论文部分内容阅读
为了减少传统的截断混合矩阵求逆(ITMM)算法在个别时频点会丢失数据或者产生噪声信号的概率,提出了一种基于频域的2步欠定瞬时盲分离算法.由于现实中存在大量软稀疏(稀疏度不是很大)混合信号,将分离过程分解为ITMM和矩阵补偿2个步骤.首先估计出混合矩阵和利用经典的ITMM算法对混合信号进行初步恢复,然后对初步估计的信号时频矩阵进行矩阵补偿处理,从而达到修补丢失数据和去除多余数据(去噪)的效果.实验仿真验证了所提出的2步分离法相对于传统的ITMM算法能够得到更好的分离效果.此外,对算法的时耗问题进行了研究,相对于传统的ITMM算法,所提算法的时耗增加不到四分之一,却能够得到更好的分离效果.
In order to reduce the probability of ITMM algorithm losing data or generating noisy signals at individual time-frequency points, a two-step underdetermined instantaneous blind separation algorithm based on frequency domain is proposed in this paper. Soft sparse (sparse degree is not very large) mixed signal, the separation process is decomposed into two steps of ITMM and matrix compensation.First, the hybrid matrix is estimated and the original ITMM algorithm is used to preliminarily recover the mixed signal, and then the initial estimated signal Frequency matrix matrix compensation processing, so as to repair the missing data and remove redundant data (denoising) effect.Experimental simulation shows that the proposed two-step separation method compared to the traditional ITMM algorithm can get a better separation.In addition, the Compared with the traditional ITMM algorithm, the proposed algorithm increases the time consumption by less than a quarter, but it can get a better separation effect.