 |
How to efficiently and reliably monitor and react to dynamic changes recorded
by networks of multimodal sensors at a large number of remote sites using
wireless infrastructure is the central focus of this proposal. Our objective is
to make significant contributions leading to the development of “Smart” sensor
networks for visual context capture and interactivity.
A couple of important highlights of the proposed research are:
1. Integrated systems oriented approach: Our research team includes
investigators representing Networking, Digital Image/Signal Processing, Embedded
Systems, and Computer Vision areas, to offer insights in a synergistic manner.
The team will pursue an integrated systems oriented research with balanced
theoretical, analytical, and experimental components.
2. Real-world
application focus: Our research will consider a particularly demanding and
important application, namely large-scale video surveillance networks.
One of our near term goals is to synthesize a comprehensive networking
architecture that can support this application, that will serve as the basis for
a prototype system that we hope to develop. The exercise for synthesizing this
architecture in an application driven context will form the basis for more
focused fundamental research, as well as future iterations on overall wireless
sensor networking architectures.
This proposal presents a systematic, two-year plan to synergistically
investigate important research problems from the distributed video analysis as
well as wireless networks fields. Our team has been active in both of these
areas in the past, in a somewhat independent manner. This proposal allows us to
develop a cohesive research agenda to develop integrative systems-oriented
solutions involving expertise from wireless networks and distributed video
capture/transmission/analysis. Networks of video cameras are expected to grow as
an essential sensory module for a range of applications, such as monitoring and
responding to activities in an “intelligent” environment. Wireless networks that
meet the needs of multimedia sensors, including video sensors, present some very
interesting challenges, opportunities, and constraints. The proposed research
shall focus on several fundamental issues.We shall assume that a large number of
sensors must be deployed over some possibly large coverage region. Some of these
sensors are stationary, may generate video traffic continuously, and have access
to AC power. Some video sensors are battery powered and mobile. Other types of
sensors (non-video) generate significantly less traffic relative to the video
sensors. Each sensor is attached to a communications node which sends/receives
sensor data, receives sensor network commands, and serves as a relay node in an
ad-hoc wireless mesh. Both centralized and distributed management and control
will be considered and compared on the basis of chosen objective criteria. We
shall also seek to develop a base of knowledge potentially applicable to any
wireless sensor network, independent of application.
The following CWC faculty are participating in this research project: Mohan Trivedi(lead PI),
Truong Nguyen, Anthony Acampora, Sujit Dey, Rene Cruz, and Ramesh Rao.
|
|
|
 |
|
Affiliated Faculty
CoRe Research Projects