一种基于JADE与小波去噪的自适应盲分离方法

Adaptive Blind Signal Separation Algorithm Based on JADE and Wavelet De-Noising

针对传统盲分离算法忽略噪声影响的问题,引入小波去噪,在JADE算法的基础上,研究了前、后去噪盲分离算法与JADE算法在不同输入信噪比条件下的分离性能。仿真表明,输入信噪小于约-5dB时,后去噪盲分离算法输出信噪比最佳;大于约25dB时J,ADE输出信噪比最佳;其它情况下,前去噪分离算法性能最佳。在此研究基础上,给出一种可估计信噪比的带噪信号自适应盲分离方法,实现对信号信噪比动态变化情形下目标信号的有效分离。

To solve the problem that noise is ignored in the classical blind signal separation algorithm, this paper introduces wavelet de-noising and studies the de-noising performance of preceding/ posterior blind signal separation algorithms and JADE algorithm under different input-SNR condi- tions. Simulation shows that the blind signal separation algorithm with posterior de-noising is the best when the input-SNR is less than about -5 dB, the JADE algorithm is the best when the input-SNR is greater than about 25 dB and the blind signal separation algorithm with preceding de-noising is the best under other conditions. Based on those results, an adaptive blind signal seperation algorithm with SNR estimation is proposed to seperate the target signal effectively under different input-SNR conditions.