Risk assessment of horizontal curves using reliability analysis based on Google traffic data

Abstract

The risk associated with horizontal curves makes them one of the most critical elements in the transportation network compared with other roadway elements. In this regard, the current geometric design practices of horizontal alignment need to be reviewed due to ignoring the uncertainty of design inputs resulting from drivers’ behavior. Besides, numerous studies utilized the regression speed prediction models developed from several studies in different countries to estimate the operating speed on horizontal curves. However, there are differences in road characteristics, environmental conditions, and drivers’ behavior that differ from one country to another. Thus, this paper presents a new technique based on Google Maps Application Programming Interface (GMAPI) to estimate reliable operating speeds and validate the regression models. It also introduces the reliability analysis approach to overcome the shortcomings of current geometric design practices and evaluate the risk levels on different horizontal curves. Reliability analysis is performed on five horizontal curves located along Gamassa-Al Mansoura road, in Dakahlia, Egypt. The results confirm that the GMAPI method is more accurate than the regression models. It also proves that the regression models are not recommended for all countries. Moreover, this study provides the designers with risk level charts for the design of horizontal curves based on the reliability approach. These charts could be very useful for highway designers to easily predict the safety margin for different design alternatives.