Paper Type

Master's Thesis


College of Computing, Engineering & Construction

Degree Name

Master of Science in Electrical Engineering (MSEE)



First Advisor

Dr. Chui Choi

Second Advisor

Dr. Zornitza Prodanoff

Third Advisor

Dr. Susan Vasana

Department Chair

Dr. Murat M. Tiryakioglu

College Dean

Dr. Mark A. Tumeo


Wavelet analysis is a powerful mathematical tool commonly used in signal processing applications, such as image analysis, image compression, image edge detection, and communications systems. Unlike traditional Fourier analysis, wavelet analysis allows for multiple resolutions in the time and frequency domains; it can preserve time information while decomposing a signal spectrum over a range of frequencies. Wavelet analysis is also more suitable for detecting numerous transitory characteristics, such as drift, trends, abrupt changes, and beginnings and ends of events. These characteristics are often the most important and critical part of some non-stationary signals, such as physiological signals. The thesis focuses on a formal analysis of using wavelet transform for noise filtering. The performance of the wavelet analysis is simulated on a variety of patient samples of Arterial Blood Pressure (ABP 14 sets) and Electrocardiography (ECG 14 sets) from the Mayo Clinic at Jacksonville. The performance of the Fourier analysis is also simulated on the same patient samples for comparison purpose. Additive white Gaussian noise (AWGN) is generated and added to the samples for studying the AWGN effect on physiological signals and both analysis methods. The algorithms of finding the optimal level of approximation and calculating the threshold value of filtering are created and different ways of adding the details back to the approximation are studied. Wavelet analysis has the ability to add or remove certain frequency bands with threshold selectivity from the original signal. It can effectively preserve the spikes and humps, which are the information that is intended to be kept, while de-noising physiological signals. The simulation results show that the wavelet analysis has a better performance than Fourier analysis in preserving the transitory information of the physiological signals.