Developing a strong working knowledge on signal processing algorithms for modeling discrete-time signals, designing optimum digital filters, estimating the power spectrum of a random signal, and designing and implementing linear and nonlinear adaptive filters. Ability to implement the studied algorithms in Matlab standalone and hardware-oriented applications.
Prerequisiti
Basic concepts in analog signal processing, spectral analysis and filtering.
Metodi didattici
The course is based on lectures, case studies, and project examples, aimed at describing applications of statistical digital signal processing to practical utility projects. Lectures (hours/year in lecture theatre): 45
Verifica Apprendimento
The final exam is an oral test devoted to score the student knowledge by means of four/five questions, starting from a candidate-selected topic, and covering most of the course program.
As an alternative to the oral examination, a project may be presented, the topic of which is to be defined together with the course instructor, which may include: analysis and understanding of the literature references, description of the architectures of the system under study, parts of the project possibly implemented in Matlab or through other tools used in class.
The minimum score to pass the exam is 18, the top score is 30 cum laude.
Testi
Monson H. Hayes: Statistical Digital Signal Processing and Modeling. John Wiley & Sons Inc.
Simon Haykin: Adaptive Filter Theory, Pearson.
Simon Haykin: Neural Networks and Learning Machines, Pearson.
Contenuti
Introduction to discrete-time signal theory.
Discrete time signals, sampling theorem, linear shift invariant digital systems.
Analysis of digital systems in the Fourier and Z transform domains.
Discrete-time random processes.
Digital filtering of deterministic and stochastic signals.
Deterministic and stochastic signal modeling, Spectrum estimation.
Wiener Filter: linear prediction, white noise filtering, unwanted signal canceling.
Linear and Nonlinear Adaptive filtering: LMS, RLS and Kalman algorithms, neural networks.
Application examples in Matlab and programmable hardware platforms: Most of the course topics will be applied to an example of a complete digital wireless transceiver.