COURSE UNIT DESCRIPTION
Transkrypt
COURSE UNIT DESCRIPTION
Poznan University of Technology European Credit Transfer System Faculty of Electronics and Telecommunications Title Code POZ04WTS2ICC15 Advanced Signal Processing Algorithms Field Year / Semester Electronics and Telecommunications 1 / spring Specialty Course Hours Number of credits core Lectures: 2 Classes: 2 Laboratory: - Projects / seminars: - 5 Lecturer: prof. dr hab. inż. Ryszard Stasiński Katedra Systemów Telekomunikacyjnych i Optoelektroniki tel. +48 61 665 3839, fax. +48 61 665 3830 e-mail: [email protected] Faculty: Faculty of Electronics and Telecommunications ul. Piotrowo 3A 60-965 Poznań tel. (061) 665-2293, fax. (061) 665-2572 e-mail: [email protected] Status of the course in the study program: Compulsory course on Electronics and Telecommunications studies, specialization Information and Communication Technologies Objectives of the course: Advanced topics in digital signal processing: idea of linear prediction/Wiener filtering and its applications, and multirate systems. Course description: The theory of linear prediction and Wiener filtering – modeling of stationary stochastic processes (ARMA, AR, MA), Wiener-Hopf equations and fast algorithms for solving them (Levinson-Durbin, Schur), lattice filters, Wiener filtering – FIR and IIR case. Adaptive filters – their connection with Wiener filters, gradient algorithms, in particular dynamic behaviour of LMS algorithm, least squares (LS) error formulation, recursive LS (RLS) algorithms – Kalman, square-root, fast RLS, and their lattice-ladder versions, comparison of gradient and RLS algorithms. Parametric spectrum estimation – its connection with linear prediction, Yule-Walker, Burg, and unconstrained LS methods, eigenanalysis approach to spectrum estimation (Pisarenko, MUSIC, ESPRIT). Multirate systems – sampling rate conversion (decimators, interpolators, approximate methods), multiplierless modulation and demodulation, polyphase filter structures, filter banks (general, uniform, critically sampled, perfectly reconstructing), QMF filters, and their use for implementation of discrete Wavelet transform. Introduction to spatiotemporal analysis – short-time Fourier transform, Gabor and wavelet transforms. Initial knowledge: Analog and digital signal processing theory. Teaching methods: Lectures supported by multimedia presentations. Assessment methods: Exam ending the lecture. Bibliography: 1. 2. Digital Signal Processing, J.G. Proakis, D.G. Manolakis, Pearson – Prentice-Hall (there are several editions of this book). Discrete-Time Signal Processing, A.V. Openheim, R.W. Schafer, Prentice-Hall, e.g. 1999. page 1 of 1