted streets, data before and after installing speed bumps were analyzed using SPSS software and regression analysis. After conducting the initial analysis, it was found that none of the linear regression models were suitable for predicting the accidents. Therefore, the data were analyzed using a nonlinear multiple regression model (R^2=0.87). The results of the analysis showed that the use of speed bumps can have a significant impact on reducing accidents in streets and intersections, provided that proper safety measures are taken into account in their design and installation. In fact, speed bumps can lead to a reduction of approximately 17% in accidents.