Abstract
A large portion of current and previous research on locomotor and vehicle navigation tends to assume that people choose a goal or destination in the visual world and then generally look where they are going. There exists, however, considerable anecdotal evidence and observational data suggesting that humans also will tend to go where they are looking. Considering the amount of time a pedestrian or driver spends looking away from his precise heading point, this tendency has received relatively little experimental attention. The goal of the present research is to determine how gaze eccentricity affects drivers' abilities to steer a straight course. A high-performance virtual reality theater was used to simulate the motion of a car through a textured environment with the participant controlling direction of travel via a forced-feedback steering wheel. Participants (n=12) were asked to fixate a Landolt-C figure in one of 7 positions (0, +/- 15, 30, or 45 degrees from center) and drive down the center of a perfectly straight road. The Landolt-C was located just above the horizon in a fixed position on the viewing screen. Throughout each trial, the orientation of the figure varied randomly between 4 possible positions (0, 90, 180, and 270 degrees) and, in order to ensure fixation, subjects were required to respond to particular orientations. Lateral position of the driver was recorded for each of the different gaze eccentricities. Significant deviations from straight ahead were found for side of presentation when compared to fixation at 0 degrees (p<0.01). That is, when participants fixated to the left, for example, they systematically steered in the same direction. These results are compared to another study in which subjects' performance was measured while their head movements were restricted using a head-strap and chin-rest. The similar pattern of results in both conditions will be discussed in terms of the influence of retinal flow on the control of locomotion.
Supported by a J. William Fulbright Fellowship to WOR.