Code: BE2M31DSP Advanced DSP methods
Lecturer: prof. Ing. Pavel Sovka CSc. Weekly load: 2p+2c Assessment: Z,ZK
Department: 13131 Credits: 6 Semester: S
Description:
The course follows the basic course in signal processing and introduces advanced methods of analysis and digital signal processing. Graduates will learn the methods of digital signals analysis and be able to practically use them. They learn to know the conditions of use of correlation, spectral and coherent analysis of random signals. They will became familiar with methods of signal decomposition and independent component analysis and the time-frequency transformations. Emphasis will be placed on an ability to interpret the results of signal analyses.
Contents:
1. Modeling and representation of linear systems in time-, correlation- and spectral-domain
2. Measurement of the delay using correlation and spectral analysis
3. Coherence, partial coherence and their use
4. Cepstral analysis and its use for signal deconvolution
5. Spectral and cepstral distances and their use
6. Methods of additive and convolution noise reduction and signal restoration
7. Methods of 1-D signal interpolation
8. Principal component analysis and its use for lossy compression of signals
9. Principles of methods of blind source separation
10. Principles of methods of blind signal deconvolution
11. Implementation of the discrete wavelet transform using filter bank, quadrature filters
12. Granger causality and Hilbert-Huang transform
13. Robust estimates of characteristics of random signals
14. Reserve
Recommended literature:
Saeed V. Vaseghi: Advanced Digital Signal Processing and Noise Reduction, Wiley,2009, ISBN: 978-0-470-75406-1
Monson Hayes: Statistical digital signal processing and modeling. Wiley, 1999, ISBN: 978-0-471-59431-4.
Keywords:
Correlation and coherence function, cepstrum analysis, noise reduction, principal component analysis, blind source separation, discrete wavelet transform, Hilbert-Huang transform, Granger causality