Navigational Training in Virtual and Real Buildings

Abstract

Three experiments explored aspects of virtual reality (VR) that may influence its utility for training in navigating architectural spaces. Taken together, these experiments suggest that the utility of virtual reality depends on characteristics of the user-virtual environment interface. In experiments testing the importance of movement control, full participant control of rotational and forward movement in VR led to nearly optimal training on a medium-sized building, but limited control in VR was worse than no control. Transparent walls, a VR possibility that goes beyond real-world options, were beneficial during training and on a direct measure of learning spatial layouts. Finally, in an experiment examining both training and transfer, VR training was similar to real-world training, and was quicker and transferred better to navigating a real-world building than floor-plan training. However, this held only when the testing route was traveled in the trained direction.