ECE6271 -  Adaptive Filtering  (3-0-3)

Prerequisites:  ECE 4270

Corequisites:  None

Catalog Description:  Basic theory of adaptive filter design and implementation. Steepest descent, LMS algorithm, nonlinear adaptive filters, and neural networks. Analysis of performance and applications.

Textbook(s):  Sayed, Ali H., Fundamentals of Adaptive Filtering, Wiley and Sons, 2003. (optional)

Topical Outline:  Background (1 week) Eigenanalysis Review of Discrete-time random processes FIR Wiener filters (1 week) Derivation of the Wiener-Hopf equations Principle of orthogonality Problems and applications Solving the Wiener-Hopf equations. The Discrete Kalman Filter (1 week) Gradient-based adaptive filters (4 weeks) Steepest descent The LMS algorithm Performance Analysis Variations on the LMS algorithm Examples and comparison of techniques Applications Gradient Adaptive Lattice Filter (0.5 weeks) Recursive least squares (1.5 weeks) Transversal filters Lattice filters - optional Performace of the RLS algorithm Tracking of time-varying systems (0.5 weeks) - Chapter 16 of Haykin Adaptive IIR filters (1.5 weeks) IIR LMS Fientuch and Horvath algorithms HARF and SHARF Examples and applications Nonlinear adaptive filters (3 weeks) - Chapters 18-20 of Haykin Order statistic adaptive filters and Volterra systems Blind deconvolution - decision directed feedback Back propagation learning Radial basis function networks Examples Other Applications - optional Adaptive line enhancement Adaptive spectrum estimation, frequency tracking Adaptive signal modeling