Evaluating the safety risk of roadside features for rural two-lane roads using reliability analysis.

The severity of roadway departure crashes mainly depends on the roadside features, including the sideslope, fixed-object density, offset from fixed objects, and shoulder width. Common engineering countermeasures to improve roadside safety include: cross section improvements, hazard removal or modification, and delineation. It is not always feasible to maintain an object-free and smooth roadside clear zone as recommended in design guidelines. Currently, clear zone width and sideslope are used to determine roadside hazard ratings (RHRs) to quantify the roadside safety of rural two-lane roadways on a seven-point pictorial scale. Since these two variables are continuous and can be treated as random, probabilistic analysis can be applied as an alternative method to address existing uncertainties. Specifically, using reliability analysis, it is possible to quantify roadside safety levels by treating the clear zone width and sideslope as two continuous, rather than discrete, variables. The objective of this manuscript is to present a new approach for defining the reliability index for measuring roadside safety on rural two-lane roads. To evaluate the proposed approach, we gathered five years (2009-2013) of Illinois run-off-road (ROR) crash data and identified the roadside features (i.e., clear zone widths and sideslopes) of 4500 300ft roadway segments. Based on the obtained results, we confirm that reliability indices can serve as indicators to gauge safety levels, such that the greater the reliability index value, the lower the ROR crash rate.