Handbook of Blind Source Separation

Chapter 14

Nonlinear mixtures

C. Jutten, M. Babaie-Zadeh and J. Karhunen

14.1 Introduction

Blind source separation problems in linear mixtures have been intensively investigated and are now well understood; see for example [26,53,29]. Since 1985, many methods have been proposed [60,28,30,21,27,10,54], especially those based on independent component analysis (ICA)  [31] exploiting the assumption of statistical independence of the source signals. These methods have been applied in various domains, and many examples can be found in application; see Chapters 16–19 of this book.

The linear mixing model, either without memory (instantaneous mixtures) or with memory (convolutive mixtures), is an approximated model, which is valid provided that the nonlinearities ...