Unified formulation of closed-form estimators for blind source separation in complex instantaneous linear mixtures

The blind separation of unknown independent source signals from sensor observations is addressed in environments of instantaneous linear mixtures. After pre-whitening the sensor data, the source extraction reduces to the identification of certain parameters defining a unitary transformation. This contribution develops the algebraic devices which allow us: 1) to provide a unified formulation of the closed-form estimators of the separation parameters in the complex-mixture scenario — typical of multi-user digital communications —, and 2) to disclose the remarkable parallelism existing between the real and the complex problem in the context of their analytic solutions.