15.5. ESTABLISHING A REAEARCH AGENDA FOR VIRTUAL REALITIES IN EDUCATION AND TRAINING

Since virtual reality is such a new technology, establishing a research agenda, identifying the important issues for research, is an important first step in exploring its potential. So far, work in virtual reality has focused primarily on refining and improving the technology and developing applications. Many analysts suggest that VR research needs to deal with far more than just technical issues. Laurel (1992) comments, "In the last three years, VR researchers have achieved a quantum leap in the ability to provide sensory immersion. Now it is time to turn our attention to the emotional, cognitive, and aesthetic dimensions of human experience in virtual worlds." Related to this, Thurman (1993) recommends that VR researchers need to focus on instructional strategies, because "device dependency is an immature perspective that almost always gives way to an examination of the effects of training on learners, and thereby finetune how the medium is applied." To date, not much research has been conducted to rigorously test the benefits --- and limitations ---- of learning and training in virtual reality. This is especially true of immersive applications. And assessing the research that has been carried out must take into consideration the rapid changes and improvements in the technology: improved graphics resolution, lighter head-mounted displays, improved processing speed, improved position tracking devices, and increased computer power. So any research concerning the educational benefits of virtual reality must be assessed in the context of rapid technological improvement.

Any research agenda for virtual realities must also take into consideration existing research in related areas that may be relevant. Many analysts (Henderson, 1991; Laurel, 1991; Biocca, 1992a, 1992b; Heeter, 1992; Pausch, Crea, & Conway, 1992; Piantanida, 1993; Piantanida, 1994; Thurman and Mattoon, 1994) have pointed out that there is a strong foundation of research and theory-building in related areas --- human perception, simulation, communications, computer graphics, game design, multimedia, ethology, etc. --- that can be drawn upon in designing and studying VR applications in education and training. Increasingly, research and development in virtual reality is showing an overlap with the field of artificial intelligence (Badler, Barsky, & Zeltzer, 1991; Waldern, 1994, Taubes, 1994a). And Fontaine (1992) has suggested that research concerning the experience of presence in international and intercultural encounters may be valuable for understanding the sense of presence in virtual realities. This example in particular gives a good indication of just how broad the scope of research relevant to virtual realities may be.

Furthermore, research in these foundation areas can be extended as part of a research agenda designed to extend our understanding of the potentials of virtual reality. For example, in terms of research related to perception that is needed to support the development of VR, Moshell and Dunn-Roberts (1993) recommend that theoretical and experimental psychology must provide:

1. Systematic measurement of basic properties
2. Bbetter theories of perception, to guide the formation of hypotheses --- including visual perception, auditory perception, movement and motion sickness, and haptic perception (the sense of force, pressure, etc.)
3. Ccareful tests of hypotheses, which result in increasingly valid theories
4. Constructing and testing of input and output devices based on empirical and theoretical guidelines
5. Evaluation metrics and calibration procedures.

Human factors considerations will need careful attention (Pausch, Crea, & Conway, 1992; Piantanida, 1993; Piantanida, 1994). Waldern (1991) suggests that the following issues are vital considerations in virtual reality research and development: (1) Optical configuration; (2) Engineering construction; (3) Form; (4) User considerations; (5) Wire management; and (6) Safety standards. According to Waldern, the single most difficult aspect is user considerations, which includes anthropometric, ergonomic and health and safety factors. Waldern explains: "If these are wrong, even by a small degree, the design will be a failure because people will choose not to use it." One issue that has come under scrutiny is the safety of head-mounted displays (HMDs), especially with long-term use. This issue will need further study as the technology improves. Wann, Rushton, Mon-Williams, Hawkes, & Smyth (1993) report:

Everyone accepts that increased screen resolution is a requirement for future HMDs, but equally we would suggest that a minimum requirement for the reduction of serious visual stress in stereoscopic presentations is variable focal depth.

Thurman and Mattoon (1994, p.56) comment:

It is our view that VR research and development will provide a foundation for a new and effective form of simulation-based training. However, this can be achieved only if the education and training communities are able to conceptualize the substantial differences (and subsequent improvements) between VR and other simulation strategies. For example, there are indications that VR is already misinterpreted as a single technological innovation associated with head-mounted displays, or sometimes with input devices such as sensor gloves or 3-D trackballs. This is analogous to the mistaken notion that crept into the artificial intelligence (AI) and subsequently the intelligence tutoring system (ITS) community in the not too distant past. That is, in its infant stages, the AI and ITS community mistakenly assumed that certain computer processors (e.g., lisp machines) and languages (e.g., Prolog) constituted artificial intelligence technology. It was not until early implementers were able to get past the "surface features" of the technology and began to look at the "deep structure" of the concept that real inroads and conceptual leaps were made.

This is a very important point for VR researchers to keep in mind.

It will be important to articulate a research agenda specifically relating to virtual reality and education. Fennington and Loge (1992) identify the following issues: (1) How is learning in virtual reality different from that of a traditional educational environment? (2) What do we know about multisensory learning that will be of value in determining the effectiveness of this technology? (3) How are learning styles enhanced or changed by VR? and (4) What kinds of research will be needed to assist instructional designers in developing effective VR learning environments? Related to this, McLellan (1994b) argues that virtual reality can support all seven of the multiple intelligences postulated by Howard Gardner --- linguistic, spatial, logical, musical, kinesthetic, interpersonal and intrapersonal intelligences. VR researchers may want to test this notion.

A detailed research agenda concerning virtual reality as applied to a particular type of training application is provided by a front-end analysis that was conducted by researchers at SRI International (Boman, Piantanida, & Schlager, 1993) to determine the feasibility of using virtual environment technology in Air Force maintenance training. This study was based on interviews with maintenance training and testing experts at Air Force and NASA training sites and at Air Force contractors' sites. Boman, et al (1993) surveyed existing maintenance training and testing practices and technologies, including classroom training, hands-on laboratory training, on-the-job training, software simulations, interactive video, and hardware simulators. This study also examined the training-development process and future maintenance training and testing trends. Boman, et al (1993) determined that virtual environments might offer solutions to several problems that exist in previous training systems. For example, with training in the actual equipment or in some hardware trainers, instructors often cannot see what the student is doing and cannot affect the session in ways that would enhance learning.

The most cited requirements were the need to allow the instructor to view the ongoing training session (from several perspectives) and to interrupt or modify the simulation on the fly (e.g., introducing faults). Other capabilites included instructional guidance and feedback to the student and capture the playback of a session. Such capabilities should be integral features of a VE system. (V. II, pp. 26-27)

Boman, et al (1993) report that the technicians, developers, and instructors interviewed for this study were all in general agreement that if the capabilities outlined above were incorporated in a virtual environment training system, it would have several advantages over current training delivery methods. The most commonly cited advantages were availability, increased safety, and reduced damage to equipment associated with a simulated practice environment. Virtual reality was seen as a way to alleviate the current problem of gaining access to actual equipment and hardware trainers. Self-pacing was also identified as an advantage. For example, instructors could "walk through" a simulated system with all students, allow faster learners to work ahead on their own, and provide remediation to slower students. Boman, et al (1993) report that another potential benefit would be if the system enforced uniformity, helping to solve the problem of maintaining standardization of the maintenance procedures being taught.

Boman, et al (1993) report that some possible impacts of virtual environment simulations include:

1. Pportraying specific aircraft systems
2. Evaluating performance
3. Quick upgrading
4. Avoiding many hardware fabrication costs
5. Disassembling in seconds the computer-generated VR model
6. Configuring the VR model for infrequent or hazardous tasks
7. Incoporating the VR model modifications in electronic form

Their findings indicate that: (1) A need exists for the kind of training virtual reality offers and (2) virtual environment technology has the potential to fill that need. To provide effective VR maintenance training systems, Boman, et al (1993) report that research will be needed in three broad areas: (1) Technology development to produce equipment with the fidelity needed for VR training; (2) Engineering studies to evaluate functional fidelity requirements and develop new methodologies; (3) Training/testing studies to develop an understanding of how best to train using virtual reality training applications. For example, Boman, et al (1993) recommend the development of new methods to use virtual environment devices with simulations, including:

1. Eevaluating methods for navigating within a simulated environment, in particular, comparing the use of speech, gestures, and 3-D/6-D input devices for navigation commands
2. Evaluating methods for manipulating virtual objects including the use of auditory or tactile cues to detect object colision
3. Evaluating virtual menu screens, voice, and hand gesture command modes for steering simulations
4. Evaluating methods for interaction within multiple-participant simulations, including methods to give instructors views from multiple perspectives (e.g., student viewpoint, God's-eye-view, panorama)
5. Having the staff from facilities involved in virtual environment software and courseware development perform the studies on new methodologies.

In sum, virtual environments appear to hold great promise for filling maintenance and other technical training needs, particularly for tasks for which training could not otherwise be adequate because of risks to personnel, prohibitive costs, environmental constraints, or other factors. The utility of virtual environments as more general-pupose maintenance training tools, however, remains unsubstantiated. Boman, et al (1993, Vol,IV,pp.12-16) make a number of recommendations:

• Develop road maps for virtual environment training and testing research
• Identify and/or set up facilities to conduct virtual environment training/testing research
• Conduct experimental studies to establish the effectiveness of VE simulations in facilitating learning at the cognitive process level
• Develop effective principles and methods for training in a virtual environment
• Assess the suitablity of VE simulation for both evaluative and aptitude testing purposes
• Develop criteria for specifying the characteristics of tasks that would benefit from virtual environment training for media selection
• Conduct studies to identify virtual environment training system requirements
• Develop demonstration systems and conduct formative evaluations
• Conduct studies to identify guidelines specifying when and where virtual environment or other technologies are more appropriate in the total curriculum, and how they can be used in concert to maximize training efficiency and optimize the benefits of both
• Develop integrated virtual environment maintenance training system and curriculum prototypes
• Conduct summative evaluation of system performance, usablity, and utility, and of training outcomes

This study gives a good indication of the scope of the research still needed to assess the educational potentials of virtual realities. As this study indicates, a wide gamut of issues will need to be included in any research agenda concerning the educational potentials of VR. Virtual realities appear to hold great promise for education and training, but extensive research and development is still needed to refine and assess the potentials of this emerging technology.