摘要
独立向量分析(IVA)是解决频域卷积盲分离排序模糊性最好的方法之一,但存在迭代次数较多、运算时间较长、分离效果易受分离矩阵初值影响的局限性。该文提出一种基于步长自适应的IVA卷积盲分离算法,该算法使用特征矩阵联合近似对角化(JADE)算法对分离矩阵进行初始化,并对步长参数进行了自适应优化。JADE初始化能够使分离矩阵具有合理的初值,避免局部收敛的情况;步长的自适应优化能够显著提升算法的收敛速度。仿真结果表明,该算法进一步提升了分离性能,并显著缩短了运算时间。
Independent Vector Analysis （IVA） is one of the best methods to solve the sort ambiguity of convolutive blind separation in frequency domain. However, it needs more iterations and computing time, and the separation effect is susceptible to the initial value of the separation matrix. This paper proposes an IVA convolutive blind separation algorithm based on step-size adaptive, which uses Joint Approximative Diagonalization of Eigenmatrices （JADE） algorithm to initialize the separation matrix and optimizes adaptively the step step-size parameters. JADE initialization can make the separation matrix have an appropriate initial value, thus avoiding the situation of loca! convergence; step-size adaptive optimization can significantly improve the convergence speed of the algorithm. Simulation results show that this algorithm improves the separation performance and shortens the operation time significantly.
作者
付卫红
张琮
FU Weihong;ZHANG Cong(School of Telecommunication Engineering,Xidian University,Xi＇an 710071,China;Collaborative Innovation Center of Information Sensing and Understanding,Xi＇ an 710075,China)
出处
《电子与信息学报》
EI
CSCD
北大核心
2018年第9期2158-2164,共7页
Journal of Electronics & Information Technology
基金
国家自然科学基金(61201134)~~
关键词
盲源分离
卷积混合
独立向量分析
特征矩阵联合近似对角化
步长自适应
Blind Source Separation （BSS）
Convolutive mixture
Independent Vector Analysis （IVA）
Joint Approximative Diagonalization of Eigenmatrices （JADE）
Step-size adaptive
