Organizers:
http://www.ece.ncsu.edu/dept/directory/egrant
http://www.science.uva.nl/research/ias/people/f.c.a.groen/
http://www.decis.nl/content/view/39/18/
Uwe
Hanebeck (University of
Karlsruhe, Germany)
http://isas.uka.de/site_en/people/hanebeck/hanebeck.html
Thomas
C. Henderson (University of
Utah, USA)
Art
Sanderson (Rennselaer
Polytechnic Institute, USA)
http://www.ecse.rpi.edu/homepages/acs/
Scope of the Tutorial
Recently
researchers have begun examining how to best exploit strong models in
order to
better utilize sensor networks; these models may be either of the
environment
being sensed, or of the sensor network itself. The
models are developed in such a way as to provide
solution techniques otherwise difficult or not possible, to improve the
accuracy
or efficiency of the solution, or to provide a framework in which
uncertainly
can be explicitly characterized and manipulated.
The
goals of this tutorial include:
·
Survey
state-of-the-art in this area by researchers active in the area and
representing a diversity of approaches.
·
Provide
a statement of the core intellectual challenges in this area.
·
Consider
which applications may best benefit from this approach.
·
Determine
what benchmarks and datasets may be useful and made publicly available
for the
research community.
In
addition to a brief introduction to sensor networks prepared mainly as
a
handout, we propose to cover the following areas from our own research
activities:
·
(Grant)
Acoustic models and the embedding of sensor networks in fabric.
·
(Groen)
Bayesian inference models for sensor networks.
·
(Hanebeck)
Probabilistic and statistical methods in sensor networks.
·
(Henderson)
Exploitation of PDE models of physical phenomena in computational
sensor
networks.
·
(Sanderson)
Model exploitation in sensor networks and underwater robot coastal
water
monitoring and model-based adaptive sampling.
Our
goal is to do this in such a way that novices and experts will benefit.
Much
work has been done in sensor network devices, and operating systems as
well as
some basic information processing areas (data aggregation, gradient
calculation, etc.). The topic of
higher-level model based exploitation of sensor networks is an exciting
new research
area and a timely topic for the robotics and automation community.