The objectives of
this study are to investigate and quantify the relationship between
design consistency and road safety for two-lane highways in the West
Bank. This study produced speed prediction method using real time
traffic speed data obtained from Google Earth maps, which were used to
estimate the 85th percentile speed along an alignment that includes both
horizontal curve sections and tangent sections. A comprehensive crash
and geometric design database of two-lane rural highways has been used
to investigate the effect of several design consistency measures on road
safety.
Previous studies showed that the most promising consistency measures
identified in previous research fall into four main categories, namely:
operating speed, vehicle stability, alignment indices, and driver
workload. Five crash prediction models, which relate design consistency
to road safety, have been examined. The generalized linear regression
approach has been used for model development. All models adopted in this
study showed acceptable levels of goodness of fit and over-dispersion.
The developed models verified that the main design consistency measures
have an important impact on safety. The consistency measures used in
model development are: variation between the design speed and the
operating speed, absolute difference of the 85th percentile speeds
between successive design elements, difference between side friction
supplied and demanded, average radius of curvature, average tangent
length, maximum radius of curvature to minimum radius of curvature,
curvature change rate, ratio of individual curve radius to average
radius of the section, and visual demand of familiar drivers of the
section.
Validation step was performed; the goal was not only to compare the
accuracy of different models developed, but also to evaluate the overall
accuracy of Crash Prediction Models for use on rural two-lane highways
in the West Bank. Validation requirement was to demonstrate that a model
is appropriate, meaningful, and useful for the purpose for which it is
intended.
The models can be used as a quantitative tool to evaluate the impact of
design consistency on road safety. An application is presented where the
effectiveness of crash prediction models, which incorporate design
consistency measures, is compared with those, which rely on geometric
design characteristics. The study concluded that models, which
explicitly consider design consistency, can identify the inconsistencies
more effectively and reflect the resulting impacts on safety more
accurately than those which do not. Finally, a systematic approach to
identify geometrically inconsistent locations using the safety
consistency factor has been proposed.