Coursebooks

Nonlinear signal modeling and prediction

EE-714

Lecturer(s) :

Vesin Jean-Marc

Language:

English

Frequency

Every 2 years

Remarque

Every two years. Next time : Spring 2020

Summary

The literature on nonlinear signal processing has exploded, and it becomes more and more difficult to identify the most useful approaches for specific contexts. This course presents promising developments for the practical application of nonlinear signal models in various fields of engineering.

Content

1. Introduction
2. Summary of linear AR and ARMA modeling
3. Nonlinear AR and ARMA modeling, polynomial models and their estimation
4. Specific nonlinear models (threshold AR, ...)
5. Neural network based modeling and prediction
6. Model selection
7. Chaos theory and applications
8. Kernel-based approaches
9. Laboratory exercises: application of nonlinear modeling/prediction to synthetic and experimental data

Keywords

Signal modeling, Signal prediction, Nonlinear autoregression, Parameter estimation.

Learning Prerequisites

Recommended courses

Statistical signal processing

Assessment methods

Multiple.