基于仿射投影--独立分量分析的盲源分离  

Blind source separation based on affine projection and independent component analysis

作　　者：李雄杰[1,2] 周东华[2] 

机构地区：[1]浙江工商职业技术学院电子信息工程系,浙江宁波315012 [2]清华大学自动化系,北京100084

出　　处：《计算机应用研究》2017年第6期1721-1725,共5页Application Research of Computers

基　　金：国家自然科学基金资助项目(61210012)

摘　　要：仿射投影算法(APA)重复利用数据,可提高算法的收敛速度。针对现有盲源分离(BSS)收敛慢问题,以BSS的独立分量分析(ICA)为基础,结合APA思想,设计出BSS的APA-ME、APA-MMI、APA-EASI新算法。在这些新算法中,输出向量数据被重复利用,向量式数据转变成矩阵式数据,从而加快了BSS的收敛速度。仿真结果表明,APA-ICA类的BSS算法是有效的。Using data in a repeated mode can improve the convergence speed of the affine projection algorithm （APA）. Ai- ming at the problem of slow convergence in the existing blind source separation （BSS） , based on the independent component analysis （ICA） for BSS, this paper designed new APA-ME, APA-MMI and APA-EASI algorithms for BSS by using the idea of APA. In these new algorithms, output vector data was utilized in a repeated fashion, and the vector data was thus converted into matrix data. The convergence rate of BSS was accelerated. Simulation results show that the effectiveness and applicability of APA-ICA class BSS algorithm.

关 键 词：盲源分离 仿射投影算法 独立分量分析 最大熵 最小互信息 

分 类 号：TN911.7[电子电信—通信与信息系统]