Abstract
Following on from our work investigating gaze and steering along simulated roadways (Wilkie & Wann, 2003; 2004) we examined gaze behaviour in a steering task that supported a greater degree of route selection. We integrated a bicycle with our simulation rig that allowed more precise measurement of the paths taken, due to narrowness of the bicycle wheel. Participants steered around a series of obstacles in a simulated slalom of varying complexity. We recorded patterns of gaze sampling alongside the mean passing distance from obstacles, the overall variability of paths, and the smoothness of each trajectory. Gaze responses were categorized as either gaze ‘fixations’ on the most proximal obstacle (On), gaze ‘polling’ of more distal obstacles (a saccade to On+1 and then refixation of On) or gaze ‘shifting’ (saccade and fixation of On+1 with no refixation of On). Increased course complexity caused more steering errors, with greater path variability and reduced steering smoothness. The mean timing of gaze shifts did not change, though there was variability across obstacles (reflecting the changing demands of the task) and participants (related to their steering skill). Gaze polling was only intermittently observed, except in one participant who was particularly well practiced. She also displayed the greatest accuracy and smoothness in steering, suggesting that polling is only employed at an advanced state of locomotor control. We then examined how enforced timings for gaze fixations and shifting disrupted steering. These results and the relationship between gaze polling, fixation shifting and steering performance are discussed in the context of our current steering model (Wann & Wilkie, 2004).
Research supported by the UK EPSRC GR/S86358.