JOURNAL OF NETWORKS (JNW)
ISSN : 1796-2056
Volume : 4    Issue : 1    Date : February 2009

Multi-Threshold Based Data Gathering Algorithms for Wireless Sensor Networks
Jun-Zhao Sun
Page(s): 30-41
Full Text:
PDF (230 KB)


Abstract
Sensor networks have recently attracted significant attention for many military and civil applications,
such as environment monitoring, target tracking, and surveillance. A high-level abstraction of sensor
networks forms the distributed database view, in which query is adopted to retrieve data from the
network. Continuous query is an important approach that is commonly used for collecting real-time
data. Sensor nodes have limited energy resources and their functionality continues until their energy
drains. Therefore, query for sensor networks, and especially continuous query, should be wisely
designed to extend the lifetime of sensors. This paper presents a query optimization method for
continuous data gathering in sensor networks. The method is based on the flexible setting of
thresholds and the analysis on the content to be sent over the network. Thresholds represent the
accuracy requirement of the application issuing the queries. Thus at each single sensor node,
instead of direct delivery of each result data, data communication is needed only when the current
data is out of the range constrained by the thresholds. A series of static and dynamic algorithms are
proposed for the flexible setting of threshold values. In particular, the paper studies deeply the
situation where phenomena under monitoring appear a periodical feature. An initiative data model
is first created, and then at each single sensor node, instead of direct delivery of each result data,
data communication is needed only when the current data is out of the range constrained by the
periodical data model. The performance of the proposed algorithms shows that the proposed
methods can achieve better performance than without performing the optimization.

Index Terms
Wireless sensor networks, continuous query, aggregation, threshold.