基于M估计变步长自适应仿射投影算法的稳健时延估计

Robust Time Delay Estimation Using Variable Step-Size Affine Projection Algorithm with M-estimators

该文提出了一种基于M估计变步长自适应仿射投影方法的稳健时延估计(TDE)算法。该算法将自适应仿射投影算法应用于时延估计,无须事先假定信号和噪声的统计特性,自适应调整自身参数;应用稳健M估计理论,抵消重尾噪声干扰。数值仿真表明,在高斯噪声、非高斯噪声甚至冲激噪声的干扰下,该文算法比高阶统计量法和最小均方自适应法有更强的稳健性和更高的估计精度。

Robust time delay estimation is proposed using variable step-size affine projection algorithm with M-estimators. The novel algorithm, using the structure of adaptive filter and M-estimators, is free of the assumption that the statistical characteristics of singal and noise is known. Simulation experiments under three different noise environments, Gaussian, non-Gaussian and impulse noise, are conducted to compare with the proposed TDE method using the existing Higher-Order Spectra（HOS） and Least Mean Square（LMS）, and the result showes that the proposed TDE method is more robust and precision.