IMPROVING THE ESTIMATION ACCURACY OF SIGNALS WITH LOW SIGNAL-TO-NOISE RATIO USING OPTIMAL COMBINING METHODS

IMPROVING THE ESTIMATION ACCURACY OF SIGNALS WITH LOW SIGNAL-TO-NOISE RATIO USING OPTIMAL COMBINING METHODS

By Timothy Arthur Falkner

Thesis Advisor: Dr. Ali Abedi

A Lay Abstract of the Thesis Presented

in Partial Fulfillment of the Requirements for the

Degree of Master of Engineering

(in Engineering Physics)

May, 2010

            The effectiveness of modern medical equipment depends largely on the interpretation of the data that these machines provide. A thorough understanding of the output signals of these devices is essential in providing timely and accurate diagnoses of patients. For example, an electroencephalogram (EEG) can provide insight on brain function by recording electric voltage levels from different parts of the brain. EEG signals that are properly resolved can be used to diagnose a variety of brain impairments. This thesis is the result of collaborations among the Wireless Sensor Networks (WiSe-Net) laboratory in the Electrical and Computer Engineering Department, the Graduate School of Biomedical Sciences (GSBS) at the University of Maine, and the Maine Institute for Human Genetics and Health (MIHGH).

            One of the goals of this thesis is to enhance signal processing capabilities of these research groups to better investigate the developmental abilities of newborns who have been exposed to opiates in utero. To accomplish this, advanced signal processing techniques are utilized to improve the quality of the recorded of brain activity.

            Due to the age and conditions of the infants, the act of recording brain activity can be challenging. An improvement in our ability to better characterize weak signals will help us understand what processes are taking place in the infants' brains. The dynamic nature of this particular dataset has presented some unique challenges. We focused on developing methods that help remove noise and increase the reliability of a wide range of signals. This will allow usage of the research results in a variety of applications. The main goal is to maximizethe amount of information that can be gained from multiple experiments.