Abstract
In full-cue, real-world environments, people are accurate at visually-directed actions indicating absolute egocentric distance to targets on the ground. The same distance estimation tasks conducted in virtual environments using head-mounted display (HMD) systems have shown that people act as though the environment is smaller than intended. In some cases, judged distances are near 50% of intended distance. The explanation for this difference between actions in real and virtual environments is unknown. Our current experiment investigates the hypothesis that factors involved with wearing a head-mounted display, such as reduced field of view, increased mass, torques, and moments of inertia around the head of the participant might be contributing to the misperception of target location. This study involved creating a mock HMD that has the same fit, feel, and approximate field of view as the real HMD. We outfitted a HMD shell with a distribution of weights that matched the mass and moments of inertia in the real HMD. Visual occluders were used in place of the optics to afford participants a similar field of view. Subjects performed either a direct walking task or a triangulated walking task without vision to previously seen targets on the ground in either a real room with the mock HMD, or in a model of the same room with the real HMD. On par with previous work, both tasks conducted in the virtual environment with the real HMD showed underestimation of distances. For the tasks conducted in the real-world room using the mock HMD, an underestimation of distance was also observed, though of lesser magnitude than in the corresponding virtual world conditions. The results suggest that factors associated with wearing the HMD may account for some, but not all, of the apparent compression of space in virtual environments.
This work was supported by NSF grants IIS-00-80999 and IIS-01-21084.