| Course title |
Spectral Signal Analysis |
| Acronym |
13E053SAS |
| Study programme |
Electrical Engineering and Computing |
| Module |
Signals and Systems |
| Type of study |
bachelor academic studies |
| Lecturer (for classes) |
|
| Lecturer/Associate (for practice) |
|
| Lecturer/Associate (for OTC) |
|
| ESPB |
6.0 |
Status |
elective |
| Condition |
Digital signal processing, Stochastic systems and estimation |
| The goal |
Introduce students to basics methods for time-series analysis and spectral estimation. Enable students to practically implement and interpret the results of these algorithms using Matlab/Octave and Python. |
| The outcome |
Students will understand theoretical and practical aspects of classical and parametric algorithms for estimating power spectra of wide-sense stationary stochastic signals. Students will be enabled to properly choose, practically implement and adequately tune the spectral estimation algorithms, and to interpret the results obtained by applying these methods to realistic signals. |
| URL to the subject page |
https://automatika.etf.bg.ac.rs/sr/13e053sas |
| Contents of lectures |
Stochastic processes. Classical methods for spectral analysis: periodogram, Blackman-Tukey method. Parametric time-series modelling: AR, MA and ARMA models, linear prediction. Spectral estimation of AR models: Yule-Walker equations, Levinson-Durbin recursion, lattice filters, autocorrelation/(modified) covariance/Burg methods. Effect of measurement noise. Model order selection. |
| Contents of exercises |
In class, with the teachers supervision and aid, the students will implement and verify methods covered in lectures. As a homework assignment, each student will be given samples of a signal, and their assignment will be to individually estimate and analyse the spectrum of the sampled signal. |
