α-稳定分布噪声环境下的非线性回声消除研究

Research on Nonlinear Echo Cancellation in the α-Stable Noise Environment

针对广泛存在的非线性回声,以及在非高斯噪声环境下,传统回声消除器中自适应算法性能衰退,继而导致回声消除效果下降的情况,本文提出了一种基于最小离差准则的协同函数链接型自适应滤波回声消除方法.该方法使用归一化最小l_p范数算法更新线性和非线性函数扩展的自适应滤波器的权值,并将输出信号协同组合,以消除非线性回声.仿真实验结果表明,该方法在α-稳定分布噪声且非线性回声存在的情况下比传统回声消除方法具有更高的回声衰减增益.

In the condition of the widespread nonlinear echo and the non-Gaussian noise environment,the performance of the adaptive algorithm in the traditional echo canceller is degraded,which leads to the decline of the echo cancellation effect.In this paper,aiming at this question,a collaborative functional link adaptive filtering echo cancellation method based on the minimum dispersion criterion is proposed.In this method,the weights of the linear and nonlinear functional expanded adaptive filters are updated by using a normalized minimum l_p norm algorithm,and the output signals are cooperatively combined to eliminate the nonlinear echo.The simulation results show that the proposed method can achieve higher echo return loss enhancement in the presence of the α-stable distributed noise and the nonlinear echo.