The course reviews classical and modern methods and algorithms for computer-based spectral analysis of signals. Also, it gives an overview of various applications in communications, systems engineering, radar, and biomedicine. After the course, the student will:
understand the spectral estimation problem and the meaning of spectrum
understand the differences between non-parametric and parametric approaches to the spectral estimation problem
master several non-parametric methods for spectral estimation, both periodogram and correlogram-based methods as well as data adaptive filter-bank methods, and be able to use this knowledge to solve real-world problems
master several parametric methods for estimation of line spectra as well as rational spectra, and be able to use this knowledge to solve real-world problems
be able to decide what methods ( for example, parametric or non-parametric) are suitable for a specific problem
be able to solve spectral estimation problems and to visualise their solutions using the MATLAB software
be prepared to use the tools of spectral analysis of signals for solving practical problems in a diversity of areas such as control system modelling, wireless communication systems, radar and sonar signal processing, genomic data mining, image processing, magnetic resonance spectroscopy, acoustic imaging, biomedical signal processing, economic, geophysical, astronomic (etc.) data processing, Internet traffic analysis, and so forth.
Basic definitions and overview of the spectral estimation problem. The periodogram and correlogram methods. Improved methods based on the periodogram. Filter-bank methods. Parametric methods for rational spectra and for line spectra. Review of selected applications.
Lectures and computer-based exercises
Graded homework assignments and passed computer exercises.
week 27, 2017
Spectral analysis of signals
Upper Saddle River, N.J.:
Pearson Prentice Hall,