声源定位中的时延估计技术

Time Delay Estimation Techniques in Source Location

时延估计(TDE)是阵列信号处理中的一项关键技术,其目的是要估计出同源信号到达不同传感器时,由于传输距离不同而引起的时间差。现有的TDE算法主要包括经典的广义互相关(GCC)方法、自适应最小均方(LMS)方法、基于子空间的特征值分解(EVD)方法和基于传递函数比(ATF-s ratio)的方法等。这些算法因其抗噪和抗混响性能不同,有着各自的应用场合。本文首先对现有的各种TDE算法进行分类论述,进而通过综合比较揭示了它们各自的优缺点,最后给出了进一步的研究方向。

Time delay estimation （TDE） is one of the key techniques in the array signal processing. Since microphones in array are usually placed at different places, TDE is used to derive the time difference of signal propagation from the source to the spatially separated micro- phones. The popular TDE algorithms mainly include the classical generalized cross correlation （GCC）, the adaptive least mean square （LMS）, the subspace based eigenvalue decomposition （EVD）and the acoustic transfer functions ratio （ATF-s ratio） methods, etc. For these TDE techniques, different approaches show different anti-reverberation and anti-noise capability, thus using for diverse acoustic scenarios. This paper gives the state of the art TDE techniques. By comparison of the referred algorithms, distinct features of each method , including advantages and disadvantages, are revealed. Finally the future researches on TDE techniques are pointed out.