This paper introduces a collaboration setting where an expert in VR gives instructions to a technician using MR being in front of the real object. For this study, we develop a system in which users communicate while building a specific structure using Lego Duplo blocks. Here, only the expert knows the final configuration and instructs the technician equipped with an MR headset about the assembly. Four communication methods are evaluated: voice chat, avatars, the possibility of moving virtual blocks similar to the real ones, and an arrow functionality. This work proves the usefulness of such a VR/MR instruction system and provides insights into the advantages of each proposed communication method.

Redirected Walking (RDW) is a common technique leveraged to allow real walking for exploring large virtual environments in constrained physical tracking spaces. Effective RDW is challenging due to its complexity and disturbance factors (e.g., spontaneous user behavior). Existing techniques range from combinations of simple motion scaling to more elaborate curvature injections and reactive, predictive, or scripted steering concepts. However, many of these approaches were evaluated in simulation only, and researchers argued that the findings would translate to real scenarios to motivate the effectiveness of their algorithms. Using the Redirected Walking Toolkit and its virtual path generator, a randomized waypoint-based path generator has been common practice, although its built-in simplifications assume sequential user behavior regarding translation and rotation.In this paper, we argue that pure simulation-based evaluations employing such simplified path generators require critical reflection. We demonstrate RDW simulations that show the chaotic process fundamental to RDW, in which altering the initial user’s position by mere millimeters can drastically change the resulting steering behavior. This insight suggests that RDW is more sensitive to the underlying data than previously assumed. Thus, we rigorously analyze the influence of commonly used synthetically generated paths on multiple state-of-the-art steering concepts and compare them against previously recorded real paths.

The main goal of Virtual Training Environments (VTEs) is to maximize training effectiveness, which can be achieved by increasing the degree of immersion. While prior work mainly focused on the visual, auditory, and navigational aspects of immersion, proprioceptive aspects may be particularly important. In this work-in-progress paper, we explore this potential by implementing an industrial VTE, which can be interacted with, using VR gloves and a tracked real tool. Further, we evaluate the VTE in a pilot study with industry apprentices, providing initial evidence that participants experienced high presence and low task load, while being generally satisfied with the training.

One of the main characteristics of virtual reality (VR) is immersion, which leads to comprehensive illusions of reality. Accordingly, VR is used in many applications like entertainment, marketing, and training. Especially in training applications, the effect of immersion on training success is still not entirely clear, since too much immersion may cause side effects such as users experiencing high mental demand whereas too little may disturb users’ well-being. To further investigate the matter, we developed two virtual training environments, wherein users train a typical industrial assembly task either in low or high immersive VR. In a controlled pilot study, we additionally introduced a third condition, the control group, which justifies the necessity of the training. Immediately after the VR training session, each participant completed the corresponding real assembly task in which their performance was measured. Preliminary results from our pilot study show that participants trained in high immersive VR performed better, while negative side effects could not be detected.

In predictive redirected walking applications, planning the redirection and path prediction are crucial for a safe and effective redirection. Common predictive redirection algorithms require many simplifications and limitations concerning the real and the virtual environments to achieve a real-time performance. These limitations include for example that the tracking space needs to be convex, and only a single user is supported. In this paper, we present a novel approach called PReWAP which addresses many of these shortcomings. We introduce artificial potential fields to represent the real environment which are able to handle non-convex environments and multiple co-located users. Further, we show how this new approach can be integrated into a model predictive controller which will allow various redirection techniques and multiple gains to be applied.

In predictive redirected walking applications, reliable path prediction is essential for an effective redirection. So far, most predictive redirected walking algorithms introduced many restrictions to the virtual environment in order to simplify this path prediction. Path prediction is time-consuming and is also prone to errors due to potentially impulsive and even irrational walking behaviour of humans in virtual environments. Therefore, many applications confine users in narrow virtual corridors or mazes in order to minimise larger deviations from intended and predictable walking patterns. In this paper, we present a novel approach for short-term path prediction which can be applied to virtual open space predictive redirected walking. We introduce a drop-shaped trajectory prediction which is described using a Lemniscate of Bernoulli. The drop's contour is discretised and we show how this is used to determine potential user trajectories in the virtual environment.

For the contest held at IEEE VR 2018 3DUIs for three tasks have to be designed. The tasks include climbing a ladder, fly a drone in first-person perspective and stacking objects. The goal of our design is to find an easy yet intuitive solution to the given tasks. In that way, for the ladder task, a 3DUI is implemented closely mimicking a real ladder ascent or descent using only the controllers to interact with the environment. For the drone flying, we propose a solution which uses the head orientation to control the drone. The third task concluded with a no gravity, temporary storage space which is used to prepare the final stack that can be released with a trigger press.

Redirected walking applications allow a user to explore large virtual environments in a smaller physical space by employing so-called redirection techniques. To further improve the immersion of a virtual experience, path planner algorithms were developed which choose redirection techniques based on the current position and orientation of the user. In order to ensure a reliable performance, planning algorithms depend on accurate position tracking using an external tracking system. However, the disadvantage of such a tracking method is the time-consuming preparation of the physical environment which renders the system immobile. A possible solution to eliminate this dependency is to replace the external tracking system with a state-of-the-art inside-out tracker based on the concept of Simultaneous Localization and Mapping (SLAM). In this paper, we present an approach in which we attach a commercially available SLAM device to a head-mounted display to track the head motion of a user. From sensor recordings of the device, we construct a map of the surrounding environment for future processing in an existing path planner for redirected walking.

Existing redirected walking applications use accurate tracking systems to determine the position and orientation of the user within a designated tracking space. In order to plan redirection and to ensure the user's safety, it is necessary to define the walking area in advance. However, when using ad hoc redirected walking, this is not possible, because the user's surroundings are not known beforehand.