Probabilistic, safety-explicit design of horizontal curves on two-lane rural highways based on reliability analysis of naturalistic driving data
Accident Analysis and Prevention , Volume 123 p. 200- 210
The high collision rates on horizontal curves compared to other roadway elements make them one of the most critical elements in a transportation network. In this regard, it is important to develop models to predict the safety performance of the horizontal curves. A considerable number of studies have been conducted to develop safety performance functions based on several concepts such as geometric characteristics, design consistency, reliability analysis, and comfort threshold. However, these models do not account for all horizontal curve design criteria or consider several cases such as driving in adverse weather conditions or on pavement of low available friction. This paper develops a probabilistic, safety explicit approach of horizontal curve design using reliability analysis of four design criteria: vehicle stability, driver comfort, sight distance, and vehicle rollover. Two situations were considered in the analysis: driving in clear weather (dry pavement) and raining weather (wet pavement) to develop safety performance functions for annual and five-year collision frequency. Four types of regression models, Poisson, negative binomial, zero-inflated Poisson, and zero-inflated negative binomial, were used in the analysis. The AIC, BIC, and Vuong test were used in evaluating the developed models.
|Horizontal curves, Naturalistic driving study, Probabilistic design, Probability of failure, Reliability analysis, Road safety, Safety performance function, Safety-explicit design|
|Accident Analysis and Prevention|
|Organisation||Department of Civil and Environmental Engineering|
Dhahir, B. (Bashar), & Hassan, Y. (2019). Probabilistic, safety-explicit design of horizontal curves on two-lane rural highways based on reliability analysis of naturalistic driving data. Accident Analysis and Prevention, 123, 200–210. doi:10.1016/j.aap.2018.11.024