Paper Type

Master's Thesis


College of Computing, Engineering & Construction

Degree Name

Master of Science in Electrical Engineering (MSEE)



NACO controlled Corporate Body

University of North Florida. School of Engineering

First Advisor

Dr. O. Patrick Kreidl

Second Advisor

Dr. Alan Harris

Rights Statement

Third Advisor

Dr. Thobias Sando

Fourth Advisor

Dr. Brian T. Kopp

Department Chair

Dr. Osama M. Jadaan

College Dean

Dr. William F. Klostermeyer


Under similar conditions, products that are designed and used for similar tasks fail similarly. Developers may become aware of various product failure modes during the initial stages of new product generation, where redesign and failure mitigation processes can occur with minimal detriment to consumer safety. Developers strive to mitigate the potential for catastrophic failures. This thesis concentrates on when these failures occur outside of controlled conditions, specifically where the development of processes feature low accuracy sensing techniques that impact the safety and operation of the end user. This thesis develops a set of statistical analysis simulation techniques using two existing methods: Sequential Analysis and Quickest Detection. Through the comparison of method-specific features, this thesis aims to assist future researchers unfamiliar with these methods to understand the individual characteristics of each as they pertain to failure mitigation. Each detection method is subjected to investigation via a pair of sensor models, a strong sensor and a weak sensor. Variable detection settings are used to quantify the operational characteristics of these sensors and their individual means of analysis. This thesis then compares both statistical techniques to recognize their overall usefulness to the topic of product failure analysis and mitigation pertaining to lower accuracy sensing processes that require longer sampling periods for better informed decisions. It is ascertained that the Sequential Analysis technique is best used when the initial system state is not yet known to the observer. The Quickest Detection method should be utilized when the initial state of a system is known and it is imperative to detect, with minimal delay, the occurrence of a random change-point in the operational status of the system.

ETD1627AFA.pdf (3241 kB)