An analog optoelectronic broadband independent component analysis smart antenna system

The paper discusses optoelectronic processing for broadband microwave signals. Blind signal separation (BSS) refers to the process of determining the original signals by measuring linear combinations of those signals, where the coefficients of the linear combinations are unknown. One solution to the BSS problem is independent component analysis (ICA). A circuit performing ICA needs to have a nonlinearity that produces higher order signal powers and performs independence estimation for very high bandwidth signals. Our approach is to use the high bandwidth and dynamic holography offered by nonlinear optoelectronics. We have demonstrated a principal component analysis (PCA) smart antenna with the same technology (Anderson, D. et al., IEEE Trans. Antennas and Propag., p.607-17, 2002). PCA requires second order statistics, and the ICA analog processor requires a different approach due to the required nonlinearity for 4th higher-order independence determination. We apply the demonstrated system to multiple signals modulated onto multiple radio-frequency carriers that are incident on an antenna array from unknown directions. The presented system can process signals of very high bandwidths (hundreds of MHz), well above the current state of the art using digital or analog electronic circuits.