基于高阶统计量的最小均方误差自适应回声消除算法(英文)

Least Mean Square Adaptive Echo Cancellation Algorithm Based on High-Order Statistics

自适应算法的原理是选取一个误差标准 ,通过迭代过程使误差渐近最小 ,从而得到期望的优化值 .基于二阶统计量的误差标准的自适应算法对附加噪声非常敏感 ,在有噪情况下估计值与期望值相差甚远 .利用高阶统计量的对噪声不敏感的特性对自适应算法加以改进 ,使得在有噪的情况下估计值与无噪情况下相同 .

Error criterion is essential in adaptive algorithm.We get expected optimal value through iterate process.Since adaptive algorithms based on second-order statistics are sensitive to additive noise,the estimated results under noisy environment are different from that under noiseless environment.In this paper we utilize the property that high-order statistics are insensitive to noise to get a novel adaptive echo cancellation algorithm,which can overcome the limitation.