3D Sound Rendering in a Virtual Environment to Evaluate Pedestrian Street Crossing Decisions at a Roundabout

Crossing streets is a potentially hazardous activity for pedestrians, and it is made more hazardous when the complexity of the intersection increases beyond a simple linear bisection, as it does in the case of a roundabout. We are interested in how pedestrians make decisions about when to cross at such intersections. Simulating these intersections in immersive virtual reality provides a safe and effective way to do this. In this context, we present a traffic simulation designed to assess how pedestrians make dynamic gap affordance judgments related to crossing the street when supplied with spatialized sound cues. Our system uses positional information to generate generic head-related transfer functions (HRTFs) for spatializing audio sources. In this paper we evaluate the utility of using spatialized sound for these gap crossing judgments. In addition, we evaluate our simulation against varying levels of visual degradation to better understand how those with visual deficits might rely on their auditory senses. Our results indicate that spatialized sound enhances the experience of the virtual traffic simulation; its effects on gap crossing behavior are more subtle.