Processamento Digital de Sinais

processamento digital de sinais

Instructors:

Jorge Salvador Marques	lectures, problems and lab	email: jsm at isr.ist.utl.pt
Bruno Guerreiro	lab	email: bguerreiro at isr.ist.utl.pt

Office hours: JSM - Tuesday (11:00-12:30) and Thursday (9:30-11:00), 7^th floor of North Tower. BG - Tuesday (15:30-17:00) and Friday (14:30-16:00), room 5.15, 5th floor of North Tower.

Students interested in meeting with the professors should send an email 24h in advance.

Objectives:

This course is an introduction to the processing and manipulation of discrete signals using digital computers.

We wish to develop student’s ability to solve signal processing problems and manipulate real signals.

See the following paper in which 3 renowned experts (one of them is Portuguese) discuss what Signal Processing is and what is the future of Signal Processing.

J. Moura, J. Flanagan, N. Jayant, The discipline of Signal Processing, IEEE Signal Processing Magazine, 174-176 Nov 2013.[link]

Work requirements

The DSP course has lectures (3h/week), problem and lab sections (1.5h/week on average).

It is assumed that the student spends 4h/week solving problems and preparing the Lab. We believe this is the involvement required to achieve good results in this course.

Syllabus:

The syllabus is organized into two parts comprising topics from classic digital signal processing and model based signal processing.

Part 1 – Signal Transforms e Filtering

1. Discrete Systems and Signals

Elementary signals. Linear and time-invariant systems. Convolution. Difference equations. Frequency response of LTI systems.

2. z Transform

Definition and region of convergence. Properties. Inverse transform of rational functions.

3. Discrete Fourier Transform

Definition. Properties. Periodic convolution. Filtering based on the DFT. Applications.

4. Digital Filters

FIR and IIR filters. Canonic forms. Design of IIR filters based on continuous filters. Design of FIR filters based on windows. Applications.

Parte 2 – Model Based Signal Processing

5. Random Signals

Random signals. 2nd order characterization. Gaussian signals.

6. Classic parameter estimation

Characterization of estimators. Crámer-Rao bound. Minimum Variance method. Maximum likelihood method. Least squares method. Applications to radar, sonar, echo cancellation and channel equalization.

7. Bayesian methods

Prior distribution. A posteriori distribution. EQM and MAP methods..

Textbooks:

The syllabus comprises two parts each of them being covered by a different textbook. The adopted textbooks are classic references in their area:

· A. Oppenheim, R. Schafer, J. Buck, Discrete Time Signal Processing, Prentice Hall, 1999 (OSB)

· S. Kay, Fundamentals of Statistical Signal Processing. Estimation Theory, Prentice Hall, 1993 (K)

The association between the syllabus topics and the textbook chapters is the following:

1-OSB2 2-OSB3 3-OSB8 4-OSB6 5- 6-K2,3,7,8 7-K10

Grading:

Student grading is based on continuous assessment with two components:

· Theory: evaluated by 2 tests (minimum requirements: 9.5/20) with a weight of 70%.

· Lab works: 6 lab works (2 pages reports prepared during the lab session) (minimum requirements: 9.5/20), with a weight of 30%.

There is a recovery exam for those who don’t meet the minimum requirements in the tests.

Dates: 1st Test: April 20; -- 2nd Test: June 6;-- Recovery exam: June 29.

Lab:

Lab sessions are based in Matlab. The first session is an introduction to Matlab and it is not evaluated. The other sessions are associated to 6 Lab works each of them with a different processing problem.

These works are done in groups of 2 students and by the end of the session each group should deliver a small report in pdf format with figures and interpretation of the results (max: 2 pages).

There will be 2 extra sessions for the recovery of previous works. These sessions will take place after the 3^rd and 6^th works.

Problem series:

1 2 3 4 5 6 7