Date of Award

5-2014

Document Type

Thesis

Degree Name

Master of Science (MS)

Legacy Department

Computer Science

Committee Chair/Advisor

Hallstrom, Jason

Committee Member

Malloy , Brian

Committee Member

Sorber , Jacob

Abstract

Due to an increasing demand to monitor the physical world, researchers are deploying wireless sensor networks more than ever before. These networks comprise a large number of sensors integrated with small, low-power wireless transceivers used to transmit data to a central processing and storage location. These devices are often deployed in harsh, volatile locations, which increases their failure rate and decreases the rate at which packets can be successfully transmitted. Existing sensor debugging tools, such as Sympathy and EmStar, rely on add-in network protocols to report status information, and to collectively diagnose network problems. Some protocols rely on a central node to initiate the diagnosis sequence. These methods can congest network channels and consume scarce resources, including battery power. In this thesis, we present Corl8, a system for analyzing diagnostic traces in wireless sensor networks. Our method relies on diagnostic data that is periodically transmitted to a network sink as a part of the standard sensor payload to enable fault diagnosis. Corl8 does not require any specific data to be present in the system, making it flexible. Our system provides an interactive environment for exploring correlated changes across different diagnostic measures within an individual node. It also supports processing on a batch level to automatically flag interesting correlations. The system's flexibility makes it applicable for use in any wireless sensor network that transmits diagnostic measures. The analysis methods are user-configurable, but we suggest settings and analyze their performance. For our evaluation, we use data from five real-world deployments from the Intelligent River(R) project consisting of 36 sensor nodes.

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.