Paper presented at the meeting of the American Psychological Society, 1996 in San Francisco, CA.
Abstract
Touching a virtual object improves the quality of the virtual experience. For some objects, subjects saw a virtual object and touched its real counterpart. Compared to visual VR only, converging evidence from both visual and tactile senses increased the illusion of "being in a place" when experiencing the virtual environment.
Introduction.
VR systems can create a compelling illusion that one is inside a computer-simulated environment, rather than standing in a lab room wearing a VR helmet. The level of presence experienced is one measure of the quality of the interface. We investigated the effect of touching virtual objects on subjects' sense of presence in the virtual environment. We placed a real object (e.g., a rubber ball) within the participant's grasp at the location of the virtual object. As their cyberhand explored the illusory ball in cyberspace, their real hand explored the real ball in the real world, a technique we call "tactile augmentation" (see Figure 1). If the visual image is dominant, the additional tactile cues might be attributed to the virtual object, enhancing subjects sense that they are in the computer-simulated environment. In contrast, touching the real object could distract attention back to the labroom, decreasing subjects' sense of presence in VR. Based on perception literature showing a strong tendency for visual input to dominate in the process of unifying visual and tactile sensory input (e.g., Miller, 1972), we predicted that tactile augmentation would enhance presence, improving the quality of the human-computer interface.
Method
Subjects. Fourteen students from the U. of W. participated. Materials and equipment Eight real objects (e.g., butter knife) were modelled in 3-D, texture mapped, and scaled to size. The VR system consisted of a Division ProVision 100, coupled with a Division dVisorTM head mounted display.
Design and Procedure. A position sensor attached to the subject's hand allowed synchrony between subject's hand movements, and cyberhand movements. A within-subjects design was used1. For the "see and touch" condition, subjects saw a virtual image, and were also able to physically touch it (e.g., a carefully placed rubber ball, see Figure 1). For "see only" objects, subjects saw a 3-D virtual image (e.g., a blue ball) that they attempted to touch with their cyberhand (to no avail, see Figure 2). Each subject "saw and touched" 4 objects, and "saw only" 4 objects. In the course of the experiment, each object was assigned to each of the two conditions approximately equally often The order in which the objects were presented was randomized. Subjects then completed a questionnaire requiring them to rate on a scale from 1 (low) to 7 (high) how present they had felt in each of the conditions.
Results and Discussion
A mean presence rating for "see and touch" trials was calculated and entered into the analysis for each subject, as was a mean for "see only" trials. Subjects rated their sense of presence higher for "see and touch" trials compared to "see only" trials (means = 4.9 vs 3.9 respectively). A Wilcoxon, signed-rank test showed a highly significant difference between the two conditions, Z = 2.95, two-tailed p < .005. The pattern was the same for each of the five presence questions used.
Conclusion
As predicted, input from real objects, a form of "mixed reality" (Milgram, & Kishino, 1994), enhanced subjects' sense of presence in VR. In contrast to expensive, computer generated force-feedback and teletaction (e.g., Caldwell, Wardle and Godwin, 1994) tactile augmentation costs very little money, requires less computer processing, and the physical textures of the real objects (e.g., the rough feel of a coconut) are hard to reproduce in computer simulations.
The HITLab is exploring ways of making VR as effective as possible for training, for information display (e.g., optimizing situational awareness in fighter pilots), for learning (e.g., the U.S. West Virtual Reality Roving Vehicles project studying educational applications of virtual reality), medical applications, architectural design, and for immersive telecommunication (the GreenSpace project). Tactile augmentation is one way to enhance the quality of VR as a human-computer interface.
References
Caldwell, D.G., Wardle, A., and Godwin, M. (1994). Telepresence: visual, audio, and tactile feedback and control of a twin armed mobile robot. Proceedings of IEEE International Conference on Robotics and Automation, San Diego, CA. IEEE Computer Society Press. Los Alamitos, CA.
Milgram, P. & Kishino, F. (December, 1994). A taxonomy of mixed reality visual displays. IEICE Transactions on Information Systems, Vol E77-D. Miller, E.P. (1972). Interaction of vision and touch in conflict and nonconflict form perception tasks. Journal of Experimental Psychology, 96, 114-123.
Acknowledgements
This research was supported by AFOSR grant #F49620-93-1-0339. and by the U.S. West Foundation. Thanks to Boris Kogon and Keith Hullfish for help programming the virtual objects and room.
1. A replication of this experiment using a between-subjects design (to reduce demand characteristics) is in progress.