UC Berkeley

The goal of this work is to develop an approach to the design of natural and immersive interaction methods for three-dimensional virtual environments. My thesis is that habitation and presence in any environment are based on a continuous process of perception and action between a person and his/her surroundings. The current practice of virtual environments, however, disconnects this intrinsic loop, separating perception and action into two different `worlds'--a physical one (for perception) and a virtual one (for action). This research is aimed at bridging the gap between those two worlds.

Being drawn from perceptual philosophy and psychology, the theoretical study in this dissertation identifies three embodiments of natural perception-action loop: direct perceptual acts, proprioceptive locomotion, and motor intentionality. These concepts form the basis for the interaction methods proposed in this work, and I demonstrate these methods by implementing pertinent prototype systems. First, I suggest a view-dependent, non-planar display space that supports natural perceptual actions, thereby enhancing our field of view as well as depth perception. Second, I propose Interactive Chair, which provides an intuitive locomotion control based on our sense of self-body movements. Third, I argue that pointing-based interaction techniques inhibit our motor-intentional behaviors, therefore demonstrate a line-of-sight, direct object manipulation method. I finally conduct a series of user tests to evaluate the proposed methods and systems, and confirm the contribution of suggested interaction design approaches to the natural experience of virtual environments.