SAFETY ASPECTS OF INDIVIDUAL DESIGN ELEMENTS AND THEIR INTERACTIONS ON TWO-LANE HIGHWAYS: INTERNATIONAL PERSPECTIVE

The results of an extensive literature review of the safety performances of low- and intermediate-traffic-volume, two-lane rural highways are presented. The effects on traffic safety, as measured by accident rates, of pavement width, radius of curve/degree of curve, gradient, sight distance, traffic volume, and design speed on curved sections of two-lane rural highways are covered. The following are some of the main findings. There is a distinct tendency for accidents to decrease with increasing pavement width up to about 7.5 m (25 ft). There exists a negative relationship between radius of curve and accident rate. The sharper the radius of curve, the higher the number of run-off-the-road accidents. Curves that dictate a significant change in operating speeds and that cause nonhomogeneity in road characteristics are especially dangerous. The most successful parameter in explaining the variability in accident rates was degree of curve (United States) or curvature change rate (Europe). Gradients of up to about 6% have a relatively small effect on the accident rate. A sharp increase in the accident rate was noted on grades of more than 6%. There exists a negative relationship between available sight distance and accident risk. However, other influencing parameters, such as wide pavements and gentle radii of curve, might also play a part in the observed positive effect of sufficient sight distances on the accident situation. For narrow road sections, an increase in sight distance could favorably affect traffic safety. A negative relationship between traffic volume and road traffic accidents was established. Run-off-the-road accidents were found to decrease with increasing average annual daily traffic up to 10,000 vehicles per day. Recent investigations reported a U-shaped distribution between accident rate and traffic volume. An accident rate of 2.0 accidents per 10 to the 6th power vehicle km (3.2 accidents per 10 to the 6th power vehicle mi) was proposed as a breakpoint between levels of safety and unsafety. This breakpoint was derived from relationships between single design parameters and accident rate as well as from the superimposition of the design parameters. Limiting values for a number of design parameters are also proposed. If these limiting values were exceeded, the proposed breakpoint, in relation to the accident rate, would be exceeded.