CS 5678: Topics in Mixed Reality

Symbolic input is a key challenge for Mixed Reality (MR) interaction, as the standard point-and-click QWERTY keyboard is inefficient and error-prone. Here, we implement and evaluate alternative methods for entering text using hand-held controllers. We propose ZoomText, a single-joystick novel solution that utilizes a circular layout with dynamic zooming and will compare its results to existing text entry methods. We compare its performance at three different zoom factors to point-and-click QWERTY, and two varieties of PizzaText. We find that ZoomText with a medium-to-high zoom factor is more accurate than all other techniques assessed and competitive with two-handed PizzaText on speed. However, point-and-click QWERTY was much faster than all other methods and greatly preferred by users. We conclude that there is reason for further exploration in circular, controller-based symbolic input techniques in MR, but that there are significant hurdles in user experience compared to more familiar methods.

Virtual Reality allows users to completely immerse themselves in a virtual world and experience these spaces as if they were truly part of it. Movement within these spaces are crucial for proper exploration, but continuous movement tends to cause cybersickness for a substantial proportion of users. Although methods such as teleportation and opaque vignettes have been popularized to avoid this problem, these may limit the level of immersion and awareness users should be experiencing while traveling within the virtual world. Here, we propose a method that restricts a user’s peripheral view with gradual blurs that decrease the perception of periphery movement while still maintaining a certain amount of spatial immersion and awareness for the user. Our result shows that the blurred vignette method reduces the level of cybersickness compared to the traditional continuous movement method, but did not achieve a significant higher presence score compared to opaque vignettes and teleportation.

Although raycasting is usually an intuitive and accurate selection technique in virtual reality environments, its performance deteriorates when a high degree of precision is required, such as when selecting small objects in densely-packed areas. To deal with this issue, some techniques such as ARM (Absolute and Relative Mapping) Raycasting and PRISM (Precise and Rapid Interaction through Scaled Manipulation) adjust the control/display (CD) ratio - either manually or dynamically - between the real and virtual hands to allow for higher precision. We propose an adaptation of raycasting, Density-Based Selection (DBS), where the CD ratio is dynamically adjusted according to the target density in the vicinity of the raycast, allowing for greater selection precision for densely-packed objects.

In this project, we analyzed the effects of replacement vestibular stimulation to forward motion in order to reduce the effects of motion sickness and increase immersion in VR. We examined two methods of body-centered locomotion, specifically jumping and a novel technique: “penguin waddle”, and compared them to the standard controller-centered techniques of smooth motion and teleportation. We then assessed the effects that these body-centered locomotion techniques had on user motion sickness, as well as the impact on the user’s immersive experience. We found no statistically significant difference between any of the conditions in terms of immersion. We found a statistically significant increase in motion sickness for the penguin waddle and smooth motion conditions compared to teleportation. We also found that our jumping condition did not increase motion sickness a significant amount compared to teleportation.