Abstract
Recent research has suggested that both egocentric direction and optic flow are used to guide locomotion to a goal (Warren et al., 2001; Wood, et al., 2001; Harris & Carre, 2001). However, the contribution of the optic flow strategy should depend on the velocity of the flow field. In a pair of experiments, we examined the relative contributions of egocentric direction and optic flow while varying the effective flow field velocity. Participants were tested in a 12 m × 12 m virtual environment, wearing a head-mounted display (40 deg H × 60 deg V), with head position measured by a hybrid sonic/inertial tracking system. To dissociate the two strategies, the heading direction specified by optic flow was shifted 10 degrees (randomly to the L or R) from the actual direction of walking. In the first experiment, participants walked to a target located 9 meters away in physical space, placed at one of three angles (3, 8, or 13 degrees) away from the initial heading. Velocity of optic flow was manipulated by changing the “gain” between physical and virtual space so that the flow velocity was either 0.5, 1, 2, or 4 times the physical walking speed. Preliminary results indicate that the optic flow strategy was more dominant at higher flow velocities than at low velocities. In experiment 2, participants again walked to a target located 9 meters away in physical space while avoiding one to three obstacles en route, in the same visual gain conditions. Fajen & Warren's (submitted; VSS 2001) dynamic model of steering and obstacle avoidance will be expanded to incorporate both egocentric direction and optic flow strategies.
NIH EY10923, NSF LIS IRI-9720327