Blind Separation for Cabin Acoustic Signals in Complex Environment

Abstract

Separation of mixed acoustic signals is important both in theory and practice. In complex cabin environment, it's hard to get the source signals exactly from the mixed acoustic signals. Blind source separation (BSS) method can be used in the separation of cabin acoustic signals. The basic concepts of blind source separation were introduced. Three typical algorithms of blind source separation were compared and analyzed, based on the characteristics of complex cabin acoustic signals. An effective blind source separation method was explored on the separation of complex cabin acoustic signals. The identification and separation of source signals in cabin was realized by the method. The practical analyses results show that, the FastICA algorithm can effectively separate a variety of complex cabin acoustic signals. This method provides a parameter foundation for the further research about reducing the cabin noise, improving the cabin environment and realizing the equipment state monitoring and fault diagnosis.