Graduate Courses BIOE 3528

BIOE 3528 (EE 3528) - Time-Frequency Signal Analysis

Instructor: Loughlin

Credits: 3

Term: Spring

Description: Many signals exhibit frequency characteristics that change over time. Examples include natural signals such as speech, marine mammal sounds, heart rate, electromyographic activity, among numerous others, as well as man-made signals like machine vibrations, sonar, radar, and communications signals. Understanding these changes is important because they are often indicative of the underlying processes that generated the signal, and/or the changes encode information. Time-frequency analysis, also called time-varying spectral analysis, is a technique for studying the time-dependent spectral changes in a signal. This course covers the fundamentals of time-frequency analysis, and its application in a variety of areas including biosignal analysis, acoustics, and machine vibrations. Topics include instantaneous frequency and instantaneous bandwidth, why spectra change, the AM/FM model of signals,the short-time Fourier transform/spectrogram, the Wigner distribution, and the generalized Cohen-class time-frequency distributions. Material will be presented from the current and classic literature and a text on the subject. Students will have the opportunity to implement some of these techniques and apply them to data from their research and/or to data provided, and present their work in a final written report. Students will also give an oral presentation of a journal paper of their choosing from the subject literature. The final course grade will be determined by the written project report (30%), the oral journal paper presentation (20%), a midterm exam (30%) and homework/quizzes (20%).

Prerequisites: Basic knowledge of Fourier analysis and digital signal processing with Matlab (e.g., ECE2523 or equivalent).

Required Texts: Time-Frequency Analysis, by Leon Cohen (Prentice Hall, 1994)

Course Objectives:

Topics Covered:

Class/Labratory Schedule: