Application Of Blind Source Separation In Separation Of Communication Signals

In signal processing, useful signals are often mixed with interfering signals. With knowing the information of source signals and transmission channel, we can carry on processing to the signals, filter out interfering signals and get useful signals. Unfortunately, they are not known in many situations, and the source signals can not be restored by the traditional methods. Blind source separation is put forward to solve the problem well.This paper describes the basic theory of blind source separation and natural gradient algorithm in detail. In the natural gradient algorithm, the convergence speed and stability error are two important indicators, but these two indicators are of inherent contradictions. Variable step-size natural gradient algorithm which enjoys faster convergence speed and smaller stability error is proposed to solve the shortcoming of natural gradient algorithm. However, the performance of variable step-size natural gradient algorithm is still less than ideal. In this paper, an improved variable step-size natural gradient algorithm is studied on the basis of traditional algorithm, and the new algorithm enjoys faster convergence speed and better experimental results than variable step-size natural gradient algorithm.To avoid slow separation speed and poor anti-noise performance, this paper describes the non-orthogonal joint diagonalization algorithm and studies the improved one on the basis of triangle parameter algorithm. The improved algorithm enjoys better anti-noise performance than the other non-orthogonal joint diagonalization algorithm and achieves the separation of communication signals at lower SNR.