基于EMD的单通道通信信号盲分离算法研究

Research on Blind Source Separation Algorithm of Single Channel Communication Signal Based on EMD

针对单通道盲信号处理领域先验知识不足的多信号分离问题,提出了一种基于交叉验证技术、独立分量分析和经验模式分解的单通道通信信号盲分离算法解决方法。该方法对单通道观测信号进行经验模式分解得到一系列固有模态函数和余量,利用交叉验证技术的思想确定出单通道源信号的数目,再根据源信号数目选取合适的固有模态分量组合成多通道信号,然后利用独立分量分析的Fast-ICA算法实现通信信号的盲分离。最后利用2ASK和BPSK两路混合信号进行实验验证。结果表明,该方法可以有效地估计单通道通信信号的源数目并实现分离。

In the light of the multi signal separation problem of prior knowledge deficiency in single channelblind source separation field,a novel method based on Empirical Mode Decomposition（EMD）,the cross-val-idation technique and Independent Component Analysis（ICA） is proposed. Firstly the EMD is applied tosingle-channel observation signal to obtain a series of intrinsic mode function and residue. The number ofsource signals in a single channel is acquired by using cross-validation techniques,and a new multi-channelsignal is reconstructed according to the number of source signals. Then blind separation of communicationsignal is realized by using independent component analysis of Fast-ICA algorithm. Finally,a mixed signalby using 2ASK and BPSK is adopted to validate the algorithm. The experimental results show that the pro-posed method is effective and feasible to estimate the sources number of single-channel communication sig-nal and to realize the separation.