Optimal Signalbehandling, OSB, ETT074
Electrical and Information Technology, Lund University.

- Optimal Signal Processing 2010,
Preliminary Course description-

Introduction

Welcome to the course in Optimal Signal Processing. The course will he held by Bengt Mandersson (phone: 2229021, room E:2540).

Lectures: 14 h. Exercises: 28 h. Laboratory work: 8 h.

Aim: The course presents the fundamentals on statistical signal processing and treats the development of optimal methods and how these methods can be applied. The traditional approach to designing filters, usually based on specifications on "passband/stopband", is replaced by a design based on the statistical properties of the signal and the noise.

Description: Discrete-time signals, matrix representation, eigenvalues and eigenvectors, stochastic processes, correlation matrix, signal modeling e.g. by. Prony's method, stochastic models (AR and MA), normal equation, Levinson recursion, lattice filters, forward and backward prediction, Wiener filtering (FIR and IIR structures), spectrum estimation - non-parametric and parametric methods, frequency estimation (MUSIC).

 

Litterature

Monson Hayes, "Statistical digital signal processing and modeling", John Wiley, 1996. ISBN: 0471594318.

Extra material which can be downloaded from the homepage.

Information during the course

Information is  found at the homepage www.eit.lth.se. 

Lectures and exercises

Lectures and exercises: Bengt Mandersson (room E:2540, bm@eit.lth.se).
Lectures: Tuesday 15.15-17.00, room E:1406.

Exercises: Wednesday 10.15-12.00, room E:1147, Thursday 13.15-15.00, room E:3316. 

Laboratory work

Laboratory work is compulsory. Please sign up for date and time on the homepage on September 16 2010.
Information: Birgitta Holmgren, tel. 2229020, bh@eit.lth.se.

Schedule: 
Laboratory work 1:   Gruop A: Monday         September   27,    13.15-17.00
                            Group B: Wednesday    September   29,     13.15-17.00
                            Group C: Friday           October      01,      08.15-12.00

Laboratory work 2:   Grupp A: Monday        October       04,    13.15-17.00
                            Group B: Friday          October       08,     08.15-12.00
                            Group C: Monday        October       11,     13.15-17.00

Computer exercises

Two exercises use computers, (11/9 and 25/9) and will be in room E:4115. Please download the manual for this from the homepage in advance.

 

 

Schedule for the lectures

Tuesday 15.15-17.00, room E:1406.

 

F1 Chapter 1,2,3. What is optimal signal processing , Models, Applications

F2 Chapter 4. Signal modeling

F3 Chapter 5. Levinson-Durbins recursion

F4 Chapter 6. Lattice filter

F5 Chapter 7. Optimal filters - Wienerfilter, prediction

F6 Chapter 8. Spectral analysis

F7 Chapter 8. Spectral analysis

Schedule for the exercises

V1 2.5, 2.14, 2.16, 3.1, 3.3, 3.5abc, 3.6b, 3.8ab, 3.10, 3.15a-e, 3.17, 3.25, 3.26

V2 4.2, 4.6, 4.10, 4.18, 4.19, (4.21, 4.11)

V3 4.24, 5.1, 5.6, 5.12ab, 5.17, 5.20ab, (5.10)

V4 5.22, 6.1, 6.3, 6.8, 6.10a (6.4ab)

V5 7.1, 7.2, 7.7, 7.10, 7.12, 7.15

V6 7.4, 7.5, 8.3, 8.5 8.7, 8.13ab,

V7 8.22ab, 8.28a+partly b, problems from old exams
.

Examination

The examination consists of a written exam. The written exam consists of 5 problem. Each problem will give maximum 1 mark. Limits 0-1.9 not passed, 2.0-2.9 gives grade 3, 3.0-3.9 gives grade 4, 4.0-5.0 gives grade 5. The date for the final exam is October 21, 14.00-19.00, 2010. You may use the textbook during the exam. To pass the course, you must pass the final exam and fulfil the laboratory works.