Adaptive Signal Processing ET2415 (ETC004)
Course start: January 20, 2009. Time: 15:15-18:00, in Gradängsal 2304A
Curriculum (English translation)
The purpose of the course is to provide both knowledge and background in adaptive and optimal systems as well as provide understanding and experience of applied signal processing problems involving these systems. The goal is that the student should be able to recognise situations where adaptive systems are likely to provide a fruitful solution and to be able to implement this type of solution in the industry as well as in continued studies. The course is given in the form of lectures, exercises, home assignments and laboratory experiments as well as projects. The project is a large part of the course. It can be either be a software based (Matlab) project or a more applied project based on a digital signal processor. For the theoretically interested student it is possible to carry out a more theoretical project.
The following material is available or will be available during the course:
- Presentation of available projects 2007
- Project presentations
- The Filtered-x LMS algorithm
- Limited Numerical Precision and The LMS Algorithm, and The Leaky LMS - Solution
- Table of formulas
- Study instructions and exercise list
- Answers/Soulutions to selected exercises
- Computer laboratory experiments
- The mean squared error as a function of the coefficient vector
- The RLS algorithm
- Home Assignment, Week 6.
- Home Assignment, Week 8.
- Home Assignment, Week 11.
- Solution Home Assignment, Week 6.
- Solution Home Assignment, Week 8.
- Solution Home Assignment, Week 11.
Monson H. Hayes
Statistical Digital Signal Processing and Modeling Wiley 1996
Material från institutionen
Old important exam problems will be solved in room 2415 on Thursday, 15/3-2007, 14:00-17:00!
Oral presentations of the projects will be held in room 2214 in Karskrona on the 17/6 from 13.00 17.00.
List of projects:
A short presentation of the projects planned for spring 2006. In a dialogue with the participants a more detaild project desription will be derived. Also, study the information concerning the DSP part of the projects on Jörgens project page.
1. Adaptive Notch Filter Implementation on a DSP.Implement the adaptive notch filter for one or more sinusoidal disturbances by using the LMS filter-method. A typical application could be suppression of acoustic feedback in microphone-amplifier-speaker systems. The principle for the adaptive notch filter is described in Widrows book Widrow's book (pdf 700k, contact Lars for the password). Project manual
2.Detecting fetus heartbeat from signals recorded on pregnant woman.This project aims at detecting the hearbeats of a fetus by examining a recording made from a pregnant woman. A dominating disturbance is in this case the heartbeats of the mother. In addition, other disturbances like 50 Hz hum from the power grid, is present.
3. Demonstration of adaptive algorithms using Matlab.In this project, your task is to develop a demonstration program for the adaptive algorithms covered in the course. The demo program should be interactive and able to present vital data, e.g. error functions, learning curves, coefficients convergence etc. Matlab offers the possibility of creating powerful graphical user interfaces and this should be utilized in the demo program. The Clarksson book described some different LMS algorithms (pdf, 1800k) as well as RLS (pdf, 1000k).
4. Adaptive Channel Equalization.Many occupations of today requires the usage of personal preservative equipment such as a mask to protect the employee from dangerous substances. The goal of this project is to investigate the possibility of placing a microphone for communication purposes inside such a preservative mask and perform a digital channel equalization on the speech path in question in order to enhance the speech intelligibility.
||MLS e-mail: email@example.com
5. Active Control of Noise.A classic DSP experiment. Noise originating from one loudspeaker is to be cancelled, in a point of reference (a microphone), by creating counter-noise from another loudspeaker. This project employs the Filtered-X-LMS algorithm, depictured in Widrow page 292-294 (288k) and in the tutorial on the Filtered-X LMS-algorithm. Project manual
Some previous given exams, in PDF-format
Regarding the course book:
Home page for Hayes: Statistical Digital Signal Processing and Modeling
Matlab Primer, for Matlab 4 (pdf, 320k)