Adaptive Filtering

The course is a substitute for:

Adaptive Filtering (X31ADA)

Lecturer:

Tutor:

Supervisor:

Department of Circuit Theory

Synopsis:

The aim of this subject is to give the overview of adaptive filtering methods and their applications to speech and biological signal processing, noise cancellation, signal separation and signal analysis. Seminars are devoted to the implementation of algorithms in MATLAB. It is highly recommended to be acknowledged with the bases of digital signal processing given in subjects UCZ, CFS, PSP or in similar ones.

Requirements:

Syllabus of lectures:

1. Batch and recursive signal analysis, basic terms

2. Batch parametric methods, least squares

3. Recursive parametric methods, accuracy of estimation

4. Memory shaping, RLS and LMS algorithms, properties, prediction

5. Estimation and signal modelling, RLS/LMS FIR estimator

6. Adaptive line enhancement, conditions, application to speech

7. Equalisation of signal path, sign LMS algorithm

8. Analysis of LMS algorithm, error surface, stability, normalised LMS

9. LMS in non-stationary conditions, gradient noise, convergence control

10. Narrow-band signal separation, LMS FIR/CIIR, resonator

11. Noise compensation, conditions of proper function

12. Symmetric adaptive structures (SAD) for sources separation

13. Hyperstable algorithms, LMS IIR, echo cancellation

14. Frequency domain adaptive filters, lattice filters

Syllabus of tutorials:

1. Introduction to MATLAB, basic sequences generation

2. Short-time spectral analysis, spectrograms

3. Parametric methods, basic functions

4. Signal analysis using parametric methods

5. Synthesis of AR non-stationary processes and their analysis

6. Implementation of LMS estimator

7. Implementation of LMS predictor, noise suppression

8. Application of LMS algorithm for equalisation

9. Comparison of LMS, SLMS, NLMS and RLS algorithms

10. Frequency analysis with application to speech formants

11. Implementation of noise cancellation methods

12. Beamforming and applications

13. Blind separation using symmetric adaptive decorrelator

14. Projects presentation, credit

Study Objective:

Study materials:

1. Widrow, B., Stearns, S.D.: Adaptive Signal Processing. Prentice-Hall, Inc., New Jersey, 1985

2. Young, P.: Recursive Estimation and Time-Serie Analysis, Springer-Verlag, New York, 1984

3. Haykin, S.: Adaptive Filter Theory. Prentice-Hall, Inc., New Jersey, 1991

Note:

Further information:

No time-table has been prepared for this course

The course is a part of the following study plans:

• Biomedicínské inženýrství - inženýrský blok (elective specialized course)
• Biomedicínské inženýrství - inženýrský blok (elective specialized course)