自适应回声消除算法的仿真研究  

Simulation and Research on Adaptive Echo Cancellation Algorithm

作　　者：乐彦杰[1] 陈隆道[1] 祁才君[1] 

机构地区：[1]浙江大学电气工程学院,浙江杭州310027

出　　处：《电声技术》2012年第2期67-71,74,共6页Audio Engineering

摘　　要：介绍几种自适应回声消除算法:NLMS,PNLMS,PNLMSDU,IPNLMS,MPNLMS,并在Matlab环境下进行了仿真比较。采用高斯噪声作为输入,分别在回声路径稀疏、回声路径中等稀疏、回声路径完全非稀疏的条件下进行仿真。仿真结果表明,PNLMS算法是针对稀疏回声路径的有效快速算法,MPNLMS算法做了进一步的优化。IPN-LMS算法和PNLMSDU算法融合了NLMS算法和PNLMS算法的优点,既保证有较快的初始收敛速度,又保证后阶段收敛速度不明显下降。Several adaptive echo cancellation algorithms including NLMS, PNLMS, PNLMSAU, PNLMSDU, IPNLMS and MPNLMS are introduced. Some simulations are done under Matlab to compare the performance of these algorithms. White Gaussian noise with zero mean and unit variance are used as the input and simulations are finished under the condition of sparse impulse responses quasi sparse impulse response and dispersive impulse response. According to the result of simula- tions,PNLMS is the effective and fast algorithm which is applied for the sparse echo path and MPNLMSdo further optimiza- tion in the basis of PNLMS. IPNLMS and PNLMSDU which contains respective advantages of NLMS and PNLMS, ensure the fast initial convergence speed and simultaneously make it not slow down dramatically after the initial period.

关 键 词：回声消除 NLMS PNLMS PNLMSDU IPNLMS MPNLMS 

分 类 号：TN911[电子电信—通信与信息系统]