Abstract
Control of steering is an essential component of successful high speed locomotion. Land & Lee (1994) recorded gaze behaviour during driving and observed extensive use of the tangent point of upcoming bends. In Wilkie & Wann (2003), however, we observed that gaze exploration of the centre of the road 1–2 seconds ahead was associated with accurate steering. Here we examine the relationship between gaze and steering behaviour by simulating locomotion along roads of varying width and curvature. Nine participants were asked to steer as close as possible to the centre of a computer-simulated roadway when traveling at 13.8m/s (50km/h). Two road widths (3m or 6m) and two bend curvatures (constant radii of 60m or 120m) were used. Each participant steered initially down 24 roads with randomly interleaved width and curvature (6 trials of each) and were free to look wherever they like in the scene. We then added a fixation requirement to the task and repeated these conditions. A fixation cross could either be on the tangent point of the bend, or on the centre of the road the same distance ahead as the tangent point. We also ran conditions with fixation on the outside edge of the road, or in the space between the centre and the road edges, or 3m beyond the road edges. Throughout trials we recorded position relative to the road centre and gaze behaviour. The results show that steering performance was influenced by both road curvature and width (the largest errors occurring on tight wide roads). The point of fixation also influenced steering, causing systematic biases in the direction of gaze. We conclude that the tangent point of the road does not provide information which is uniquely useful, but rather that looking to the point you wish to pass through contributes to successful steering.
Research supported by the UK EPSRC GR/R14644 & EP/D055342/1