SGN-21006 Advanced Signal Processing, 5 cr
Additional information
Suitable for postgraduate studies.
Person responsible
Ioan Tabus
Lessons
| Implementation | Period | Person responsible | Requirements |
| SGN-21006 2019-01 | 2 |
Ioan Tabus |
Final examinaion and a homework assignment |
Learning Outcomes
Student will learn advanced signal processing methods, especially linear optimal filter design, adaptive filters, spectrum estimation, nonlinear filters and how to select proper methods for signal processing tasks at hand. After completing the course, the student - Is familiar with the most important advanced signal processingr generic problems: optimal design, convergence, recursiveness in time, spectrum estimation; - Is able to start from the formulation of a problem formulation and utilize a number of typical algorithmic tools to derive the solution; - Knows what are the most important algorithms for optimal and adaptive filters: LMS, NLMS,RLS etc. - Acquires practice on simulating optimal and adaptive algorithms with given input data and extracting useful performance indices helpful in comparing various algorithms. - Knows how to integrate an optimal or adaptive filter in a number of important applications: echo cancelation, noise cancellation, channel equalization etc.
Content
| Content | Core content | Complementary knowledge | Specialist knowledge |
| 1. | 1. Deterministic and random signals: review of Fourier transform, Z transform, random variables, random signals, correlation, AR,MA, ARMA | ||
| 2. | 2. Optimal filter design (Wiener filter, Least squares, essentials of estimation, MLE, CramerRao) | ||
| 3. | 3. Adaptive filter design (LMS, NLMS, RLS ) | ||
| 4. | 4. Application areas of Optimal filter design and Adaptive filter design | ||
| 5. | 5. Spectrum estimation:Frequency spectrum (needed in machine function regime diagnosis, finding periodicities in time series), Direction of Arrival spectrum | ||
| 6. | 6. Nonlinear filters (median and order statistics filter family) |
Instructions for students on how to achieve the learning outcomes
The course is graded on the basis of answers to exam questions. Very good grade is obtained when exam questions are correctly answered and homework is accepted. Course acceptance threshold is approx. half of the maximum exam points. By volunteering to show exercise solutions and solving bonus items in the homework will be rewarded with additional points, to be added to the exam points, and finally the sum will determine the final grade.
Assessment scale:
Numerical evaluation scale (0-5)
Study material
| Type | Name | Author | ISBN | URL | Additional information | Examination material |
| Book | Adaptive Filter Theory | Simon O. Haykin | No | |||
| Book | Optimum Signal Processing | S. J. Orfanidis | No | |||
| Book | Spectral analysis of signals | Petre Stoica and Randolph Moses | No | |||
| Lecture slides | Ioan Tabus | Yes |
Prerequisites
| Course | Mandatory/Advisable | Description |
| SGN-11000 Signaalinkäsittelyn perusteet | Advisable | |
| SGN-11007 Introduction to Signal Processing | Advisable |
Correspondence of content
| Course | Corresponds course | Description |
| SGN-21006 Advanced Signal Processing, 5 cr | SGN-2607 Statistical Signal Processing, 6 cr |