摘要
基于信号的白化变换及矩阵的QR分解理论，提出了一种新的盲分离算法。该算法将对混叠矩阵的辨识转化成对一系列Givens矩阵的辨识，从而减少了对n2-c2n个未知矩阵元素的辨识，降低了其计算量。另外，在对未知元素进行辨识时，采用了改进的罚函数算法，这样又大大提高了算法的收敛性能。仿真结果表明，该算法不仅缩短了分离时间，而且有很好的分离效果。
A new blind separation algorithm based on signals' whiten transform and QR decomposition of matrix is proposed. By which mixture matrix is changed into a series of Givens matrix, so that the recognition of matrix element is reduced by n2-c2n , the computation is decreased. On the other hand, the recognition of unknown elements is done by penalty function algorithm, the convergence speed is improved greatly. Simulation results show that the separation time is reduced and the separation effect is very good.
出处
《计算机工程》
CAS
CSCD
北大核心
2003年第17期55-57,共3页
Computer Engineering
基金
国家自然科学基金项目(60274006)
广东省自然科学基金项目(990892)
教育部重点科研基金项目(02152)
广东省优秀人才基金资
关键词
罚函数
Givens变换
盲分离
Penalty function
Givens transform
Blind separation
