Virtual equivalence : matching visual scene and treadmill walking speeds in virtual reality

Abstract

(cont.) The last experiment showed that individual subjects PER values changed over time intervals as short as ten minutes, and revealed the importance of the subject's prior experience in PER experiments. This suggests that limitations of working memory may effect the repeatability of the PER measure. The definition and measure of PER for each subject may also provide a means for quantifying the magnitude of a clinical condition known as oscillopsia, where we perceive the world as non-stationary, such as moving independently of our head motions. These findings are important for a perceptually sensitive environment, such as virtual reality. Designers of virtual environments that utilize self-motion perception should consider calibrating PERs within a session for each individual user and be aware that that the subject's calibration may change over time.

If we walk on a treadmill and are looking in the direction of motion of a moving virtual environment, the perceptions from our various senses are harmonious only if the visual scene is moving in a narrow range of speeds that are, typically, greater than our walking speed. This observation has been reported when we project a virtual environment through a display with a restricted field-of-view, such as a head-mounted display (HMD). When the subject feels that the scene-motion is natural for their walking speed, the ratio of the speed of his visual surround to that of the treadmill walking speed is defined as his perceptual equivalence ratio (PER) in that setting. Four experiments explored a set of conditions under which the PER measured on a treadmill is constant. The experiments, motivated by several hypotheses, investigated the relationship between PER and display type (HMD vs. either desktop monitor or on-screen projection), sense of presence in the virtual environment, and the magnitude of illusory self-motion (vection). We also investigated differences among subjects, and the stability of PER over time due to the limitations of working memory. Most experiments considered more than one hypothesis. The first two experiments found that PER was affected by the type of display used, but found no correlation of PER with the sense of presence reported by the subject. A third experiment showed that PER was nearly the same whether we manipulated visual or treadmill speed (and asked the subject to match the other.) While PER values were often constant versus treadmill speed for any individual subject, they were very different from subject to subject. PER appears to be relatively stable over a short test session, but may be highly variable over extended periods of time.