Relationship between design speed and operating speed at horizontal curves on suburban arterials

Abstract

Much research has been conducted in recent years to examine the relationship between design speed and operating speed at tangent and horizontal curve sections. The majority of previous research has focused on rural two-lane highways and urban residential streets and collectors. Little if any research has considered the problems related to horizontal curves on suburban arterials. Lack of information and the results of previous studies prompted this research focusing on current design practices on suburban four-lane arterials. Roadways of the suburban arterial functional class were identified and reviewed to decide their appropriateness for inclusion in the study. The experimental design included a procedure for collecting speed data on both tangent and horizontal curve sections to examine driver behavior associated with these curves. Two primary independent variables, curve radius and approach density, were examined throughout this study. Regression analyses were performed to examine their effect on 85th percentile speeds on tangents and curves. The results indicated that the 85th percentile speeds on tangents decreased as approach density increased. Comparisons were made between design and operating speeds for the curves examined. For the comparisons made on suburban arterials, the 85th percentile driver usually exceeds the design speed and corresponding side-friction factor on curves with inferred design speeds less than 70 km/h (43.5 mi/h). The evaluations of driver speed behavior on horizontal curves suggested that curve radius is a good predictor of 85th percentile speed on curves. It was determined that a curvilinear relationship exists between curve radius and 85th percentile speed on curves. For the curves examined, the 85th percentile speed on curves increased as curve radius increased. The results also show that approach density is significant in contributing to the explanatory power of the relationship. Additional analyses were performed to examine the effect of curve radius and approach density on mean speed reductions at curves. Previous research suggests that mean speed reductions decrease as curve radius increases. The regression results do not indicate a statistically significant effect of curve radius and approach density on mean speed reductions. The combined variability of speeds on curves and speeds on tangents probably caused the poor results, given the small number of sites examined.