This study investigates methods of estimating visual demands during driving in road traffic environments. Accurate estimation of visual demands is expected to promote the identification of driving situations in which drivers have sufficient capacity to divide their attention to in-vehicle systems without resulting in distraction or inattention. Two experiments were conducted to identify occlusion methods suitable for assessing the visual demand variations derived from traffic conditions and road structures. First, four occlusion methods were compared to estimate visual demands during driving with and without lead vehicles: (1) depression of a switch allowed the road scene to be viewed for 0.6 seconds, (2) depression of a switch blanked out the driver’s vision for 1.5 seconds, (3) the driver was allowed to view the road while depressing a switch, and (4) the driver’s vision was blanked out while depressing a switch. The results of the first driving simulator experiment suggested that differences in the proportions of viewing time to driving time were higher in occlusion methods 2 and 4. Second, we estimated the visual demands of road structures (straight sections, curves, etc.) using the two occlusion methods identified in the first experiment. The second simulator experiment suggested that drivers viewed the roadway throughout the curve in occlusion method 2, while some drivers occluded the forward scene for a very short time even during the curve in method 4. The findings suggest that the occlusion method in which the scene is invisible for a specific time at the driver’s request may contribute to establishing the conditions in which the driver can safely operate in-vehicle systems. (Author/publisher)
Abstract