PhD Thesis Proposal Defence


"Pattern-Based Event Detection in Sensor Networks"

by

Mr. Wenwei Xue


Abstract:

Many sensor network applications, such as disaster monitoring and object
tracking, require effective techniques for detecting events of interest in
the physical world. Despite the simplicity in implementation, existing
threshold-based event detection approaches for sensor databases cannot
fully express the semantics of real-world events, and the threshold values
are difficult to be set accurately. In this proposal, we present a novel
pattern-based approach to event detection in sensor networks and integrate
the approach into an SQL-based in-network sensor query processing
framework. The key observation of our proposal is that events in sensor
networks can be abstracted into spatio-temporal patterns of sensory data
and that the problem of event detection can be converted into a pattern
matching problem. In addition to handling general spatio-temporal patterns
for events, we define three types of basic spatial patterns as well as
five types of basic temporal patterns that are useful to capture many
real-world events in our pattern-based event specification. We encapsulate
the matching process of these patterns into system built-in methods and
develop execution techniques for the event-driven queries in our
framework. We adopt the contour maps of attributes as the building blocks
to support spatial pattern matching and propose communication-efficient
techniques for in-network construction and incremental update of these
maps. Considering the limited computation and storage on sensor nodes, we
design a cache manager on each node to maintain the historical data
required for temporal pattern matching. Our initial results with
real-world and synthetic workloads demonstrate the effectiveness and
efficiency of the proposed approach.


Date:     		Friday, 7 July 2006

Time:                   2:00p.m.-4:00p.m.

Venue:                  Room 5506
			lifts 25-26

Committee Members:      Dr. Qiong Luo (Supervisor)
                        Prof. Lionel Ni (Chairperson)
			Prof. Dik-Lun Lee
			Prof. Frederick Lochovsky


**** ALL are Welcome ****