Release of Distributed Collaborative Virtual Wind Tunnel
Steve Bryson, Bryan Green, David Whitney, Sandy Johan, Leslie Keely, Michael Gerald-Yamasaki, Creon Levit |
|
The Distributed, Collaborative Virtual Wind
Tunnel (DCVWT) is an immersive virtual-reality-based
system for use in the investigation of simulated
airflow by geographically distributed, collaborative
teams (figure 1). Participants in the DCVWT interact
with the airflow simulation as if they were in the
same room interacting with a model, even if they are
in fact spread across the country. Simulations ranging
from ocean temperature through commercial aircraft
to the Reusable Launch Vehicle can be examined in
the DCVWT. The DCVWT plays two roles, providing
an immersive environment for collaborative design,
as well as acting as a test bed for developing new
methods. In addition to the Office of Aero-Space
Technology, the results of the DCVWT effort also
benefit the Human Exploration and Development of
Space and the Earth Sciences Enterprises. Programs
like the Intelligent Synthesis Environment Program
use technology and techniques pioneered in the
DCVWT in the development of distributed, collaborative
design environments.
The DCVWT is based on a client-server model,
in which the data to be examined resides on a single
server. Various representations of these data, such as
particle traces in the flow or surfaces of constant
temperature are computed on the server as three-dimensional
graphical objects. Multiple-user clients
receive these graphical objects, which are drawn by
the client from a point of view determined by the
user (figure 2). In this way each user can have his
own view of the same data. Each user can interact
with the data representations, for example, moving a
particle trace source, and all users will see the result.
All users are "peers," each able to interact with the
data, showing things to the other users. Interaction is
performed via a "direct manipulation" paradigm,
where each user can directly "pick up" a data
analysis tool via a three-dimensional interface. Such
a direct manipulation interface requires very high
performance in order to give the sense of the data
analysis tool moving in response to the user's
motions. Such interaction involves information
traveling from the user client to the server, computation
of the new data analysis representation on the
server, sending the new representation to all the
clients, and drawing from all the clients. Because the
complete round trip must occur in about one-tenth of
a second, careful system design and a high-performance
network are required.
The DCVWT presents a new paradigm in collaborative
design. Rather than all participants seeing
the same "master" view controlled by one user, all
participants in the DCVWT have equal status, each
with his own view of the same three-dimensional
data. When one person changes the data representations,
all participants see the result from their own
points of view. This is how the DCVWT creates the
effect of all participants being in the same room even
though they may be actually widely geographically
distributed. The DCVWT supports a variety of
interface hardware, ranging from conventional
workstation and mouse to immersive virtual reality
interfaces, allowing each client to tailor its display to
requirements or budget.
The DCVWT has been demonstrated in cross-country
operation, with four user clients, two in
Washington, D.C., at NASA Headquarters, and two at
Ames. The server for this demonstration was at Ames.
The performance was very satisfactory, with very
good responsiveness, giving a sense of moving
objects in virtual space.
Point of Contact: S. Bryson
(650) 604-4524
bryson@nas.nasa.gov
Back To Top
Previous Paper
Return to Revolutionary Technology
Next Paper |
|
Fig. 1. The Distributed Collaborative Virtual Wind
Tunnel, using the Responsive Workbench virtual
reality display, showing an analysis of simulated
airflow around the Space Shuttle.
|
|
Fig. 2. The Client-Server architecture of Distributed
Collaborative Virtual Wind Tunnel. Simulation data
reside on the server (top), where visualization
geometry is computed and sent to multiple-user
clients (bottom). User commands are sent from each
client, causing new visualization geometry to be
computed by the server, which is then sent to all
clients. Note that the clients can have different types
of display and each sees the data from his own point
of view.
|
|
|
