International Technology Alliance
This is a project
funded by the IBM led ITA Consortium
IBM Research is leading the International Technology Alliance,
a consortium of 24 institutions from industry and academia in both
the U.K. and the U.S. created in May 2006. The International Technology
Alliance opens a new era of collaborative, multi-disciplinary research
which spans multiple universities and industrial research labs.
A number of leading researchers have come together to perform joint
research, forming a virtual organization of some of the best minds
in both countries over the next ten years. Among them, Dr. Boleslaw
Szymanski of RPI is leading Project 9, one of the 12 projects of the
alliance. For more information about the alliance click
here.
Overview
The research of the alliance is comprehensive and ranges from finding
fundamental limits of the technology to novel protocols and application
algorithms to security and privacy issue, to performance optimized to the
particular goals of the network owners to integration with and interfaces to
humans. The integral component of the program is also rapid transition of
research innovations into military and civilian applications. Hence, the
research will result in paradigm shift in which the networking and
distributed processing will become a pervasive and ubiquitous part of
infusion of technology into military and society. The design principles for
autonomous, self-organizing and self-repairing networks, including sensor
networks, will be the most tangible outcome of research and wide military and
commercial deployment will follow as the result of transition efforts.
The program is unprecedented in terms of its geographical scope (UK-US),
length (ten years) and depth of industry-academia collaborations. The four
broad topics that we proposed are interconnected and developments in each one
will influence others. The numbers of university and industrial partners in
both countries and the collective expertise and leadership represented by the
PI's ensure that the consortium will be one of the leading forces in wireless
networking and sensing in the next decade. Being a part of such a team is the
most exciting aspect of this program.
Scope
The scalability problem of large-scale online
simulation for network management has to be tackled in several dimensions,
and the scalability gains achievable in these several dimensions have to
be closely integrated. The development of this integrated, scalable
online simulation system is the core of our approach.
The alliance will perform research in the four areas of (i) network theory
(ii) security across system of systems (iii) sensor information processing and
delivery and (iv) distributed coalition planning and decision making.
Dr. Szymanski's contributions will focus on area 3 in which he leads Project 9.
In the context
of sensor information processing and delivery, we will focus on schemes that
can characterize and improve the quality of information obtained from sensor
networks and mechanisms that can allow sensor networks to adapt themselves to
the operational context dynamically while reducing the human burden in
managing them to best serve the objectives of military operations.
Project 9 Summary
A key tenet of the network-enabled capability concept is that sensors and
actors are decoupled, unlike the traditional platform-centric model. Sensors
from one platform may provide information that enters the network and is
ultimately used by a variety of actors and decision makers in
operationally-diverse contexts. For several reasons, there is a need for
intelligent information fusion and semantic integration as well as for
sophisticated feedback loop from the actors and decision makers to networks and
ultimately sensors that controls sensor.s activities. First, sensor information
designed for use by a particular battle-space entity must be available for
wider use. Second, the tradeoffs of the sensor operation points must be
resolved based on the current needs of the actors and decision makers.
Third, the future strategic battle-space will feature a range of information
sources that extend well beyond the traditional notions of battlefield
sensor systems. To make sense of all collected information, the tools are
needed for countenancing increasingly sophisticated fusion-related processes
to enhance situation awareness and to avoid information overload. Quickly
changing and essentially unpredictable needs of each mission make it equally
important to enable mission commanders and warfighters to provide a feedback
on the current mission needs and priorities, while considering the
constraints posed by wireless and sensor networks.
Hence, there are several key aspects of managing data and infrastructures
in a distributed, resource-constrained, multi-modal sensor environment,
including:
-
Preventing information overload while capturing important information
through intelligent distributed sensor data fusion. This involves defining
and using a metric for quantifying the importance of information.
-
Respecting resource constraints of the operating environment, especially
energy and bandwidth in sensor networks.
-
Managing and controlling the sensor infrastructure.
-
Providing mission command with an operator control paradigm which lets them
decide parameters defining the constraints and objective functions of the
associated constrained optimization problems. These should be exposed as
meaningful higher-level parameters as opposed to detailed low-level system
parameter settings.
The first three items eventually enable the last one, which is the
high-level goal. Hence we discuss them in more detail next. Furthermore, we
discuss the state-of-the-art, and describe the open problems and challenges
which will be addressed by this project.
Point-of-Contact:
Boleslaw
K. Szymanski, Ph: 518-276-2714