Abstract
We examined the use of heading, path trajectory, tau-equalization, target drift canceling, and target egocentric direction strategies in visual control of steering toward a goal. The display (110°H × 95°V) simulated an observer walking over a ground plane at 2 m/s. Observers used a joystick to change the curvature of their traveling path to steer toward an environmental target under two display conditions (sparse flow: the ground was composed of random dots; dense-flow: the ground was textured). In Experiment 1, the simulated observer gaze direction was fixed on the target. The target position was thus fixed on the screen and observers could not perform the task by canceling the target drift or centering the target with respect to their perceived straight ahead. Observers did not rely on the tau-equalization strategy (equate the time-to-closure and the time-to-passage) or a path-based strategy (point their expected path at the target) to perform the task. Instead, observers steered their heading toward the goal with a shorter response delay for the dense than the sparse flow displays. In Experiment 2, the simulated observer gaze direction was fixed at 10° to the left or right of the observer's instantaneous heading. The target thus moved on the screen during steering control. Three groups of observers were tested with one group instructed to steer toward the target, one group to steer to center the target with respect to their perceived straight ahead, and one to steer to cancel the target drift. Steering to cancel the target drift generated a different pattern of performance data from steering toward the target or to center the target. Mean final heading errors for the latter two tasks were similar and significantly less than 10°. We conclude that humans use both heading and egocentric direction strategies in steering toward a goal and discrepant optic flow can adapt our perceived straight ahead direction.
Supported by: Hong Kong Research Grant Council, HKU 7478/08H.