Database Reference
In-Depth Information
spatio-temporal constraints and can be non-spontaneous. Meanwhile,
like general event streams, sensor event streams can be generated with
very high volumes and rates. Primitive sensor events need to be filtered,
aggregated and correlated to generate more semantically rich complex
events to facilitate the requirements of up-streaming applications. Mo-
tivated by such challenges, many new methods have been proposed in
the past to support event processing in sensor event streams. In this
chapter we will mainly focus on complex event processing over sensor
streams including RFID data streams, as the trends show that primi-
tive event processing is gradually moved to the edges of event sources.
Next we will present an overview of sensor event processing techniques,
including event specification languages, event detection models, event
processing methods and their optimizations. During the discussion, we
will pay special attention to the distinct challenges of event processing
over sensor streams and RFID streams.
2. Event Processing in Sensor Streams
Event detection approaches in sensor networks can be categorized
into statistical methods [1], topographical techniques [2-4], and edge
detection algorithms [5-7].
Statistical methods.
A statistical method is presented in [1] for
detecting generic homogeneous regions without the benefit of an a priori
predicate to identify events. Instead, it uses a kernel density estimator
to approximate the probability density function of the observations. It is
suggested that the detection routine be rerun periodically to accommo-
date the scenario of any new regions or holes that evolve in the midst of
tracking. Even so, there is not an elegant way to handle new detections
and persistent tracking at the same moment.
Topographical methods.
An example of the topological and con-
tour mapping technique is Iso-Map [2], which builds contour maps based
solely on the reports collected from intelligently selected “isoline nodes”
in the network. This approach is limited to a plane. Another technique
[3] collects time series of data maps from the network and detects com-
plex events through matching the gathered data to spatio-temporal data
patterns. Essentially the work provides a basic infrastructure and then
outsources the problem solution to the user, instead of directly solving
the event tracking problem. SASA [4] uses a hole detection algorithm
to monitor the inner surface of tunnels, where sensor nodes may be dis-
placed due to collapses of the tunnels.
