同频同调制通信信号的卷积混合盲源分离

Convolutive blind source separation applied to the communication signals that with same carrier frequencies and modulation

针对卷积混合同频同调制通信信号,提出一种新的卷积混合盲分离算法,利用滑窗Z变换将时域卷积混合形式信号通过数学变换转换到Z域瞬时混合形式,通过成熟的线性混合盲分离算法分离出Z域源信号,估计出时域上的源信号。该方法适用于适定情形下的卷积混合盲分离问题,需要设置的分离参数只有一个,就是滑动窗口长度,从而降低了计算复杂度,提升了分离性能,易于硬件实现,可移植性强。

Against the convolutive mixture communication signals with same carrier frequencies and modulation,a new algorithm for convolutive blind source separation was presented.This algorithm transformed the time domain convolutive mixture into the Z domain instantaneous linear mixture by using sliding window Z transform and then separated the Z domain source signals by using the blind source separation algorithm for instantaneous linear mixture.The time domain source signals were estimated.The method is suitable for the cases of well-posed model of convolutive blind source separation,for there is only one parameter to set up,that is,the length of sliding window.It thus reduces the computational complexity and improves separation performance,with strong portability and easy implementation of hardware.