Machine Learning-Aided Management of Motorway Facilities Using Single-Vehicle Accident Data

Management of expressway networks has been mainly focused on defect management without looking at the correlations with accidental risks. This causes unsustainability in expressway infrastructure maintenance since such defects may not be a contributing factor toward public safety. Thus it is necessary to incorporate accidental events for decision-making in infrastructure management. This study has developed a novel approach to machine learning (ML) that incorporates actual primary data from the last 10 years of single-vehicle accidents (SVA) by collisions with motorway facilities, or so-called single-vehicle collisions with fixed objects. The ML is firstly aimed at identifying the influential factors of SVA in relation to finding effective countermeasures for accidents by integrating the correlation analysis, multiple regression analysis, and ML techniques. The study reveals that wet pavement conditions have a significant effect on SVA. The results show that improvement of the skid resistance is the most effective method to reduce SVA when the average vehicle speed (AVS) is less than 60 km/h. At the locations with gentle curve radii, ML indicates that it is crucial to redesign the speed-through management. Interestingly, the real data over 10 years indicate no relationship between equivalent single axle load (ESAL) and skid resistance, although many other studies have demonstrated the inverse relationship. In this study, the novel ML mean demonstrates excellent capability in providing suitable countermeasures for a reduction of SVA under a variety of uncertain and road quantitative aspects. The ML-based mitigation policies can also be applicable to other motorways and can contribute to their road safety, underpinning sustainable transport systems.