相关背景噪声下ETDE算法研究

An Explicit Time Delay Estimation(ETDE) Algorithm for Correlated Noise Environment

相关背景噪声下的时延估计问题是现阶段的一个研究热点。在时延估计的众多算法中,属于“约束”类算法的ETDE(Exp licit T im e D elay E stim ation)算法具有适用性广、可靠性高、计算量小和实时性好等优点。文中首次将ETDE算法扩展到相关背景噪声领域。从理论上分析了在相关背景噪声下ETDE算法的估计效果,得出了时延估计的期望表达式。指出在此种情况下,ETDE算法得到的是时延的有偏估计,其偏差量与时延真值、信噪比以及背景噪声间互相关函数值的大小有关。通过计算机仿真验证了理论结果。

Time Delay Estimation （TDE） is an important research problem in array signal processing. Conventional TDE methods are always based on the assumption of uncorrelated background noise. Among them, the ETDE （explicit time delay estimation）algorithm, which belongs to a kind of “constrained” algorithms, has many merits such as wide applicability, high reliability, low computation cost and good real-time property. Our aim is and we believe we are the first to extend the ETDE algorithm to the case of correlated noise environment. When the signal-to-noise ratio and the cross correlation value of two channels of noise have been fixed, properly selecting the true value of time delay is essential to the convergence of ETDE. The expectation of time delay is derived under the assumption of convergence. The result validates that ETDE algorithm provides biased estimation at this time. Then the error function of time delay is derived. The bias is related to the true value of time delay, signal-to-noise ratio and the cross correlation value of two channels of noise. If the true value of time delay approaches 0, the error function can be simplified. In practice, people are mainly concerned with the decimal part of time delay. When the true value of time delay is between 0 and 1, the ETDE algorithm can always converge. So the research results of this paper have strong practical meaning. At last, simulation results prove preliminarily that the new conclusions we found in our extension of ETDE algorithm to the case of correlated noise environment are indeed correct.