The course focuses on Advanced Statistical Digital Signal Processing and covers the following topics:
- Review of basics & fundamentals (8h): linear algebra & factorizations (Tel); z trasf. & filters, random processes (Mat); multivariate Gaussian (All); constrained optimization (Tel).
- Block-processing (6h): regression, filtering, interpolation, direct & inverse problems
- Estimation theory and performance limits (24h): Min Var. & BLUE, LS, MLE for Gaussian distributions and hints on non-Gaussians, Cramer Rao Bound, numerical examples: regression, sinusoids & PLL, TOA, xIxO identification and deconvolution.
- Bayesian estimators (14h): a-posteriori estimation (MAP, MMSE and LMMSE); Wiener filter; linear prediction, Yule-Walker equations and Levinson recursion; EM method (2h).
- Adaptive MMSE filters (8h): iterative LMMSE, LMS, RLS methods; examples on adaptive identification, deconvolution, and MIMO systems.
- Bayesian tracking (8h): dynamic model and Kalman filter; examples on target positioning (e.g., GPS) & tracking.
- Spectral analysis (10h): periodogram; parametric methods (MA, AR, ARMA models); line spectra and high-resolution methods (HOYW, MUSIC).
- Array processing (6h): narrowband model definition, direction of arrivals (DOA), beamforming methods and multichannel systems.
- Pattern and sequence recognition (10h): supervised and unsupervised classification, classification of signals in noise, linear discriminant, clustering methods.
- Montecarlo simulation and numerical analysis
Course has been Adopted by Google: part of the laboratory activity is in cooperation with Google-teams that introduce the students to practical problems using Google-data. The framework Google adopts Advanced Digital Signal Processing course.