Safety evaluation of horizontal curves on two lane rural highways using machine learning algorithms: A priority-based study for sight distance improvements

Abstract

Objective

Current Road geometric design processes disregard stochastic aspects, resulting in inadequate traffic safety considerations. In addition, the primary sources of crash data are obtained from police department, insurance agency and hospitals, where detailed investigation is not carried out from a transportation perspective. So, the data obtained from these sources may or may not be reliable. The main objective of this study is to account uncertainties using reliability as a tool that considers slow down the vehicle while they maneuver the curve and to develop thresholds of reliability index associated with sight distance based on operating speed design consistent measure as a surrogate safety measure rather than using crash data.

Methodology

This study proposes a thresholds of reliability index associated with sight distance for various operating speed ranges based on design consistent measure. In addition, established the relationship between the consistency levels, geometric characteristics and vehicle characteristics. In this study, classical topography survey was carried out on the field using total station. The data collected comprises speed data and geometric data of 18 horizontal curves (Lane based analysis was carried out). A total of 3042 free flowing vehicle speeds have been extracted from the video graphic survey and used in the analysis.

Results

The threshold values of reliability indices associated with sight distance is higher as the operating speed increases for the section to be a consistent design section. The results from the Binary Logit Model show that the consistency level is significantly affected by deflection angle and operating speed. Deflection angle was negatively correlated with the in-consistency level, and the operating speed was positively correlated with in-consistency level.

Conclusions

From Binary Logit Model (BLM) results, we can conclude that an increase in the value of deflection angle will significantly decrease the probability of In-Consistent level, which indicates uncertainties that cause drivers to change the path of vehicle or rate of deceleration of the vehicle while maneuvering the curve will be decreased. Whereas, increase in the operating speed will significantly increase the probability of In-Consistency level.

Keywords:

• Uncertainties
• reliability index
• sight distance
• operating speed
• consistency level

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.