Separation of independent sources from correlated inputs

Abstract

The characterization of independent stationary stochastic components (sources), is generally achieved by using the spectral matrix of partially correlated measurements, which are linearly related to the components of interest. In the general case where no assumptions are made concerning the way the sources are mixed on the measurements, the spectral matrix is not able to extract the true sources. While spectral analysis only uses second-order properties of independent stochastic sources, a procedure based on higher-order analysis (fourth-order cross cumulants) is developed. This approach leads to a complete identification of the sources.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>