Identifying hazardous road sections using a fuzzy expert system

Abstract

To identify hazardous sections of newly built mountainous highways, decision-makers require assistance to determine the sections that may pose risks to road users, thereby enhancing the level of road safety management and capital utilisation efficiency. In this research, traffic accident economic losses were used as evaluation indices to characterise 30 horizontally curved sections of roadways. These highway sections were divided into five levels based on their evaluated level of risk. With the effects of multiple factors and the characteristics of uncertainty, mathematical statistics and fuzzy expert systems (FESs) were used to determine the factors contributing to hazardous sections of mountainous highways. The membership functions and fuzzy rule base were decided according to accident data and expert experience and a method for identifying hazardous sections based on the expert systems was established. The risk values of another 35 horizontally curved sections were obtained by this method and good results were achieved using the method to identify hazardous sections. The results showed that hazardous section classification is greatly affected by the horizontal radius, grade and declination angle of the horizontal curve. The results obtained from the FESs agreed overall with actual accident data.