| Publications Page | HITL Home |
Engineering is increasingly limited by information complexity. This problem can be addressed in part by increasing the human-computer bandwidth. Virtual interfaces (VI's), which make more complete use of human sensory input than traditional interfaces, are a potential means to increase this bandwidth. However, two fundamental problems face the effective use of VI's. The first is avoiding side-effects, such as simulator sickness and postural after-effects. This problem is addressed in related research. The second problem is understanding how to design effective virtual interfaces. Optimizing virtual interfaces requires robust measures for how good they are. Presence, the sense of ``being in'' a virtual environment, appears to be related to the intuitiveness of a VI. Hence, a good presence measure may be one indication of the quality of a VI. It is suggested that a sensitive Class A (``objective'') measure for presence may be possible in terms of the ability of virtual cues to perceptually overwhelm conflicting real cues. Two pilot studies, based on conflicting visual and inertial rotation cues, were conducted. These studies each compared two visual conditions (Wide vs. Narrow field-of-view and Rightside-up vs. Upside-down). Preliminary data indicate that large between-condition differences in the ability of visual cues to perceptually overwhelm inertial cues (factor of 2-3) can be detected in about 1/3 of subjects. These differences were much less apparent on questionnaire data, suggesting that the perceptual measure may be a more sensitive means to assess presence. These results will be discussed in terms of the ``rest frame hypothesis'', which links in a particular way presence; motion sickness; and ``illusions'' such as vection and perceived tilt. More robust and convenient perceptual measures of presence are under investigation.