LMS与RLS算法仿真消噪对比研究  

Simulation De-noising Comparison of LMS and RLS Algorithm

作　　者：田玉静[1] 左红伟[2] 朱周华[3] 

机构地区：[1]青岛理工大学现代教育技术中心,山东青岛266033 [2]青岛理工大学土木工程学院,山东青岛266033 [3]西安科技大学通信工程学院,陕西西安710054

出　　处：《通信技术》2009年第12期161-163,共3页Communications Technology

摘　　要：讨论了RLS(递归最小二乘)和LMS(最小均方)自适应滤波算法及原理,对两种算法进行了系统全面的分析,对比研究了各自的优势及不足,提出了两种算法在语音消噪仿真中的算法实现,对实际语音信号进行了仿真消噪,研究表明选用算法对语音消噪是明显有效的,RLS自适应消噪算法及LMS自适应噪声抵消算法具有很强的实际应用价值。The algorithms and principles of RLS（recursive least-squares） and LMS（least mean squares） adaptive filter are studied. The two de-noising methods are analyzed systematically and comprehensively, and comparisons of their advantages and disadvantages are presented. The simulation algorithms based on the speech de-noising is given, and the simulationon de-noising of actual speech signal with white noise is made. The research result shows that the RLS and LMS algorithms for speech de-noising are obviously effective, and the RLS and LMS adaptive filter algorithms are of great practical value.

关 键 词：递归最小二乘算法 最小均方算法 语音消噪 自适应滤波 

分 类 号：TN713.7[电子电信—电路与系统]