School Seminars and Colloquia

Stochastic Routing Models in Sensor Networks

Completion talk

by Paul Keeler


Institution: Department of MASCOS/Mathematics and Statistics, The University of Melbourne
Date: Fri 5th March 2010
Time: 3:15 PM
Location: Theatre 2, Old Geology

Abstract: The technology of sensor networks promises numerous applications. The
random and dynamic structure of sensor networks has motivated greedy
data-routing algorithms. We examine a recently-developed mathematical
model for studying the advancement of messages under greedy routing. We
introduce a model framework based on homogeneous spatial Poisson
processes, and address stochastic dependencies arising in the system. A
simple power-saving sleep scheme is included, and its effects on the
system are addressed to reveal that it reduces one of the dependencies.
Expressions for the distribution of the multihop advancement of messages
are derived. These expressions involve high dimensional integrals, which
are evaluated with quasi-Monte Carlo integration methods coupled with
importance sampling. The subsequent results are shown to agree well with
those obtained via routing simulations. We conclude by examining model
extensions, such as randomly-distributed transmission radii, and giving
future research directions.

For More Information: contact - Nathan Clisby. email: n.clisby@ms.unimelb.edu.au