22-11-2012, 06:14 PM
A Tele-Immersive Environment for Collaborative Exploratory Analysis of Massive Data Sets
A Tele-Immersive.pdf (Size: 270.49 KB / Downloads: 16)
Abstract
This is a white paper outlining a methodology for
employing collaborative, immersive virtual
environments as a high-end visualization interface for
massive data-sets.
Introduction
In 1997 a series of National Science Foundation
(NSF) and Department of Energy (DOE) sponsored
workshops brought together computer scientists
specializing in high-performance computing and
scientific visualization, and domain scientists in
physics, chemistry, materials science, and
engineering. Their goal was to assess the needs of the
scientific and engineering community; to identify
current and projected computational capabilities; and
to outline a federal research and development agenda
in scientific visualization, human interface, and the
manipulation of massive scientific data-sets[1].
The findings of the workshops indicated a clear
trend- that the amount of data collected and generated
through scientific simulations was growing
dramatically and that currently available technologies
for interpreting this data are becoming increasingly
inadequate. It was estimated that in 1999 a typical
data query will access between 3-30 tera-bytes of
data. This is expected to increase to 30-300 tera-bytes
in 2001, and 1 peta-byte in 2004. Progress in data
mining can help significantly in finding the gems of
information buried in the data, however scientists are
currently still unable to articulate sufficiently smart
algorithms that can reliably find relevant features or
draw correct and relevant conclusions on their own.
TIDE : a Tele-Immersive Data
Exploration environment
Envision a scenario in which three users- one in a
CAVE, another on an ImmersaDesk and yet another
on a desktop workstation are all engaged in a typical
data exploration exercise within a virtual laboratory.
The CAVE virtual reality system is a 10 foot-cubed
room that is projected with stereoscopic images
creating the illusion that objects appear to co-exist
with the user in the room. The ImmersaDesk is a
smaller, drafting-table-like system also capable of
projecting stereoscopic images. All users are
separated by hundreds of miles but appear colocated-
able to see each other as either a video image
or as a simplified virtual representation (commonly
known as an avatar). Each avatar has arms and hands
so that they may convey natural gesture such as
pointing at areas of interest in the visualization.
Digital audio is streamed between the sites to allow
them to speak to each other. The desktop workstation
displays a data-flow model that can be used to
construct the visualization that is shared between all
three display devices. The participants in the VR
displays can use three-dimensional tools to directly
manipulate the visualization- for example in the
CAVE a user is changing the isosurface value in the
data-set. These changes are automatically propagated
to all the other visualization displays. In the
meantime the ImmersaDesk user, noticing an
anomaly in the data-set, inserts an annotation in the
data-set as a reminder to return to more closely
examine the region.
The Tele-Immersion Server
The Tele-Immersion Server’s primary responsibility
is to create a persistent entry point for the TICs. That
is, when a client is connected to the TIS, a user can
work synchronously or asynchronously with other
users. The environment will persist even when all
participants have left it. The server also maintains the
consistent state that is shared across all participating
TICs. Finally the TIS stores the data subsets that are
extracted from the external data sources. The data
subsets may consist of raw and derived data sets,
three dimensional models or images.
The Tele-Immersion Client
The Tele-Immersion Client (TIC) consists of the VR
display device (either CAVE, ImmersaDesk, etc) and
the software tools necessary to allow “human-in-theloop”
computational steering, retrieval, visualization,
and annotation of the data. The TIC also provides the
basic capabilities for streaming audio and video, and
for rendering avatars to allow participants to
communicate effectively with one another while they
are immersed in the environment. These capabilities
come as part of EVL’s Tele-Immersion software
framework called CAVERNsoft (Figure 2) [4].
CAVERNsoft has already been used in over a dozen
collaborative applications ranging from art and
design, engineering and science, and education and
training; and by clients such as General Motors,
Searle/Monsanto, Nippon Telephone and Telegraph,
the Center for Coastal and Physical Oceanography,
and the Naval Research Lab [5]. We believe it is well
suited as the base architecture for the TIC.
Closing Remarks
This paper has outlined a methodology for using teleimmersive
systems as a high-end visualization
interface for exploring massive data-sets. TIDE’s
research foci seeks to develop: new human-factors
techniques and technologies for multi-dimensional
visualization; new technologies for sustaining longterm
collaborative data exploration; and new
techniques for the interactive exploration of massive
data-sets.
We are only in the architectural design phases of
TIDE. However as TIDE's underlying architecture
will be based on CAVERNsoft much of the
architecture for supporting Tele-Immersion is already
in place. We anticipate that a functioning proof-ofconcept
will be built by the end of 1999 in which it
will already be useful for visualization a variety of
data-sets collaboratively.