
Intelligent Systems Report May 1996 Volume 2
No. 5
Computer networks have drastically changed the way businesses
operate. The use of LANs (local area networks), WANs (wide area
networks) and the Internet have enabled companies to easily perform
business transactions and transfer information not only within the
comfortable environs of the workplace itself, but internationally as
well. Historically, these networks have been used for the transfer of
one-dimensional, or at most two-dimensional, information. Now,
virtual reality (VR) has entered the networking fray.
Holding promise for industrial design as well as marketing,
collaborative networked VR permits users at the same or remote sites
to enter an immersive VR environment and manipulate objects in
real-time. The basic components of a networked VR system include a
view of the virtual world for every user and some kind of
representation of the user -- an avatar. The other participants are
also visible to each user.
According to Doug Schiff, vice president of marketing, Division Inc.
(Chapel Hill, N.C.), a provider of VR development tools, virtual
reality networks must be interactive. "That means not only moving
around the world and having interactive images, but being able to
manipulate that world. You've got to be able to create objects,
assign behaviors to them, change those behaviors, delete objects, and
manipulate them. In other words, change the virtual world. This is
critical. Just viewing virtual reality is only halfway there. Being
immersed in a world means being able to participate in that
world."
Speaking at the recent VRWorld conference in Boston, Schiff said,
"Immersion is key to an interactive virtual world. One of the main
reasons for this is the natural movements you get from immersion. You
get a sense of scale and size and location."
Uses of networked VR
The most common use of collaborative, networked VR is concurrent
engineering -- different groups creating different parts of a
mechanical design in different locations and in different parts of
the design process. One user will be building a certain component
while somebody else tests it for installation or maintainability, and
yet another user performs training on that component.
Networked VR can also be used in design reviews. Schiff said, "Right
now, at a large automobile company, they'll have groups all over the
state or city doing frequent design reviews to evaluate different
components of their design -- a piece of an engine, a piece of a car
body, the entire assembly, all different parts of the process.
Without collaborative, networked VR, you have to bring all those
people together in one room or have a video conference. What VR
enables them to do is gather around a virtual prototype and evaluate
the design together in real time."
One of Schiff's first real-world uses of collaborative VR involved
the marketing of an engineering design. "Our customers were doing
engineering design and their customers were located elsewhere. They
were able to give presentations, walk-throughs and demonstrations
from remote sites.
"Also, an important part of marketing is getting early information
about mechanical designs. With collaborative VR, an opportunity
exists to link up sales, marketing and engineering at an earlier
stage in the process."
Yet another proven use for networked VR is training, where experts
are in one location and the trainees are in another.
Implementation
To implement collaborative VR, Division uses dVISE, a VR application
that imports previously established design data, such as CAD, and
uses that data to build interactive virtual worlds. The building of a
collaborative virtual world starts with the use of some kind of CAD
model, which is translated by dVISE into a 2-D Windows-based
interface or a 3-D immersive interface. dVISE information runs over
DVS -- a run-time environment operating on a set of
multiple-independent servers, called actors. Each of these actors
handles a different part of the virtual world. For example, there is
a visual actor, whose purpose is to create images; it doesn't know
anything about collisions or interaction, it only generates images.
There is also a collide actor; all it deals with is one object
meeting another object.
An added benefit in a networked situation is that these different
actors don't need to run on the same processor. In a networked,
collaborative VR setup, these actors could be running on separate
network workstations. There are numerous advantages to this type of
architecture. One of which is that it takes advantage of
multiple-processor systems, which are becoming more commonplace.
"If you know, for example, that you want to do a lot of collision
detection, you could devote a single processor to just worrying about
the collide actor. You could then have all the other actors run on a
separate processor. It gives you the opportunity to customize your
installation to get very high performance," Schiff said.
Division has already participated in a number of networked,
collaborative VR projects, including:
Issues
"Connections are always an issue," Schiff pointed out. Right now the
Internet does not work very well with VR, though recent efforts to
develop a VRML (Virtual Reality Markup Language) standard on the
World Wide Web show a lot of promise (see
ISR,
March 1996). There is not enough bandwidth and connections do not
have the appropriate stability. ISDN will likely be the most common
method of establishing the proper connections, according to Schiff.
Another effective method of establishing such a network is with the
use of WANs.
"Networked VR is a reality," said Schiff. "This is not a research
project; this capability is already standard in our software.
Networked VR will help in design, and in marketing and training. It
will be as common as every other kind of network we have right
now."
Click
here to return to Table of Contents for the ISR May
issue