ELG 5377

Description: Description: Description: http://www.site.uottawa.ca/sitehead.gif

ELG 5377 Adaptive Signal Processing

Fall 2010

Professor: Claude D'Amours

Office: STE 5013

Phone : 562-5800 ext 6249 or 5916

Email : damours@site.uottawa.ca

Course Webpage : www.site.uottawa.ca/~damours/courses/ELG_5377

Course Description

Theory and techniques of adaptive filtering, including Wiener filters, gradient and LMS methods; adaptive transversal and lattice filters; recursive and fast recursive least squares; convergence and tracking performance; implementation. Applications, such as adaptive prediction; channel equalization; echo cancellation; source coding; antenna beamforming; spectral estimation.

Schedule

Monday 2:30-4:00 VNR2075

Wednesday 2:30-4:00 MNT204

Marking Scheme

Assignments	20%
Term Paper	30%
Final Exam	50%

Textbook

Ali H. Sayed, Fundamentals of Adaptive Filtering, Hoboken, NJ: Wiley and Sons, 2003.

Reference

Simon Haykin, Adaptive Filter Theory, 4^th Ed., Prentice Hall: Upper Saddle River NJ, 2002

P.S.R. Diniz, Adaptive Filtering: Algorithms and Practical Implementation, 3^rd Ed., New York: Springer, 2008.

Assignments

Students will be assigned 4-6 assignments throughout the semester. The assignments are due one week after they are assigned.

Term Paper

Each student will write a term paper. The subject of the term paper should be the application of adaptive signal processing to a practical problem. The term paper should have some analysis and/or simulation to demonstrate its usefulness in solving the problem at hand. Students should consider the figures of merit that have been discussed in class (training time, tracking capabilities, excess error etc). A proposal should be given to the professor no later than Feb. 8 and the final paper is due on the last day of class. If time permits, students will be asked to present the results of their papers in the last two weeks of class.

Final Exam

An open book final exam will be held following the final lecture.

Topics

1) Review of Random Variables and Optimal Estimation

2) Stochastic Models: Autoregressive Model

3) The Filtering Problem: Wiener Filtering

4) The Correlation Matrix and its properties, Eigenanalysis, Eigenfilters

5) Some applications of Wiener Filters

6) Steepest descent algorithm

7) Stochastic gradient Algorithms: Least mean squares (LMS)

8) Method of Least Squares

9) Recursive Least Squares

10) Steady State Performance of LMS and RLS algorithms

11) Tracking Performance of Adaptive Filters

Presentation on simulation of LMS, NLMS and Affine Projection Adaptive Filters

Solution to assignment 1 in word format. Class average on Assignment 1 = 88%.