Adaptive ICA for separation of complex signals with known source distributions in time-varying channels

A novel implementation of the complex block adaptive independent component analysis algorithm is presented. The new realisation is employed for the separation of complex signals with known source distributions. In practice, such situations are commonly encountered. Computer simulations show that the new algorithm converges much faster than the already existing complex fast-ICA in time-varying channels. In addition, the fast convergence of the novel technique is maintained with variation in the processing block size and is computationally cost effective.