Human centred design of 3-D interaction devices to control virtual environments

https://doi.org/10.1016/j.ijhcs.2005.08.010Get rights and content

Abstract

It is commonly acknowledged that user needs should drive design, but often technical influences prevail. Currently, there are no standard interaction devices or interfaces used in 3-D environments, and there is a lack of specific best practice guidelines to develop these. This paper discusses the process of collecting feedback on prototype designs for VR/VE interaction devices from both expert users and non-expert users, and demonstrates how the information gained from human centred evaluation can be used to further the design process. Experiment 1 examined the usability of two magnetically tracked interaction devices with three different types of menus (Sphere, Linear and Fan). Quantitative and qualitative analysis was carried out on the results, and usability problems with the menus and devices are discussed. The findings from this experiment were translated into general design guidance, in addition to specific recommendations. A new device was designed on the basis of some of these recommendations and its usability was evaluated in Experiment 2. Feedback from participants in Experiment 2 demonstrated that the design recommendations emerging from Experiment 1 were successfully applied to develop a more usable and acceptable device.

Introduction

One of the major problems when designing and developing new interaction techniques for 3-D environments is the lack of specific best practice guidelines. Reported approaches to 3-D environment design (e.g. Kaur, 1998; Gabbard et al., 1999a; Bowman et al., 2001a, Bowman et al., 2001b; Eastgate, 2001; Wilson et al., 2002) are difficult to translate into specific guidance for use by developers. In this paper, we discuss the design of menus and interaction devices used within immersive environments and how the results from evaluating such designs can be used to begin the development of guidelines.

The Hybrid 6-sided Personal Immersion System (HyPI-6) is a 6-walled 3×3×3 m cubic stereo projection space in which computer-generated images can be projected externally onto the walls. The term ‘hybrid’ refers to the fact that the system can be driven in active (in conjunction with an SGI Onyx) as well as passive (in conjunction with a PC cluster) stereo projection mode. A user standing within the cube feels integrated within a computer-simulated world. Such a projection environment is achieved by a combination of stereoscopic real-time representation, suitable interaction tools with which participants can steer the three-dimensional space around themselves, and special software. Systems which generate virtual worlds have been in use for several years although the number of systems actually installed and in working use is relatively small.

In an average office working environment most employees will be using a standard keyboard and mouse to interact with certain conventional computer hardware and software. However, there is no typical or average VR technology set-up and it is possible that for each facility using a projection display via CAVE technology, a different set of interaction devices are used (a CAVE is a spatially immersive display in the form of a cubic projection room consisting of 3–6 projection screens). Over the past decade, a number of innovative interaction devices have been proposed, but they are still far from becoming standard tools (Zhai, 1998) and selection may have more to do with cost and availability than usability and comfort.

In addition to hardware considerations, there are no standard interface metaphors used in 3-D environments. All too commonly, designers create a menu structure/device which they personally find usable, or worse, attractive, with little concern for the actual usability of these interfaces/devices and the demands which they will place on the non-specialist user. Problems discovered later in real use could easily have been avoided if usability had been a more realistic concern earlier on in the design process. As was seen in early PC technologies, the emphasis in VR/VE design is still much more on overcoming technical limitations and showcasing new capabilities, rather than on user led design. In addition, new interfaces are often designed for new models or applications, resulting in a lack of consistency and restricting the ability of the user to develop common understanding of interaction and navigation methods. As the potential for active use of visualization technologies in industrial applications increases, the need to develop interaction devices and interfaces which are intuitive and easy to use grows.

New, improved interaction devices can be designed through an understanding of usability problems associated with existing devices. The experiments presented in this paper were carried out within the European IST project, VIEW of the Future, and had the aim of guiding the development of new interaction devices. A set of prototype devices with different grip requirements was designed in order to investigate preferences. Non-specialist users cannot be expected to understand some of the technical constraints limiting the design of the prototypes, so we combined input from experts and non-expert users. Experiment 1 was conducted to evaluate the use of two devices with different interfaces. Design recommendations were made on the basis of this study and a new device was developed and evaluated in Experiment 2.

Section snippets

Interaction in virtual environments

Whilst there have been a number of studies considering interaction within immersive environments (e.g. Barfield et al., 1998; Ruddle et al., 1998; Bowman et al., 2001a, Bowman et al., 2001b), these have not been aimed at producing design guidance. Such guidance may be available from research carried out within the HCI domain, but the differences between 2-D and 3-D interaction mean that this guidance is not usually directly transferable to virtual environments (VEs). Some researchers have

Development of new devices and concepts in VIEW of the Future

The success of the end product in human centred design depends on an iterative design and evaluation process. It is important for the design process to allow flexibility and facilitate creativity whilst also taking into account guidance of technical, task and user constraints. Fig. 1 illustrates the stages of the iterative design process employed within the VIEW project.

However, there are sometimes difficulties with actually achieving human centred design. Firstly, it may be difficult to get

Expert testing of new prototype devices

Although often an afterthought in HCI research and computer interface design, perhaps being seen as trivial compared to the cognitive issues, physical ergonomics of devices is in fact critical for successful interaction in VR/VE. Anthropometric measurements of the human hand are essential for the design of handheld devices (Bandera et al., 1986), and there are numerous ergonomic grips which can be afforded by hand-held devices by the use of different types of handles, all of which have their

VR system

A five-sided Hybrid Personal Immersion System was linked to an SGI Onyx running on Lightning. CRT projection with a resolution of 1024×1024 and active stereo (57 Hz) with magnetic tracking (wireless motion star) and StereoGraphic shutter glasses were used to display the VEs.

Experimental virtual environments

Three VEs were used—two modeling applications (Linear menu and Fan menu) and one visualization application (Spherical menu). In the two modelling VEs participants were required to complete a number of tasks, creating and

Design recommendations

Participants in this experiment experienced a number of usability problems. Table 4 translates some of the specific findings from this experiment into more general design guidance. We acknowledge that these recommendations are based on the data from a small sample of participants, however, this guidance can be tested in the future to examine its usefulness.

Design of the Hornet

A number of ergonomic and other design issues emerged from the qualitative analysis of Experiment 1, which enabled designers to highlight the problems they would address for a new generation of interaction devices. They aimed to improve upon existing devices by incorporating the good design features of both the Mike and the OC in the new device, whilst addressing the negative features and issues emerging from Experiment 1. The basic shape was a combination of the OC and the Mike, and

Discussion

Ideally, in order to build a working knowledge of best practice design of VE systems and devices, each design decision and the reasoning behind this decision should be documented. Even when VEs have been evaluated for usability, comparison across different studies is very difficult due to the range of evaluation techniques which are in use and the differences in requirements of VR applications. A standardized tool to measure usability would be valuable. A combination of expert usability testing

Conclusions

Interaction devices designed for use in VEs need to perform technically as well as be ergonomic and easy to use in order to be accepted by end-users. This paper has demonstrated a practical example of this approach. User acceptance of the device is important in its perceived usability, and for effective and efficient operation within a VE. Individual user preference for usability and aesthetics of a device must be taken into account, but guidelines could be provided in the future to suggest

Acknowledgements

The authors would like to thank Eleanor Marshall and Anastasis Tzoumpas for their help in conducting the experiments. This work has been funded by the European Commission, IST Project VIEW of the Future, IST-2000-26089.

References (36)

  • D. Bowman et al.

    Design and evaluation of menu systems for immersive virtual environments

    Proceedings of IEEE Virtual Reality

    (2001)
  • D.A. Bowman et al.

    The virtual venue: user–computer interaction in information-rich virtual environments

    Presence: Teleoperators and Virtual Environments

    (1998)
  • D.A. Bowman et al.

    Testbed evaluation of virtual environment interaction techniques

    Presence: Teleoperators and Virtual Environments

    (2001)
  • D. Bowman et al.

    An introduction to 3-D user interface design

    Presence: Teleoperators and Virtual Environments

    (2001)
  • Bullinger, H.-J., 1994. Ergonomie, Produkt und Arbeitsplatzgestaltung, Teubner,...
  • W. Buxton

    There's more to interaction than meets the eye: some issues in manual input

  • S.K. Card et al.

    A morphological analysis of the design space of input devices

    ACM Transactions on Information Systems

    (1991)
  • Darken, R., 1994. Hands-off interaction with menus in virtual spaces. In: Proceedings of SPIE 1994, Stereoscopic...
  • Cited by (25)

    • The effects of egocentric and allocentric representations on presence and perceived realism: Tested in stereoscopic 3D games

      2012, Interacting with Computers
      Citation Excerpt :

      According to Steuer (1992), interactivity is defined as ‘the extent to which users can participate in modifying the form and content of a mediated environment in real time’. Patel et al. (2006) claims, there is a lack of research on what types of interactions enhance the experience in 3D display technology. Most studies on manipulation in stereoscopic 3D are related to visual feedback of user movement and execution.

    • Human Factors of Virtual Reality - Where are We Now?

      2007, Meeting Diversity in Ergonomics
    • Extending the desktop workplace by a portable virtual reality system

      2006, International Journal of Human Computer Studies
    • VEA: A virtual environment for animal experimentation

      2021, Proceedings - IEEE 21st International Conference on Advanced Learning Technologies, ICALT 2021
    View all citing articles on Scopus
    View full text