A scheme for source separation of ground borne vibrations

Abstract

The source separation of low level ground borne vibration signals is a tricky problem. This can be viewed as a complex version of the classical source separation problem, the `cocktail party problem'. An extension of the basic Independent Component Analysis (ICA), a Blind Source Separation method is proposed in this particular case, as the damping of the ground borne vibration signals is significant. The separation of a set of signals from a set of observed signal mixtures, without the information about the source signals or the mixing process is known as Blind Source Separation (BSS). The performance of the FastICA algorithm for various sources - sensor distance is analysed in this paper.