基于最小均方误差的时延估计方法研究

Research on the Time Delay Estimation Method Based on the Least Mean Square Error

文章对最小均方误差时延估计方法做了详细的讨论,研究经典LMSTDE算法和约束类ETDE算法和ETDGE算法的原理,并从非相关背景噪声和相关背景噪声两方面入手,来分析静态时延下的估计性能。在非相关背景噪声下LMSTDE算法、ETDE算法和ETDGE算法均有着良好的估计性能;在相关背景噪声下,传统LMSTDE算法不能准确估计时延,而ETDE算法和ETDGE算法依然具有时延估计的能力。

This paper discusses ＂Least Mean Square Error TDE Method＂ in detail, studies the theories of classical algorithm, LMSTDE, and two constrained algorithms, ETDE and ETDGE. The paper also analyzes the static delay capability of the three algorithms in two aspects which are uncorrelated background noise and correlated background noise. In the uncorrelated background noise, the algorithms of LMSTDE, ETDE and ETDGE all have good property in estimating delay. And in correlated background noise, the traditional LMSTDE algorithm can not estimate time delay exactly, while ETDE and ETDGE algorithms can still estimate delay accurately.