Validation of Perspective-View Concept for Estimating Road Horizontal Curvature

Abstract

In three-dimensional (3D) alignments, some road geometric parameters, such as vertical curve type (crest or sag), can cause drivers to have visual illusions in perceiving the horizontal curvature that may result in erroneous decisions. Road curvature estimation is usually made based on the perspective-view (PV) information. It is hypothesized that drivers can estimate road curvature visually based on the openness magnitude of the inside edge lines which appear to the drivers as parabolas or hyperbolas. This paper further develops the PV concept and validates it using the published results of driver perceptions of 3D alignments. The analysis shows that there are statistically good relationships between the ratio of the 3D perspective radii of the crest (or sag) and flat horizontal curves, and driver perceptions. Preliminary criteria for the design of 3D alignments based on driver perceptions are presented. The PV method provides a means of incorporating driver perception into geometric design, and therefore should be of interest to highway designers and researchers.