The impact of the combination equilibrium of horizontal and sag-vertical curves on safety

Abstract

Safe and aesthetic interaction between horizontal and vertical alignment may significantly occur when they are combined or closed to each other resulting in safety problems. Previous researches have been limited to focusing on qualitative analysis, which are difficult to implement in the design. To assess the quantitative impact of the combination equilibrium of horizontal and sag vertical curves on safety, KNN, SVR and KNN-SVR machine learning methods were applied for model training to analyze the influence of alignment combination of the horizontal curve (HC) and the sag vertical curve (SVC) on accidents. The highway alignment (a total of 1208 km), the traffic data and accident data from 2011 to 2018 of six interstate roads in Washington, D.C. have been used to train the models. The combination equilibrium of horizontal and sag vertical curves is expressed by the variables such as the radius of the horizontal and vertical curves, the length of the horizontal and vertical curves and the dislocation of horizontal and sag vertical curves. KNN model, SVR model, and KNN-SVR model were built by training the variables as well as the accident rate per 100,000,000 vehicle kilometers. The results show that for the HC-SVC alignment combination, the KNN-SVR model has higher accuracy in predicting the accident rate. At the same time, this paper also suggests the value range of the variables when the horizontal curve radius is small. The research conclusions can provide a reference for the subsequent quantitative optimization design and safety improvement of horizontal and vertical alignment combination.

Keywords

• Highway
• Combination equilibrium of horizontal and sag vertical curves
• Machine learning
• Highway safety