Abstract
Walking interfaces for Virtual Reality often produce proprioceptive, vestibular and somatosensory signals which conflict with the visual presentation of terrain conditions in virtual environments. We compared locomotion decisions made using a dual joystick gamepad with a walking-in-place metaphor. Each trial presented two choices where the visual path condition differed in one of the following aspects: (a) incline, (b) friction, (c) texture, and (d) width. Users chose one of these paths by using the locomotion interface to walk to a goal. Their decisions were recorded and analyzed as a generalized linear mixed model. The results suggest that the walking-in-place interface produces choices of visual conditions that more often reflect expectations of walking in the real world: decisions that minimize energy expended or risk of injury. Because of this, we can infer that different walking interfaces can produce different results in virtual reality experiments. Therefore, behavioral scientists should be wary that sensory discrepancies between visual presentation and other modalities can negatively affect the ecological validity of studies using virtual reality. Consideration should be taken designing these studies to ensure that sensory inputs are as natural and consistent between modalities as possible.