基于到达时间差的声源定位算法与实现

Sound source localization algorithm and its implementation based on time difference of arrival

基于到达时间差的声源定位方法由时延估计和定位估计两部分构成,第一步的误差会带入到第二步从而造成累积误差,影响最终定位精度.针对这一问题,本文提出了一种改进的声源定位方法,采用累积互功率谱相位时延估计,并在此基础上结合球形插值算法进行定位.为了验证该算法在实际声场环境下的可行性和有效性,构建了三维麦克风阵列声源定位系统.实验结果表明,本文所提出的方法可以有效减少误差,提高最终定位精度.

Sound source localization based on the time difference of arrival（TDOA） consists of two parts-time delay estimation and location estimation. These two parts will cause cumulative errors and affect the final accuracy of localization.Aiming at this problem, this paper proposes an improved accumulated multi-frame cross power spectrum phase（CSP） time delay estimation algorithm, and on this basis, it improves the spherical interpolation（SI） algorithm to derive the source location. In order to verify the effectiveness of the algorithm, a three-dimensional sound source localization system is established under the practical conditions. This result of experiments shows that the proposed method can effectively reduce the errors and improve the positioning accuracy.