On the application of complex ICA to Radar data in the frequency domain

Radar scattering center separation by means of the statistical properties of the data allows an algorithm to adapt to the scattering environment. Through the assumption of non-Gaussianity for scattering centers of interest it is feasible to employ a source separation technique base on finding the most non-Gaussian features within the data. A well known technique to accomplish extraction of components with this statistical property is Independent Component Analysis (ICA). This work explores the use of ICA in extracting scattering center components from complex frequency domain Radar data.