1993 VR Conference Proceedings

Go to previous article 
Go to next article 
Go to Table of Contents for 1993 Virtual Reality Conference 


CAVE Automatic Virtual Environment:
Projection-Based Virtual Environments And Disability

Drew R. Browning
Carolina Cruz-Neira
Daniel J. Sandin
Thomas A. DeFanti
The Electronic Visualization Laboratory (EVL)
University of Illinois at Chicago
UIC School of Art and Design
M/C 036
PO Box 4348
Chicago, IL 60680
E-Mail: U23283@UICVM.UIC.EDU

Abstract

The Electronic Visualization Laboratory at the University of Illinois at Chicago has developed a new virtual reality interface called the CAVE (CAVE Automatic Virtual Environment). It surrounds the viewer with projected images of a virtual environment. Three rear-projection screens make up three walls of a ten-foot cube that all but disappear when illuminated with computer graphics. A fourth data projector illuminates the floor for complete immersion. The viewer can move around the virtual environment and see his own body as he interacts with real and virtual objects.

This paper describes the use of a projection-based virtual reality (VR) interface (the CAVE) for persons with disability. It compares the projection paradigm with the more common VR technologies of head or boom-mounted displays and their associated position sensing techniques. In particular, advantages and disadvantages of the CAVE are discussed in terms of disability issues. These include shared or guided experiences, physical access to the technology, intrusiveness on the user and inclusion of real world objects in the environment. Finally, appropriate disability related applications for projection based virtual environments are considered in light of the inherent properties.


VR Defined

There are many varying definitions of Virtual Reality or degrees of complexity. For purposes of this paper a virtual reality system will be considered to be a three dimensional computer graphics system with real-time interactive control and viewer centered perspective. Additionally it has panoramic binocular display with a large angle of view. These features are essential for a powerful virtual reality experience:

These features generally describe the immersive technologies of head-mounted displays, boom-mounted displays and surround-screen projection-based displays. A Head-Mounted Display (HMD) consists of a pair of miniature displays positioned in goggles or a helmet strapped to the user's head so that each eye sees one display. A boom-mounted display is like a head-mounted goggle display but is suspended from an articulated arm and is held to the viewer's face with handles. The arm acts as a counterbalance and also measures head position. Projection-based displays use rear projection screens to surround the viewer with large scale images. Although other technologies are described as virtual reality systems they do not include one or more of the above features. Such systems would be monitor-based (desk-top VR), wide-screen projection (IMAX, OMNIMAX etc.) and 2D computer graphic virtual interfaces.

A VR system can also involve the other senses: hearing, touch and smell. Three dimensional (localized) sound and tactile feedback can greatly enhance the degree of immersion but are beyond the scope of this paper.


CAVE Background

The CAVE was conceived of in the summer of 1991 by Tom DeFanti and Dan Sandin, co-directors of the Electronic Visualization Lab at UIC. Since then, development of the CAVE has been the primary focus of doctoral student Carolina Cruz-Neira. Sound capabilities have been developed by doctoral student Sumit Das. The CAVE was designed to address some of the shortcomings of other immersive VR systems. These goals include the need for high resolution surround vision without geometric distortion, less sensitivity to head-rotation induced errors, the ability to mix virtual objects and real ones in the same environment, the need for collaborative virtual experiences, and the ability to apply successive refinement to a virtual environment [2]. The CAVE had its public debut SIGGRAPH '92. It was exhibited as part of "Showcase," a juried presentation of work by leading-edge computational scientists. Interactive CAVE visualizations included the anatomy of a seven-week-old human embryo, a virtual snowstorm and a fractal exploratorium [3]. The name is a recursive acronym and also a reference to "The Simile of the Cave" found in Plato's Republic.


CAVE Details

The CAVE is a ten-foot cubical room where three of the walls are rear projection screens and the floor is used as a screen with projection from above. The fourth wall could be a screen with provisions for entering and exiting such as hinging the wall. If the floor and ceiling are both rear projected, the ceiling could be a screen as well. Projectors are high resolution data type projecting color stereographic images on alternating fields. Active liquid crystal display (LCD) stereo shutter type glasses are worn by the viewer to separate the alternate fields to the appropriate eyes. Head position is tracked by an electromagnet sensing device. High-end workstations generating the computer graphics are connected to each projector and an additional workstation is used to synchronize the graphics to each screen. Various input devices have been developed for interacting with virtual environments including a wand and joysticks.


The Surround-Vision Projection Paradigm vs. Head-Mounted Or Boom-Mounted Systems

Although the immersive technologies in comparison here share the essential features of a VR system mentioned above, there are significant inherent differences that can affect applications to people with disabilities. The areas of greatest concern are the user's connection to the real world, quality of the images and the impact these two have on cybersickness [7] (symptoms of visually induced motion sickness in virtual environments: nausea, disorientation, dizziness, etc.) and finally access to the technology.


Connection To The Real World

One of the most significant differences of head-mounted and boom-mounted systems from projection-based VR is the fact that in the former two the user is isolated from the real world when the display is in place.


Inclusion of Real World Objects

With typical head-mounted or boom mounted displays the viewer cannot see his/her own body in a virtual environment unless it is monitored by positioning hardware and then recreated graphically [3]. This is often done with a representation of the hand using a glove input device. Other real world objects must also be modeled in computer graphics to include in the virtual space. Such objects might involve control devices like wheelchair controls, a steering wheel and a dashboard. These virtual objects would be included at the expense of rendering time.

Use of half-silvered mirrors in head-mounted and boom-mounted displays can allow the user to see through the display and have virtual objects superimposed on a real environment [3]. The viewer would be able to see his/her own body but it might have the undesirable effect of the virtual environment superimposed on it. A useful application of this feature would be virtual overlays on real control panels for instructional purposes. For example the sequence of operations in using a wheelchair could be highlighted with animated virtual overlays.

The CAVE also allows for combining real world objects in a virtual environment but with full surround vision and real objects unobstructed by the virtual environment. This last point is an advantage and a disadvantage. Although real objects will never have virtual images obstructing them there are times when it would be appropriate. If for example a virtual object was intended to be in front of a real object or if one's hand passed behind the virtual object then the real object would occlude the virtual one. These situations could be avoided with planning and although virtual overlays would not be possible highlighting can happen with other means such as virtual outlining of a real object.

Shared or Guided Experiences - Another very significant advantage for the CAVE is the ability to have multiple viewers. Although only one viewer would be position-tracked, as many as 12 viewers have routinely experienced the ten-foot CAVE at one time. Additional viewers need only wear the stereo glasses to see exactly what the other people see. For more than one person to share in the virtual reality experience with head-mounted or boom-mounted systems the display hardware must be duplicated for each user [3]. This is very important for collaborative work such as analyzing design models for accessibility or evaluation of a disabled client for reach limitations. For training or therapy situations the client can be guided through a session in a truly shared experience with effective communication, not second hand interpretation. Guided sessions are also important for the potential occurrence of cybersickness and the need for intervention.


Quality Of The Images

Quality of the images presented in a virtual reality system can greatly influence the effectiveness of the experience. Limitations of the specific technologies affect the size of the visible image (angle of view) and the amount of detail in that image (visual acuity).

Angle of View - Head-mount or boom-mount displays have an angle of view of 140 degrees or less. If half-silvered mirrors are used the field of view is restricted to less than 90 degrees [3]. The CAVE has full surround vision for a possible 360 degree angle of view.

Although the stereo glasses will limit the binocular display area, some peripheral vision is still active.

Visual Acuity - Resolution when measured with angle of view is known as visual acuity. The head-mount display with its low-resolution LCD view screens rates poorly. In comparison the CAVE rates fair (same as a desk-top system). The boom-mounted system with its high resolution black & white only displays and narrower field of view rates high. However, the CAVE has low distortion, where as the head-mounted and boom-mounted displays sometimes use optics to increase field of view and resolution at the center of the screens. This can cause distortion in the geometry of the virtual objects [3]. Obviously these visual issues would potentially reduce the effectiveness of the virtual reality experience for users with vision disabilities.

Successive Refinement - With common VR systems real-time interaction requires sacrificing detail in the computer models. By pausing viewer interaction and freezing the image, detail can be added to the models and re-rendered by the computer. In the CAVE the viewer is surrounded by the new detailed panoramic image and free to inspect and analyze it. On command the viewer can then continue to interact or navigate at lower image quality. With head-mounted or boom-mounted systems the refined image could not be panoramic without interaction. Furthermore, freezing the image in a head-mounted display would cause disorientation [2].

First time users of virtual reality devices are often surprised and disappointed by the crudeness of imagery. In fact "at present, over 95% of the population that is 'aware' or 'interested' in the technology has yet to experience it [1]." This fact combined with the fact that the expected incidence of cybersickness in the general population is greater than that of military pilots (20 to 40%) [7] brings into question the likelihood of general diffusion of the medium [1]. It also points to the importance of high quality images and non-intrusive VR experiences.

Cybersickness - Although most research on visually induced motion sickness has been on sickness induced in vehicle simulators (simulator sickness), it is assumed that the problems and findings generalize to other virtual environments [6]. Furthermore " . . . simulator sickness is a subset of the motion sickness experienced from travel through virtual environments, for which we suggest the more general term 'cybersickness.' " [7].

Without a real world reference, getting lost in a virtual world (common with navigation capabilities) can be very disorientating. This can lead to fear in the viewer and potential nausea [2]. In addition when the viewer is isolated from the real world care must be taken to avoid collision with real objects in the environment. The boom-mounted display makes reorientation convenient by allowing the user to let go of the handles and back away from the display.

In a projection based system like the CAVE disorientation and nausea are less of an issue because the user's view is not isolated by the stereo glasses. The viewer can be immersed in a virtual environment but still be conscious of the real world surroundings.

Another possible source of cybersickness is position-tracking errors. Because the user can change position faster than current tracking hardware can report, the virtual world may appear to lag behind the user's movement causing sensory conflict i.e. what is "seen" may be different from what is "felt" [5] [1]. In the CAVE only user displacement position changes cause major tracking errors. Rotational position change causes minor projection errors [2] because the projection plane does not move with the viewer's position and angle (as in head-mounted displays). Lack of image quality or realism can also possibly contribute to simulation sickness [5] [1] due to sensory conflict.

These issues of cybersickness will be very important in applications involving people with disability, particularly those disabilities that affect balance and equilibrium. Mobility training, for example, could take advantage of the ability to present different virtual environments conveniently to a client.

But if that client is presented a situation that causes sensory conflict on top of an already unstable condition then the exercise may be invalid if not harmful.


Access To The Technology

Access to virtual reality interfaces by persons with disabilities is similar to the problems of access to computers in general. Adaptive aids have been developed over the last 13 years that accommodate people of all disabilities. Although many adaptive input devices are keyboard related, the ones that can be used for VR systems would likely be cursor steering or navigation devices like joysticks, trackballs or simple switches that can be placed conveniently where the person could most easily use it. A technique becoming much more common is speech recognition. Another common adaptive input device that would be easily implemented in the CAVE is the Head Pointer. Used by persons with quadriplegia this ultrasonic device is strapped to the user's head for cursor steering. Since a position tracker would normally be mounted to the CAVE user's head, gaze monitoring is automatic.

A popular input device for VR systems is the data glove. This 3D position-tracked device also reads the articulation of each finger. It is used for both navigation, by pointing a finger in the direction of desired travel, and manipulation by grasping virtual objects. In the CAVE a position-tracked wand input device with four buttons is used in a similar way. Navigation is accomplished by pointing in the direction of travel while pushing the appropriate button. Object manipulation is accomplished by pointing at the object and intersecting it with a virtual wand extension. An alternative navigation technique would be to use the thumb operated joystick in the wand or a separate wheelchair style joystick that can be mount to the armrest of a chair.

With 3D position-tracking, gesture recognition techniques can be used for head gesture, hand gesture or wand gesture as an alternative input method. Gestures can be small and subtle or large and obvious. The system could adapt to the needs of the individual. The VR display is of course the primary output device but it could be augmented for persons with vision disabilities by localized (3D) speech output, localized (3D) sound and tactile feedback.

Access to Head-mounted display technology requires wearing hardware weighing 2.5 pounds or more. This can be a serious problem for the person with low muscle tone disability, particularly since this technology is the predominant display of VR systems. These systems increase the effective weight of the head by 20% or more [4]. Not only is it an access issue but also a cybersickness issue as well, altering static and dynamic balance when the effective center of mass is shifted.

Boom-mounted displays are counterbalanced so weight is not an issue. However the user must hold the display while interacting or maneuvering a wheelchair, a difficult if not impossible task for some people with severe disabilities. Maneuvering a wheelchair while wearing a head-mounted-display can be restricted by the tether that connects the display to the computer. Care must be taken to not run over or become tangled in the cables.

The CAVE is not wearable technology but rather an room-like construction with a projector associated with each of three walls and the floor. A ten-foot cubic CAVE occupies approximately 600 square feet of floor space and needs thirteen feet of ceiling height. For this reason the CAVE is not very portable. Entering it is like entering a room that is raised 6" off the floor (for construction reasons) and ramped for wheelchair access. So far the CAVE has been built in two sizes, a seven-foot cube and ten-foot cube. Since collision with the walls can cause damage, the person using a wheelchair must be conscious of his/her position. The ten-foot cube feels much bigger and is more comfortable to maneuver in. The stereo glasses have the head-tracking sensor mounted to them and a two-cable tether going off to an upper corner of the CAVE. The viewer using a wheelchair must be conscious of this so as not to run it over and get tangled in it. Another problem for wheelchair users is the fact that the tracking hardware uses magnetic fields to work. Large bodies of metal and electrical signals such as those found on an electric wheelchair cause interference and non-linearities in the data. For these reasons gross maneuvering is better accomplished through virtual navigation.


Cost

The advantages of the CAVE come at a price. Although number of screens implemented can be varied, the more the better. Here is a list of major hardware:


Applications

Although scientific visualization has held the spotlight as efforts focused on debuting the CAVE at SIGGRAPH '92, early in the development of the CAVE, applications for people with disability and accessible design were planned for. In collaboration with the Design Visualizationn Laboratory, research into using virtual environments to study the human factors of 3D design was conducted under a grant from the Manufacturing Research Center at the University of Illinois [8]. Current development in applications using the CAVE for persons with disability is in the areas of design and assistive technology.

In developing design applications it was apparent that a convenient way to bring Computer-Aided Design (CAD) models into the CAVE was important. Many models already existed or would be designed on 3D CAD systems for use in virtual environments. Since most applications for the CAVE were scientific visualization or algorithmically generated, there was no standardized data file format in place. The Drawing Interchange Format (DXF) developed by Autodesk, Inc. was chosen because it is widely used, flexible, portable and easy to interpret. Graduate student Lewis Siegel wrote DXFLib.a, a Library written for the Silicon Graphics programming library (GL), which when linked to a C program allows the programmer to easily load, draw and save DXF files.

To allow people using wheelchairs to navigate virtual environments a wheelchair joystick interface was built to simulate an electric wheelchair. It can be attached to an existing wheelchair, manual or electric, or to any chair. Although a "treadmill" style interface would be most appropriate for manual wheelchairs [9], a joystick approach was convenient and more universal. Both the wand and joystick are interfaced to the CAVE through a PC clone. The PC also serves as the sound interface.

An ongoing application is the design of accessible vehicle interiors. Models of the University campus shuttle bus and public transit vehicles have been built in our CAD system and converted to DXF format. By bringing these models into the CAVE we can "try them out" using a wheelchair. The wheelchair lift area can be inspected for clearance and the handrail checked for position and height. We can check the fare-box reach distance and turning clearances for entering the isle. By maneuvering into the tie-down area we can see if the securement device is within reach. Although not implement yet we will be able to make changes while in the CAVE. If the farebox if too far, we can point at it with the wand and bring it closer. If a stanchion is in the way we can move it to a new location.

In the area of assistive technology applications are being developed for evaluating somatic experiences on cognitive development in severely disabled children. Virtual reality experienced in the CAVE can allow children to explore, create, play and learn in virtual environments and virtual worlds. It is an objective reality that is malleable. They can visit places that would otherwise be impossible, impractical or too dangerous. These three dimensional panoramic worlds can be customized to the individual and their needs. The computer can monitor the interaction of a child in the environment and make changes as needed.


Conclusion

The CAVE is a virtual reality interface that differs from other immersive VR technologies in significant ways. Advantages of the surround-screen projection-based VR paradigm that most affect applications for persons with disability are the inclusion of real-world objects in virtual environments, shared or guided VR experiences, improved image quality, reduced risk of cybersickness and better physical access. Improvements need to be made with electric wheelchair interference on the position-tracker and the wires associated with it. Ideally it would be a wireless nonmagnetic technology. The effectiveness of the CAVE has been proven for scientific visualization and now we are extending its use into new applications. We must make sure that this new technology, which undoubtedly will have a profound impact on all of our lives, does not exclude people with disabilities.


Acknowledgments

The CAVE is a research activity of the Electronic Visualization Laboratory of the University of Illinois at Chicago.

Thomas A. DeFanti, Co-Director; Daniel J. Sandin, Co-Director.

Support from Argonne National Laboratory and the National Center for Supercomputing Applications at the University of Illinois at Urbana-Champaign. Equipment support from Ascension Technology Company, DataDisplay Corporation, Electrohome Projection Systems, Polhemus, Silicon Graphics Computer Systems, Stereographics Corporation, and Systran Corporation. Major funding is provided by the National Science Foundation (NSF) grant ASC-92113813, which includes support from the Advanced Research Projects Agency and the National Institute for Mental Health, NSF Grant IRI-9213822, and the Illinois Technology Challenge Grant.


References

Go to previous article 
Go to next article 
Go to Table of Contents for 1993 Virtual Reality Conference 
Return to the Table of Proceedings 


Reprinted with author(s) permission. Author(s) retain copyright.