RESEARCH COMMONS
LIBRARY

Information-driven Data Gathering In Wireless Sensor Networks

ResearchCommons/Manakin Repository

Information-driven Data Gathering In Wireless Sensor Networks

Show simple item record

dc.contributor.author Wang, Jing en_US
dc.date.accessioned 2010-11-01T21:28:46Z
dc.date.available 2010-11-01T21:28:46Z
dc.date.issued 2010-11-01
dc.date.submitted January 2010 en_US
dc.identifier.other DISS-10713 en_US
dc.identifier.uri http://hdl.handle.net/10106/5113
dc.description.abstract Since the advance of wireless technology enables the mass production of low-cost, small-sized sensor nodes, sensor nodes can be densely deployed in an area for high tolerance to node failure or to achieve better coverage statistically. The redundancy of sensor nodes results in the temporal and spatial correlation of sensory data, which motivates the information-driven data gathering approaches for wireless sensor networks (WSNs).Since existing approaches target at the sensory data that are already highly correlated with each other, little attention has been paid to the idea of changing the sampling schedules of the sensor nodes to reduce the correlation among sensory data. In this dissertation, sampling strategies and the relevant medium access control (MAC) protocol are presented to demonstrate how the correlation can be reduced through adjusting the sampling time shifts of sensor nodes.The asynchronous lossless data gathering strategy aims at extending the sampling cycle of individual node while guaranteeing the original signal to be fully recovered by the sink. Based on the correlation signal model, details of the collaborative reconstruction of the original signal are presented. An exponential temporal-spatial correlation model is introduced for presenting lossy data gathering strategies. It is justified by real data collected from wireless sensor networks. Regarding lossy data gathering applications, the sensor nodes take samples asynchronously to obtain more informative samples. Furthermore, the entropy of the joint Gaussian random variables is adopted to quantify the improvement on the quality of information obtained from the asynchronous samples. Oppeinhem's inequality is applied to prove the entropy is increased by introducing a non-zero temporal correlation parameter. A recursive algorithm is designed to solve the optimal asynchronous sampling problem with a set of sub-optimal sampling time shifts. Bounds on the performance of the three asynchronous sampling strategies are derived respectively.Motivated by the benefit of asynchronous sampling strategies, an information-driven MAC protocol is proposed to avoid the severe collisions of event reports in the event detection applications. Other than choosing a subset of nodes to report to the sink, the proposed protocol assigns sampling shifts to nodes in order to change the bursty traffic into a streamlined traffic. Consequently, the MAC performance is improved by essentially replacing the collision prone traffic with the streamlined one. An optimal probability model is adopted for selecting nodes' transmission slots that minimize collisions and in turn reduce the correlation among event reports. Through theoretical analysis and simulations, it is shown that the protocol relates the MAC performance with the information quality of event reports, which is quantified by the Cramer-Rao lower bound (CRLB) of parameter estimation. In addition to the benefit of reduced collision probability, the CRLB is lowered by the proposed MAC protocol after the nodes' s sampling time moments are shifted from each other. en_US
dc.description.sponsorship Das, Sajal en_US
dc.language.iso en en_US
dc.publisher Computer Science & Engineering en_US
dc.title Information-driven Data Gathering In Wireless Sensor Networks en_US
dc.type Ph.D. en_US
dc.contributor.committeeChair Das, Sajal en_US
dc.degree.department Computer Science & Engineering en_US
dc.degree.discipline Computer Science & Engineering en_US
dc.degree.grantor University of Texas at Arlington en_US
dc.degree.level doctoral en_US
dc.degree.name Ph.D. en_US

Files in this item

Files Size Format View
Wang_uta_2502D_10713.pdf 1.711Mb PDF View/Open

This item appears in the following Collection(s)

Show simple item record

Browse

My Account

Statistics

About Us