A Network of four Telecubicles

 

An initial network of at least four Tele-Immersion sites will be put in

place. Each of these will have a "telecubicle" - a dramatic new interface

design. A telecubicle is an office that can appear to become one quadrant

in a larger shared virtual office space. The initial telecubicle sites

will be at UIC, UNC, and USC, as well as one in the New York Area. This

configuration meets the geographical constraints set earlier in the

initiative, and includes key participants from existing tele-collaboration

efforts.

 

A telecubicle has a stereo-immersive desk surface as well as at least two

stereo-immersive walls. These three display surfaces meet, in the

formation of a desk against a corner. When a telecubicle is linked to

others on the net, the walls appear to be transparent passages to the other

cubicles. Four telecubicles can be joined in virtual space so that each

forms a quadrant of a larger virtual whole. The desk surfaces line up to

form a large table in the middle of the virtual shared room. Through the

walls can be seen the other participants as well as props, furniture, and

other physical features of their environments. Virtual objects and data

can be passed through the walls between participants, and placed on the

shared table in the middle for viewing.

 

This simple idea requires dramatic advances in technology. In order to

allow users to look into each other's spaces, head-tracked views of those

spaces must be generated. (In a head-tracked view, as opposed to a fixed

point-of-view, a scene can be viewed from a dynamically changing

point-of-view to keep up with a user as she moves her head.) This requires

sampling and resynthesis of the physical environment, as well as

the users' faces and bodies. This is a new challenge that will push a

range of emerging technologies, such as scene depth extraction and warp

rendering, to the next level.

 

There are other less obvious challenges that will also produce important

advances. For instance, the participants must see in stereo, so they will

be wearing glasses or goggles. But they must also see each other's faces

in full, including the eyes, because the complete range of communicative

cues between people, including eye movement, can be vital to collaboration.

This means when a user's face is resynthesized (which is required to

provide a head-tracked point of view), the stereo viewing equipment that

user is wearing must be made invisible.

 

In addition to pushing the envelopes of computer vision, tracking, display,

and rendering technologies, telecubicles also advance the cause of

convenience and packaging for real-world acceptance. A telecubicle will

not require more floor space than a regular cubicle. One does not have to

give up old work habits to use a telecubicle. Users will eventually be

able to bring assorted tools such as written notes on paper, physical

models, and even old-fashioned computers into a telecubicle, and

automatically share them across the network. There is no formal beginning

and ending to a telecubicle session. One simply sits down and starts to

talk and work with collaborators who are present at their own sites.

 

It is important to note that the idea and design of the telecubicle came

directly out of the Tele-Immersion meeting that took place at Advanced

Network's offices on July 21, 1997. One of the striking results of that

meeting was the discovery of an unsuspected convergence of visions of what

the next generation of immersive interface will look like. UIC, NPS, UNC,

Columbia, and USC all independently brought drawings of a cubicle with a

desk in which the user and desk were together immersed, and through the

walls of which could be seen other users at their desks. This unforeseen

shared vision, then, has served to guide the first year's plan for the

Tele-Immersion initiative.

 

Some Limitations During The First Year:

During the first year, each site will not have an equal ability to present itself to the others. For instance, UNC will take the lead on environmental sampling, meaning that other sites will be able to look into the UNC telecubicle and see an ever better representation of what is actually present in the physical cubicle. However, when a user looks out to the other telecubicles from UNC, she will probably see a less developed representation of the other physical environments.

During the first year, the various software modules, such as Sketch, Body Electric, and Alice, will be made interoperable, but the integration will be far from seamless. A first year telecubicle will have one or two conventional workstation monitors present to allow the use of these programs. By the third year, all the tools will be available either through conventional windows interfaces, or within the telecubicle, as integrated tools made out of virtual objects.

During the first year, only telecubicles will be considered. It is important to remember that the three-year goal of the initiative is significantly broader. Heterogeneous interfaces will be introduced starting in the second year, so that users in telecubicles, head-mounted displays, and outdoor ambulatory augmented reality displays can collaborate.