Go to previous article
Go to next article
Go to Table of Contents for 1993 Virtual Reality Conference
Kenneth Nemire, Ph.D.
Adam Burke, Ph.D.
Rick Jacoby, M.S.
1840 - 41st Avenue, Suite 102-315
Capitola, CA 95010
Medical advances, new legislation and changing social attitudes have encouraged more disabled individuals to enter college. Educational systems consequently must continue to respond to the unique needs of this population by developing adaptive educational tools. Some of the more promising and innovative tools in education are virtual environment technologies. Virtual environments may provide uniquely appropriate resources for the disabled. The immersive, intuitive and interactive qualities of the simulated environments give the physically disabled student opportunities for manipulative exploration unavailable in the physical environment; the dexterity, strength or physical access needed to operate in the physical environment are not required. We are examining the utility of this technology for instructional purposes. Laboratory courses requiring hands-on experience, such as physics and biology, are challenging to the physically disabled. We are developing a virtual learning environment for physics instruction that should enhance the student's conceptual as well as manipulative experience. In our presentation we will discuss the issues and technologies employed, including user needs analyses, task analyses and assessment of current technologies. The potential applications of virtual environments for education and training are vast. This is a fertile arena that we expect will provide the next generation of interactive learning systems.
There are 43 million Americans who have one or more physical or mental disabilities (USDJ, 1991). Of that number there are approximately 70,000 injury-related quadriplegics in the United States (Kalsbeek, 1980). Many of these individuals experience limited access to employment, housing, recreation, educational opportunities, and other normal life experiences. Isolated from normal opportunities they often become state dependents costing society billions of dollars a year. In addition, they experience the personal cost of reduced opportunities for a full life and the inability to contribute to the social good.
A positive change in this situation is that many more disabled students are entering college due to changing legislation, social attitudes, and efforts towards accommodation of special needs. This is significant as education can be an important element in the overall rehabilitation and social integration of many physically disabled individuals. Educational systems, however, still do not provide equitable access or opportunities for disabled students to acquire knowledge in the natural and physical sciences. That is socially problematic as a scientifically literate population is critical for our economy to remain globally competitive. If we deprive millions of individuals access to science we will pay a price as a society.
This problem of access to science education is not due, however, to an unwillingness on the part of educational systems to accommodate. They have done a significant job in working to provide access. It is more directly related to an insufficient capacity of the system to modify the physical structures and characteristics of laboratory oriented learning typically used in science education.
Lab oriented sciences are not user-friendly for the disabled. Think of a typical high school physics lab. There are high benches at work stations. The stations have high counters with drawers, sinks, gas jets, glassware, and instruments which require manual dexterity. These stations are fixed in place, with limited space between stations. The teacher is at the front behind his or her counter. Students sit on stools and watch the demonstration and then do the exercise. In addition to limitations for disabled students imposed by the physical design, the classroom activities are also limiting due to the manipulative capabilities and mobility required to perform them. In a physics experiment one might drop or roll objects to test gravitational forces. One might pull weights across the floor to measure friction. One could vibrate rope to observe principles of waves. The entire lab orientation focuses on active participation, requiring a lot of manipulation and mobility.
These laboratory tasks would be difficult for the physically disabled student to be competitive in. Consequently, over the course of the years as science education in school becomes more participatory, the disabled student would become a passive observer rather than a participant. Most of us know what that feels like. Anyone who has ever been one of the last students to be picked for something, a sport, an academic game, to dance, knows what it feels like to be left out, to be just a passive observer.
Because of this lack of meaningful involvement by students with disabilities, a natural disinclination towards the sciences may develop over time. People do not typically persist in tasks in which they cannot develop some degree of efficacy. The passive experience leads to disinterest, avoidance, and eventually, elimination of science courses. Consequently, the disabled student who is motivated to learn in school would focus attention on other cognitive domains in which he or she could be more competitive and successful. This is both a potential loss for the full intellectual development of the student as well as for the society.
Our goal is to address this gap in the education of disabled students by introducing a virtual learning environment to provide easier access and opportunities for involvement and success in the sciences. This is important for a number of reasons. In conversations with professionals who work extensively with physically disabled individuals, a number of key issues were raised consistently. These included the concerns among physically disabled students about: (1) limited mobility; (2) loss of control over the environment; (3) reduced manual capabilities; (4) isolation; and (5) limited access to education and employment. We believe that our virtual learning environment addresses all these issues to some extent. Actually we feel virtual environments are a uniquely appropriate learning technology for the physically disabled. Unlike most interactive media, virtual environments provide an opportunity for experiences which would be essentially unavailable in any other way. It is an interactive medium which provides more than another description of an idea, it actually provides access to a formerly inaccessible domain.
For our prototype learning environment, we have begun to develop a virtual physics lab. In the virtual physics lab, students will be able to perform experiments to test the very same physics principles which would be tested in a real classroom lab. Unlike a real physics class however, all students will be able to perform the exercise and be able to experience mastery within the virtual domain. It would be helpful to consider for a moment the five issues raised above to see how a virtual environment lab could address those concerns.
In considering the issue of mobility we can see how access could be dramatically improved. There are of course human factors concerns regarding the interface of the student with the hardware which would be required to provide the virtual environment experience. Those issues aside for a moment, however, we can imagine how a virtual experience provided in an unobstructed working environment would offer potentially complete access to exercises in a virtual physics lab with unhampered mobility.
Regarding the second issue, control over the environment, the virtual environment provides significant opportunities for the physically disabled student to perform feats which would otherwise be quite impossible, such as lifting a large object. In the virtual environment physics lab, the student could rather effortlessly pull springs, push levers, drop weights and do any number of other tasks quite proficiently. As an educator and psychologist, this is important to me for two reasons. On one level, it would provide the active learning experience of the physics principles in a way all students could successfully engage. But perhaps even more significantly, it could provide an actual experience of mastery over environment. This experience could counter the feelings of helplessness experienced by many physically disabled individuals and the consequent depression and sense of low self-worth which may result.
Manipulation would be enhanced by providing virtual objects with no mass or resistance. A wide variety of objects and materials could be worked with easily. If the interface were sensitive enough, the degree of specificity of manipulation could be very high.
In dealing with isolation, the virtual environment lab provides opportunities for meaningful involvement for all students. For the disabled, it could offer a unique opportunity to really participate with all students. This would be a very helpful experience in contrast to the feeling of 'can't go, can't do'. The self imposed isolation which can occur in these nonsuccess situations can heighten the sense of depression which the student may already feel due to their disability or other life concerns. Another elaboration of this participatory quality of virtual environments is to provide networked experiences so that students from various places could be working coterminously on the same experiment, thus providing cooperative learning experiences with additional benefits to learning. Obviously, the implications are for greater involvement for all students in a shared learning experience. Anyone who is familiar with the reforms currently happening in education knows that this is characteristic of the new paradigm and has been consistently shown to be a potentially effective teaching strategy for both motivation and insight.
Limited Access to Education
Finally, it is obvious how the educational experience of students would be enriched by this approach. Allowing the physically disabled to enter virtual science labs more actively in this way would enrich their understanding and interest in science. They would be active participants, not passive observers. They would have a sense of meaningful involvement. They would have a new capacity to move in and affect environments otherwise inaccessible. All this could potentially enable disabled students to participate in science courses, become even more competitive in a wider job market, and consequently, more socially productive.
One of the challenges in developing a system of this type is addressing the widely diverse user characteristics of this population. Physically disabled students are like everyone else. They come in all ages, sexes, races, personalities and learning styles. Also physical capabilities can vary widely. Consequently one approach would be to incorporate tremendous flexibility into the system to be able to match or interface each unique individual with the system. This flexibility might include single switch controls, widely diverse autonomically controlled devices, and any other fantastic design we could think of. Logically, however, we could imagine that the multiplication of such factors could make the ultimate product more cumbersome and expensive. If we want parsimony, power, and generality we would want to find a common denominator across most individuals that provides a sufficient base for the system interface. One commonalty which came up in a discussion with an active disability advocate and researcher was the wheelchair. All quadriplegics have wheelchairs and can be quite exacting in their capacity to manipulate their them. The chair can thus become a way to navigate in the virtual environment.
This is just one idea of many, an example of the type of human factors considerations which must be seriously examined to maximize a most efficient, effective, and affordable product to broaden the learning domain of physically disabled students.
Disabled students face many challenges in education. To support their efforts to succeed in science and other areas we will design, construct and test a virtual learning environment. This is the first step towards promoting an innovative technology which we anticipate will be the next generation of interactive learning systems. The unique capacity of this methodology to provide access to realistic environments will undoubtedly improve the ability of disabled students to fully participate in the educational system.
Human Factors Design
Our primary goal is to develop a cost-effective, ergonomic virtual learning environment for presenting instruction to students with disabilities. This method also will benefit all other students because our virtual environment system will accommodate both groups of people. Our initial learning environment will provide instruction in physics. Subsequent learning environments will include instruction in biological sciences, chemistry and engineering. Our virtual physics laboratory is different from that being developed by Loftin, Engelberg & Benedetti (1993). Their environment is designed to teach undergraduates difficult-to-learn physics concepts and to be used as a research tool for NASA.
Consequently, the interactive metaphors and tools have developed differently than they would if the virtual learning environment were designed initially for the needs of students with disabilities.
This paper is concerned primarily with the human factors aspects necessary for designing a virtual learning environment. Human factors is the application of knowledge about human physical, perceptual and cognitive capabilities to the design of usable, "user-friendly", tools and systems that accommodate the capabilities and limits of the human user. We will demonstrate a general human factors methodology for designing products for persons with disabilities by presenting an analysis of how we are approaching the specific problem of designing a virtual physics laboratory for students with disabilities. Design of the virtual physics laboratory requires design of the physical and perceptual interface to the virtual environment system. It also requires design of the instructional content, which involves the cognitive and emotional interface with the virtual environment. The scope of this presentation primarily will be limited to the physical and perceptual interface.
In presenting some of the specific needs for students with disabilities, we hope to facilitate a user- and task-driven development of virtual environment technology. The human factors effort must include:
1. User Needs Analysis
Go back to table
User characteristics. Our design analysis begins with describing the people who will use the virtual learning system. The analysis will help direct selection of the hardware and software components. For the prototype system, we will target students with spinal cord injury who are paraplegic or quadriplegic. However, we will attempt to accommodate the needs of students with other motor or orthopedic impairment, like those with cerebral palsy, later in the development cycle. As an aside, a survey of technologies by the U.S. Department of Education targeted virtual reality technology as being particularly suitable for meeting the needs of adolescents with severe physical impairment (Middleton, 1992).
Our analysis will include an assessment of the physical capabilities of our target population. A number of tools are available for quantitatively measuring the residual capabilities of persons with disabilities (Dryden & Kemmerling, 1990; Van der Loos et al 1988). However, each of these measures a limited range of capabilities and are not directly suitable for measuring capabilities that will be useful in the educational or laboratory environment. We will borrow some of these assessment techniques and develop others for our own purposes.
Our virtual learning environment would be designed for these measured capabilities. For example, quadriplegics fatigue easily, and those with cerebral palsy have limited physiological work capabilities (Fernandez, Pitetti & Betzen, 1990). These characteristics would lead us to design a virtual environment that places little physical demand on the user.
For the prototype system, we also targeted a population of students with physical disabilities who have good vision, hearing, intelligence and emotional stability.
As indicated earlier, persons with disabilities have specific needs or concerns for which we must design:
In addition, there are general concerns for which we must design. These are similar to those of most users of a virtual environment system. We will mention only three of these:
2. Task Description and Analysis
Go back to table
Once we have adequately described the user of the system and his or her needs, we must describe the tasks that will be performed by the user. Our human factors assessment includes a task description and analysis to describe the functions of the system and to determine equipment requirements (Drury, Paramore, Van Cott, Grey, & Corlett, 1987; Meister, 1971, 1985). A design task analysis can be used to facilitate the selection of the most appropriate components of the virtual learning system as well as to develop new hardware and software for virtual learning environments.
To choose products to aid a particular individual, one must first know the tasks that the individual wishes to accomplish. We would begin by listing the tools needed in the laboratory and the types of tasks to be performed in the laboratory. Having compiled a taxonomy of tasks, we would ask physics instructors, physics students and other knowledgeable persons to review the list and indicate the importance and frequency with which each task is usually performed.
One needed task would be moving about the laboratory. One must move through the laboratory to an assigned workstation and move from that workstation to other workstations to interact with other students or the instructor, or to use specialized equipment. Mobility and access by those in wheelchairs would require changes in a number of parameters, including aisle clearances and counter accessibility (Cotler & DeGraff, 1976; GSA, 1977). In a virtual environment, these issues of mobility are not of concern as long as the building housing the virtual environment system is accessible. However, navigation within the virtual learning environment can be free and unencumbered. For example, a user typically wears an instrumented glove to interact with the virtual environment system. Various gestures with the gloved hand allows navigation through the environment. The user can view navigation progress by viewing a disembodied hand flying through the virtual environment. This is a pleasure for most people. We can substitute a head movement, manipulation of a joystick or movement of a wheelchair for the gloved hand. Almost any image can represent the movement of the user in the virtual environment.
However, the graphical environment could be designed to literally represent interaction within an ideal laboratory. For example, wide aisles and low, accessible workspaces could be presented within the virtual environment to accommodate wheelchair dimensions. This representation may make the learning experience more compelling for the user by providing virtual interactions with virtual objects resembling those in the physical world. The user could manipulate controls to generate an image of him or herself performing laboratory tasks in or out of a wheelchair. Such an enhanced environment may provide a greater sense of control to the user than one in which disembodied gloves or bodies fly through the air. Further study is required. Other factors also must be considered, such as impact on instructional utility and cost.
Manipulation of objects in a laboratory requires operations such as grasp, translation and rotation of an object in three-dimensional space. These operations would enable a student to perform such tasks as writing notes, and lifting, pushing or examining objects.
A list of tools might include common measuring devices such as a ruler, a protractor and balance scales, other devices such as a calculator and a notebook for recording observations, and objects to manipulate to demonstrate or explore physics concepts and principles.
In the task analysis, each task element is evaluated in terms of:
Task attributes include the conditions requiring performance of the task, the planning required for execution of the task, the tools needed for the task, the operations required to perform the task, and the consequences of the task.
These include the sequence, frequency, duration and importance of the operations necessary to perform the task, and the functional relationships between the tasks and the flexibility of operation.
These include physical demands such as dexterity, reach distances, required force exertion and energy expenditure, as well as perceptual and cognitive demands of the laboratory environment.
We also need to understand other criteria for task performance, such as the necessity for accurate work, timely completion of work and requirements for maintaining good health and safety.
A thorough task analysis must also consider the situations and conditions in which the user will perform the tasks. Environmental interactions are an integral part of the learning environment. Therefore, designers may eventually need to provide a high-fidelity representation of the laboratory learning environment including instructor, students and lab syntax and culture. High simulation fidelity does not mean an exact replication of a generic laboratory. It means creation of the meaningful interactions that help create a learning situation. A simple example comes from experiencing equipment failure. Determining whether a piece of equipment has broken can lead to a more profound learning experience. However, the virtual equipment in a virtual laboratory is not likely to break and thus the student may not learn from these real-life experiences provided by on-the-job training. Does this mean designers should program virtual equipment to break occasionally? We do not know. However, the lessons learned from trouble-shooting could be provided in the instructional program.
Our design must also consider the needs of other students and the instructor when designing the virtual environment system. Eventually these systems will be networked so many students, including non-disabled students, can share the same environment. Regardless of networking, if adoption of the system entails use while others are around, care must be taken to minimize the environmental impact of noise levels and space requirements.
Once we assess the tools, tasks and contexts of a typical physics laboratory, we can determine how to design the virtual learning environment to better meet the needs of the student with disabilities. Many of these attributes can be assessed by referring to extant guidelines (Cotler & DeGraff, 1976; GSA, 1977). Other attributes must be determined for our specific application.
3. Assessment Of Current Technologies
Go back to table
For the purposes of this paper, we will not discuss selection of software for the virtual learning system, although it is an important issue (Loftin et al., 1993). Selection of the appropriate components for a virtual environment system for any purpose requires consideration of the needs of the user, the tasks to be performed and information about the available hardware for virtual environment systems. For our particular purpose in designing the virtual learning environment, the problem is how to select hardware that will meet the needs of a person with a given physical limitation in performing a predetermined set of tasks. To accomplish this goal, we must compare the requirements of the user with the capabilities of the equipment, and select equipment that can be efficiently used by the user. Because of the large number of devices available, it is important to choose the devices in some systematic and meaningful fashion.
Evaluation of the components of the virtual environment learning system must take into account the specific needs of the student with disabilities such as the need for manipulation of objects in the environment and the need for mobility in the environment. Hardware evaluation also must meet the general needs of any user such as creating a compelling simulation of the environment and creating an appropriate metaphor for interaction with the environment. We will first address the particular needs of the student with disabilities.
The virtual learning system will include displays to present visual (Bolas & Fisher, 1990; Fisher, McGreevy, Humphries, & Robinett, 1986; Fisher, Wenzel, Coler, & McGreevy, 1988; Teitel, 1990), auditory (Perrott, 1988; Wenzel, Wightman, & Foster, 1988) and somatosensory (touch, pressure, temperature, force) information, input devices and a computer (Howe, et al. 1990; Ouh-young, Beard, & Brooks, 1989; Patrick et al., 1991; Zerkus et al., 1992). We will also investigate inclusion of a motion platform to provide vestibular (sensations of balance, motion, and orientation with respect to gravity) and other somatosensory stimulation (stimulation from body senses such as sensations of movement and position arising from muscles, tendons and joints).
The function of virtual displays is to provide sufficient and necessary sensory information for the student to perceive the simulated objects. Both the needs for mobility and manipulation require a means for generating the virtual environment in which to explore. For this paper, we will only discuss visual displays. There are several choices. They can be categorized by 1) the physical interface; 2) the level of immersion allowed and 3) the characteristics of the optics.
There are many ways to present a visual virtual environment to a user. The displays are usually mounted on a helmet, head-strap, or goggles worn by the user, or by large projection screens constituting the user's entire visual field (CAVE; Cruz-Neira et al. 1992). The latter is not commercially available and will not be considered further.
Our goal is to choose or develop a device that is light-weight, easy to use, and comfortable for long duration wearing. One of the difficulties with wearing a device on the head is that the quadriplegic would be unable to reposition the device to avoid development of pressure spots and subsequent pain. The need for control over one's environment would require a device that is comfortable to wear or can be manipulated by the student with physical disabilities.
Heavy helmet-mounted displays are probably too cumbersome for this particular population to use. Light-weight goggles may be more appropriate, but as we will note later, they generally suffer from smaller fields-of-view and lack of immersion. A compromise may be boom-mounted displays. These displays can be desk-mounted, alleviating the substantial muscular strain of the neck and back that occurs with long-term use (minutes to hours) of a head-mounted display. These displays can also be coupled to head movements with a simple strap arrangement, still allowing most of the weight to be born by mechanical linkages. A potential difficulty is that manual control of these displays can be quite tiring.
Field-of-view. Perception of the spatial relationships between one's body and the external world depends upon the size of the field-of-view (FOV) of the display (Dixon, Krueger, Rojas, and Hubbard, 1989; Wells, Venturino, and Osgood, 1989). One question for our analyses would be to determine the FOV which would provide the best trade-off between resolution of the display (see below) and successful experience and cost of the display. Display resolution, for example, decreases with increases in FOV. We will need a resolution fine enough to view the instructional materials and a FOV wide enough to provide a compelling, immersive, multisensory environment.
Exclusivity. One of the ways a designer may be able to present a compelling, immersive virtual environment is to provide a display that presents only the virtual environment. Head-mounted displays and the CAVE can do this. Some of the goggles are intended for use with computer monitors. However, these displays have a relatively small field-of-view, the user is able to see the external environment, and one loses view of the virtual environment by turning away from the monitor. This lack of exclusivity may be desirable for some tasks, but we believe that the virtual learning environment would be more compelling if the user only sees the virtual environment. There are no data to support this assumption because studies have not been done.
Characteristics of the Optics.
Developers use a liquid-crystal display (LCD), cathode-ray tube (CRT) or fiber-optic cables to present visual information in virtual environments. The displays may be monocular, binocular (the same graphical image presented to each eye), or stereoscopic (two graphical images presented to represent a three-dimensional image of the scene). It is generally acknowledged that performance using some of these displays may suffer because of such problems as inadequate resolution in the display, limited field-of-view, and optical distortion (Bolas & Fisher, 1990; Crowley, 1991; Nemire, Jacoby, & Ellis, 1994; Pausch, 1991; Rash, Verona, & Crowley, 1990; Robinett & Rolland, 1992; Roscoe, 1985, 1991). However, these limitations may be task- or situation-specific and may adversely impact only some activities (Ellis & Nemire, 1993). One of our tasks will be to choose a display with characteristics sufficient for our purposes.
Stereo, CRT displays provide the kind of visual information needed for our virtual learning device. Research indicates that stereo displays result in better visual and visual-motor performance than monocular or biocular displays (Bryant & Ince, 1991; Kim et al., 1987a; Kim et al., 1987b; Pepper et al., 1981; Pepper et al.,1983). Consequently we will consider only those displays allowing stereo vision.
Currently, a CRT provides greater resolution than an LCD and lower cost than fiber-optic systems of comparable resolution. We will not further evaluate non-exclusive virtual environments because of the constraints mentioned earlier, and we will not consider heavy helmet-mounted displays because of the weight and ease-of-use problems. This leaves goggles and boom-mounted displays for further consideration.
The spatial position of the head can be provided by a mechanical, electromagnetic, optical, or ultrasonic sensing device. A computer continually monitors changes in the position of the user's head and commands the computer graphics program to update the image of the environment to create the perception that the user is moving in and manipulating objects in three-dimensions. Problems with performance in the virtual environment may arise when there are mismatches between change in head position and orientation and the updated visual representation to reflect the change. This results in slow and awkward navigation or manipulation as the user slows his or her movements to accommodate the lag in the virtual environment system. This slow operation can become tedious, awkward and annoying if the user wants to do serious work. The mismatch may also result in more serious problems such as disorientation or nausea.
Unfortunately, all of the current technologies suffer from disadvantages in position tracking. Some combination of these devices, such as combining accelerometers with electromagnetic devices, may lead to better tracking performance for some applications. For our system, we wanted certain tracking characteristics we thought would aid students in using a virtual learning system. These were: measurement of six degrees of freedom, a reasonable working volume, immunity to interference from environmental signals, fast data and update rates, low latency, high accuracy in the estimated head and hand positions, and commercial availability at a reasonable cost.
Optical spatial trackers
After an initial investigation of optical trackers, we decided against these systems for initial development of the virtual learning system because of one of several problems (cf., Meyer, Applewhite & Biocca, 1992). These were either because they are military or research products that are not easily available on the commercial market, or they have limited operational range, excess weight or susceptibility to environmental noise.
Ultrasonic spatial trackers
We also decided not to use ultrasonic spatial trackers because good tracking performance is dependent upon a clear line-of-sight between the sensor and the transmitter. Turning one's head away from the transmitter or moving one's arm between the head and the transmitter would result in gross tracking errors and consequent distortions in the virtual environment.
Mechanical spatial trackers
Mechanical spatial trackers satisfied most of our requirements except for their limited range of operation. Because many of the experiments in the virtual learning environment could be accomplished in the working volume allowed by these trackers, we decided to consider this technology.
Electromagnetic spatial trackers
Electromagnetic spatial trackers are susceptible to interference from nearby metallic objects and from electromagnetic noise generated by CRTs. Additional problems arise from significant transmission lag (Adelstein, Johnston & Ellis, 1992; Meyer et al., 1992).
Initially, we will use a mechanical device for position tracking because of its availability and excellent performance characteristics. However, as we develop our instructional software, we expect to move to an electromagnetic spatial tracker to take advantage of its greater operational range. Further development will also rely on comparisons of the performance of these trackers. Such comparisons would include objective performance measures (e.g., Casali, 1992; Radwin, Vanderheiden, & Lin, 1990) as well as performance rated by more subjective means.
A compelling experience of a three-dimensional environment may require stimulation by cutaneous cues (touch, pressure, temperature and vibration stimulation provided by the skin), proprioception (sensations of movement, position, and muscle tension arising from muscles, tendons and joints), vestibular cues (sensations of balance, motion, and orientation with respect to gravity) and auditory cues (hearing) as well as visual cues. Providing some of this nonvisual information by virtual displays may help the student more effectively interact with the simulated physics laboratory.
Furthermore, the combination of stimulation from different modalities may enhance the experience available to the disabled student. We will determine whether the combination of relatively poor spatial and temporal sensory cues may enhance perception and performance over that obtained with a single source of information (Brunswik, 1956; Massaro, 1987).
Our target population includes people with a wide range of motor function of the upper body, so we have decided to design an interface requiring only control by speech, head, and mouth movements. We have done so because this solution will enable a greater number of people to interact with the environment. However, we will also provide redundant manual controls, like a 3D mouse (e.g., Venolia, 1993) for use by those with the capability to use them. We will provide manual controls because some research has indicated that controlling a mouse cursor by hand rather than head movements results in faster and more accurate performance (Radwin et al., 1990). We want to provide this performance advantage for those with use of their hands.
There are many other ways in which the student could control various functions of our virtual learning environment. The primary concern is to choose controls, controller laws and control interfaces that best match the capabilities of the student and the device to be controlled. Eventually, we will investigate the use of eye controllers (Jacob, 1990; Knapp & Lusted, 1992; Parker & Mercer, 1987; Ware & Mikaelian, 1987; Yamaguchi et al., 1989), other ways of using the head and hand (Brooks & Bejczy, 1985; Das, 1992), as well as lingual (Tongue Touch Keypad), chin or brow switches and sip-and-puff straws (Luttner, 1981), and bioelectric signals such as EMG and EEG (Knapp & Lusted, 1992).
Speech recognition is an important means for allowing disabled persons to interact with a computer (Leifer, 1982; Hammel et al., 1989). A speech recognition system turns the users spoken words into symbols that the computer understands. The computer program can then act upon the understood words in the same way that it acts on any other input from a button, mouse, or menu. One difference between speech recognition systems is whether they recognize isolated words or continuous speech. With a system that recognizes continuous speech the user can speak at a normal rate, while with a system that recognizes distinct words the user must make a slight pause between each word.
Another difference between systems is whether they are speaker dependent or speaker independent. Speaker independent systems typically have a smaller vocabulary but require little or no training. Many speaker independent systems come with a fixed vocabulary. Speaker dependent systems can have a larger vocabulary and usually have a higher recognition rates, but require training by each user.
Ideally a system would have large vocabulary of appropriate words, require little training, recognize continuous speech and isolated words, and have a high recognition rate. We will analyze cost, speaker variability of our target users, impact of training, and vocabulary needs to determine the most cost effective system for a virtual learning environment.
General Concerns for the Design of Virtual Environments
The simulation will be represented as one of several user models or metaphors we present to the student. The user model refers to the representation which the user follows in trying to make sense of the user interface. One common model or metaphor employed in the design of user interfaces is the desktop metaphor in which elements of the interface are organized, presented and controlled like elements of a familiar desktop. These desktop elements might include a workspace, typewriter, file folders, and documents. The closer these elements of the interface are to the elements of a desktop, the easier it will be for the user to learn and use the interface to work productively. Differences between the metaphor and the interface will violate the user's expectations of how the system works and lead to difficulties in learning and using the product. Similarly, considerable evidence indicates that tasks which are more naturally organized by the learner are learned at a faster rate than are those which are less subject to organization (Kieras & Bovair, 1987). In addition, display-control compatibility has been found to affect the early learning of a new skill (Adams, 1954).
Our model is closer to a model of a natural laboratory in which students can "go to" any environment that best satisfies the needs of the investigation. As such, this model is similar to the exploration metaphor developed by McGreevy (1990). Our model must extend the typical laboratory capabilities because of the special needs of our users. Consequently, we may provide interactions that would not work in the physical environment but will work in the virtual environment.
Our task analysis will reveal how to substitute objects and means of interaction in the virtual laboratory for objects and interactions in the physical environment which would be difficult, if not impossible, for the student with a disability to perform.
For example, there is no need for pen and paper if the user has voice control of the system. The user can dictate the notes to the system. However, current voice recognition systems do not have sufficient power to understand continuous speech or to understand a large number of untrained words well (Simpson, et al., 1987; Zimmer, et al., 1991).
Similarly, we can substitute alternative measuring devices for those commonly used in the laboratory. Just as the calculator frees us from making laborious movements of pen on paper, virtual devices can free us from laborious movements in the virtual world. There is no need for a ruler; the user could simply activate a ruler tool and point on two ends of an object to get their distance. Such a virtual tool would eliminate the needs for multiple measuring tools. The same tool would replace, for example, the tape measure and the micrometer.
There is also no need for a balance scale or any other measuring device. The user would only have to apply the appropriate measuring tool and click on the relevant object and the computer could provide the answer. Just as with the calculator, the user has to know what data to provide to get an appropriate answer, or in this case, what operations or experiments to perform. Thus the tools would not replace understanding appropriate concepts, they would only eliminate the need for unnecessary manipulation of the complex physical tools.
Students should perform problem solving of many types rather than lengthy computational exercises. We will use the technology to develop a greater understanding of problem-solving and to provide feedback to the student concerning his or her progress and understanding.
One way to encourage a problem-solving approach would be to present a virtual experience as a physics problem and ask the student to manipulate various parameters to alter the experience. For example, we could provide a virtual bobsled ride complete with visual and auditory sensations of rushing down a slope at 100 miles per hour. Following the experience, we present concepts and mathematics on topics such as force vectors and ask the student to apply them to the bobsled ride. Various forms of exercises, including another ride down the chute, would provide feedback concerning understanding of the material.
The understanding of some topics in physics would be greatly increased by presenting the topic in a virtual environment. Some other topics would not benefit nearly as much from a virtual environment. We would like to identify the best-fit physics topics for a virtual environment's interaction and visualization capability, and then develop lessons for those topics.
Health and Safety
Paraplegics and quadriplegics would have other difficulties in the laboratory. Most laboratories are filled with hazardous materials. Paraplegics and quadriplegics may have limited sensitivity to temperature and touch below the waist. They could be injured by various hazards in the laboratory without knowing it. The virtual environment would prevent these kinds of hazards. However, other hazards can be introduced by the virtual environment system. These include the potential for adverse encounters with obstacles such as system cabling and physical walls because one may not be able to see the external environment while interacting with the virtual environment. These issues must be considered when designing a virtual learning environment.
4. Evaluation Of The Virtual Learning Environment
Go back to table
Principles of good system development require an iterative design and evaluation process in which evaluation of the initial prototype results in a re-design of the system and construction of another prototype (Alavi, 1984; Dieli, 1989; Gould & Lewis, 1985). Iterative testing of the components of the virtual learning environment will help ensure that they have the needed functionality. Our testing will use representative samples of students with disabilities. We will evaluate our system by comparing its performance with existing learning environments such as the traditional physical laboratory and interactive video tools. Evaluation of the virtual learning environment will be in terms of its ability to facilitate instruction in physics and the perception of usability by students. We will also evaluate the components of the virtual environment system in such task operations as grasping a tool or precisely moving a cursor.
The immersive, intuitive and interactive qualities of the virtual environments may provide a valuable means for persons with disabilities to gain greater control over their environment, greater mobility and manipulative capabilities, and greater access to education and employment opportunities. A virtual learning environment for physics instruction should enhance the student's conceptual as well as manipulative experience.
Human factors analyses are required to evaluate the needs and capabilities of the user, analyze the tasks to be performed, assess and select the components of the virtual environment systems, and evaluate the usability of the resultant system for the disabled user. Such analyses also should facilitate further development of virtual environment technologies.
Go to previous article
Go to next article
Go to Table of Contents for 1993 Virtual Reality Conference
Return to the Table of Proceedings