TEL 608E - Spectral Estimation Methods

Course Objectives

1. This course will introduce the fundamentals of the classic and modern spectral estimation techniques.
2. Techniques for comprehensive analysis and synthesis methods will be explained.
3. Advantages and disadvantages of estimation methods will be discussed.
4. For various signals, applications of the parametric and non parametric spectral estimation methods will be covered.

Course Description

An overview of concepts of spectral analysis covering traditional approaches and modern estimation methods will include: Probability density functions; expectations; processes; correlation and covariance, matrix and vector manipulations; discrete-time random signals; energy and power spectral densities. Nonparametric Spectral Estimations: Periodogram; averaged periodogram, Blackman-Tukey method, other refined periodogram methods (Bartlett, Welch, Daniell methods); Statistical properties of nonparametric spectral estimators; Rational spectral models -Autoregressive spectral estimation; Moving average spectral estimation; Autoregressive moving average spectral estimation; Yule-Walker equations; parameter estimation techniques; statistical properties. Filter-bank based approaches: Filter-bank interpretation of the periodogram, Capon method, relationship between Capon and AR methods. Advanced Concepts: Sinusiodal parameter estimation, MUSIC, ESPRIT. Introduction to Array Signal Processing: Problem statement; relationship to line spectral estimation. Array Signal Processing: Nonparametric and parametric estimation techniques; properties. Applications: Topics related to underwater acoustic, sonar, radar signal analyses and biosignal analysis.