Driver perception of horizontal curvature in sag vertical curves

Driver visual perception of the road has long been considered a very important issue in human factors research. Confusing or misleading visual cues may result in erroneous driver operation and increase the risk of accident. In previous research work, drivers have been presented with computer animations of horizontal curves superimposed on sag vertical curves and the reported perceived radius has been examined. However, further research is still necessary in order to fully understand the parameters affecting driver misperception of curvature of a horizontal curve overlapping a sag vertical curve. In this research, a suitable experiment has been designed, based on a "Graeco-Latin Square" statistical design. Photorealistic computer animations of horizontal curves superimposed on sag vertical curves, with preselected geometric design parameters (according to the "Graeco-Latin Square" statistical design), were presented to a sample of drivers, with similar curves superimposed on flat profile as reference curves, and the drivers were asked to compare the two curves with regard to their horizontal curvature. The paper discusses the findings of the experiment, regarding the parameters affecting the probability of erroneous perception of curvature and how changes in the values of these parameters are reflected to the probability of misperception of curvature.