NEW:   June 20: Letter grades are posted. (AA=40, BA=35, BB=30, ... , DD=10,FD=05,FF=01, NA=00)
NEW:   June 14: Final Exams and homeworks are graded. You can learn your scores from here. Check the class distribution from here. See the overall scores from here.

Homeworks are graded as follows: Hw 3 is out of 40, Hw4 out of 50, Hw5 out of 20. Bonus points are given for detailed solutions

              June 10: Final exam solutions are available
              June 1: Please check the page of a similar course given by Dr. Messerschmitt in UC Berkeley. You can find exam problems, hw solutions in here. You are not required to study the Berkeley course materials or the exam, use them as a suplementary reading on the topic.
               May 31: Some recent research papers from adaptive filtering literature are posted. You should be able to follow them without so many difficulties. The contents of these papers are not included in Final. If you are interested in what is going on in the current literature, you can check them out.
              May 31: 2nd midterm is graded, check the links below to get your score.
              May 31: Final Exam Day / Time : 4th June (Saturday), 13:00, EA 201
              May 27: Final exam of 2004 is available for study.

             May 11: Homework #5 is posted (due: May 20)
             May 11: Midterm-2 of 2004is available for study.
             May 2:First Midterm results are announced (check learn your grades, distribution of grades links)
             May 2: ****JAVA Applet Comparing Different Adaptive Filtering Methods****
             May 2: First Midterm results are announced (check learn your grades, distribution of grades links)
            April 26: Homework #4 (Due: May 6 postponed to : May 9 (Monday). Important: 20% bonus on your grade if you submit on May 6)

            April 21: Midterm #1 Solutions
            April 20 : Paper on IIR Adaptive Filters (Required reading), Feintuch's Paper
            April 08 : Homework #3 is assigned (Due: April 20 - postponed to 23rd) m-files : llms.m , convol.m ; Reading Assignments 1 and 2.
            March 23: Homework #2 (Due: March 30)
            March 4: Homework #1 (Due: March 11)

            Midterm #1 on 16th April 2005, Saturday, 10:30 - 12:30
            Midterm #1 of Spring 2004 (last year) available for study.

Reading Assignments:

  1.   Painless Intro. To Eigenvectors (up to conjugate directions chapter, page 21) ,
  2.   Lee - Wiener Legacy (historical notes on Wiener, optimal filtering method, early days of spectral estimation, i.e. correlation calculation machine, MIT in 1950's, collobarotors etc.)
  3.   Adaptive IIR Filters (Shynk, 1989)
  4.   Feintuch's Method - original paper with after publication comments/corrections
  5.   Some Recent Papers from Literature

Homework Assignments:


Learn Your Grades , Distribution of Grades
Reporting System is offline  

Spring 2005

 

EE 504  : Adaptive Signal Processing

Middle East Technical University

Electrical and Electronics Engineering Department

 

Course Outline:

 

In many signal processing problems, the signal of interest can be corrupted by noise or can be the output of an unknown system. For such signals the statistical signal processing methods have been developed to implement ensemble optimal (MSE) solutions. Echo cancellation, system identification, signal modeling and channel equalization are some application examples of these solutions.

 

Ensemble optimal filtering can be implemented in two ways. The first (and mostly theoretical) way of implementation is the estimation of the statistical characteristics before the algorithm design (offline estimation) and selecting the optimal set of coefficients using the estimated statistics. The second one is the implementation through adaptive methods that can both estimate-update the statistics (online estimation) and implement the solution.  

 

In this course, we examine the LMS, RLS, Kalman filter families and study different variations FIR and IIR adaptive filters.  In the first few weeks of the course, we will revisit Wiener filtering and gradually introduce the idea of adaptive filtering through the iterative solutions of linear equation systems.

 

The prerequisites are the familiarity with the concepts of basic linear system theory (especially the results on eigenvalues) and the stochastic processes.

 

 

Reference Books:

 

1.      Monson H. Hayes, Statistical Digital Signal Processing and Modelling, John Wiley & Sons, 1996.  (main reference)

 

2.      Simon Haykin, Adaptive Filter Theory, Prentice Hall, 1996.

 

3.      Athanasios Papoulis, Probability, Random Variables, and Stochastic Processes, Mc-Graw Hill, 1991.

 

 

Topics:

1. Introduction

  • Review of Random Processes
  • Mean Square Estimation Techniques, (Linear MSE estimation, optimal estimation)
  • Filtering the Random Processes
  • Moving Average (MA), Auto-regressive (AR) and ARMA processes

 

2. Wiener Filtering (Solving Wiener-Hopf Equations)

  • FIR, IIR, Causal IIR Wiener Filters
  • Iterative methods for the solution of Wiener-Hopf Equations

3. Adaptive Filters

  • LMS Filter
  • FIR, IIR , Normalized and other variations
  • RLS
  • Kalman Filters

 

4. Applications

 

 

 

Grading: Two midterms and a final, plus homeworks with Matlab assignments

 

(C.Candan)