Developing a strong working knowledge on signal processing algorithms for modeling discrete-time signals, designing optimum digital filters, designing and implementing linear and nonlinear adaptive filters, application of estimation theory and probabilistic inference to communication systems and integrated sensing and communication systems. Ability to implement the studied algorithms in Matlab standalone and hardware-oriented applications.
Course Prerequisites
Basic concepts in analog signal processing, spectral analysis and filtering.
Teaching Methods
The course is based on lectures, case studies, and project examples, aimed at describing applications of statistical signal processing to practical utility projects. Lectures (hours/year in lecture theatre): 45
Assessment Methods
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. In both cases, the ability to comprehend the subject's fundamentals and to apply the learned techniques to real-world case studies will be assessed. The minimum score to pass the exam is 18, the top score is 30 cum laude.
Texts
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.
Contents
Introduction to Discrete-Time Signal Theory. Discrete-time signals, sampling theorem, linear time-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. Modeling of deterministic and stochastic signals. The Wiener filter: linear prediction, white noise filtering, and interference cancellation. Linear and nonlinear adaptive filtering. Estimation theory, Cramér–Rao Bound, Fisher Information, factor graphs, and probabilistic inference. Application examples in MATLAB and programmable hardware platforms: most of the course topics will be applied to the design of a complete digital wireless transceiver.
