Abstract
Both retinal flow and visible road edges are potentially useful sources of information when steering down a road. Influential models of driving have identified a role for “near” and “far” road edge components, but it has been shown that flow is used even when complete road edges are present (Kountouriotis et al., 2013). The aim of the present study was to assess whether flow is used when different road edge components were available. We tested 20 participants steering through a virtual reality computer simulated environment while fixating a point on the road ~2s ahead. Flow was systematically manipulated independent of the veridical road edges, so that use of flow would lead to predictable understeer or oversteer (the flow speed was altered proportional to veridical flow by: .5, .75, 1, 1.25, 1.5). We also varied road edge information by using 7 combinations of Near road (‘N’: up to .5s ahead), Middle road (‘M’: from .5s to 1s ahead) and Far road (‘F’: from 1s to horizon) information. As predicted, steering systematically changed according to flow speed, but the magnitude of the flow-induced bias varied depending on the road-edge components that were visible. The presence of Far road increased the influence of flow, with flow bias most evident in F and N+M+F conditions. Similar results were observed when participants were free to look where they preferred. It seems that flow predominantly interacts with prospective steering control rather than the immediate correction of steering errors.
Meeting abstract presented at VSS 2015