A new type of self-adaptive blind signal separation algorithm based on feed-forward and feedback neural network

Making use of the self learning ability of neural networks to realize blind signal separation has been proven as an efficient method for signal separation. A different neural network model can produce distinct algorithm efficiency. Based on the feedforward and feedback neural network model, this paper develops a self-adaptive blind source separation algorithm and applies it to signal separation. The performance of the proposed algorithm is illustrated by theoretical analysis and computer simulation experiments.