Date of Award

May 2020

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Computer Engineering

Committee Member

Daniel Noneaker

Committee Member

Brian C Dean

Committee Member

Yongqiang Wang

Committee Member

William Harrell

Abstract

The dissertation is concerned with the efficient resolution of data congestion on wireless sensor networks (WSNs). WSNs are of increasing relevance due to their applications in automation, industrial processes, natural-disaster detection, weather prediction, and climate monitoring. In large WSNs where measurements are periodically made at each node in the network and sent in a multi-hop fashion via the network tree to a single base-station node, the volume of data at a node may exceed the transmission capabilities of the node. This type of congestion can negatively impact data accuracy when packets are lost in transmission. We propose flexible congestion management for sensor networks (FCM) as a data-collection scheme to reduce network traffic and minimize the error resulting from data-volume reduction. FCM alleviates all congestion by lossy data fusion, encourages opportunistic fusion with an application-specific distortion tolerance, and balances network traffic. We consider several data-fusion methods including the k-means algorithm and two forms of adaptive summarization. Additional fusion is allowed when like data may be fused with low error up to some limit set by the user of the data-collection application on the network. Increasing the error limit tends to reduce the overall traffic on the network at the cost of data accuracy. When a node fuses more data than is required to alleviate congestion, its siblings are notified that they may increase the sizes of their transmissions accordingly. FCM is further improved to re-balance the network traffic of subtrees such that subtrees whose measurements have lower variance may decrease their output rates while subtrees whose measurements have higher variance may increase their output rates, while still addressing all congestion in the network. We verify the effectiveness of FCM with extensive simulations.

Share

COinS
 
 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.