DOA estimation for automatic steering of a smart antenna radiation pattern using the conventional ICA-based blind source separation method

In this paper, we propose a new method based on the blind source separation concept to estimate the direction of arrival (DOA) for a smart antenna system and then steer its radiation pattern toward the desired radio-frequency emitter. Indeed, we generate a set of observations according to a mathematical model of smart antennas, and we apply various separation methods, such as ICA based ones to perform the extraction of the independent components and the mixing matrix. The tested algorithms are classified according to their performance index of separability. When the best separation is achieved, the extracted independent components are characterized using their power spectral densities which will be used to calculate the beamforming vectors.