Blind separation of sources using density estimation and simulated annealing

This paper presents a new adaptive blind separation of sources (BSS) method for linear and non-linear mixtures. The sources are assumed to be statisti cally independent with non-uniform and symmetrical PDF. The algorithm is ba sed on both simulated annealing and density estimation methods using a neur al network. Considering the properties of the vectorial spaces of sources a nd mixtures, and using some linearization in the mixture space, the new met hod is derived. Finally, the main characteristics of the method are simplic ity and the fast convergence experimentally validated by the separation of many kinds of signals, such as speech or biomedical data.