基于互相关序列和BP网络的声源定位算法  

Sound Source Localization Algorithm Based on Cross-correlation Sequence and BP Network

作　　者：曾庆宁[1] 苏盼 王红丽 龙超[1] ZENG Qing-ning;SU Pan;WANG Hong-li;LONG Chao(College of Information and Communication,Guilin University of Electronic Technology,Guilin 541004,China)

机构地区：[1]桂林电子科技大学信息与通信学院,桂林541004

出　　处：《科学技术与工程》2023年第16期6945-6951,共7页Science Technology and Engineering

基　　金：国家自然科学基金(61961009);广西省无线宽带通信与信号处理重点实验室基金(GXKL06200107);桂林电子科技大学研究生教育创新计划资助项目(2022YCXS042,2021YCXS028)

摘　　要：在基于麦克风阵列的声源定位算法中,一种常用算法的基本思路是通过麦克风接收到信号的相关序列来计算信号之间的时延,进而再根据阵列的结构确定声源的位置。在分析传统的声源定位算法基础上,针对双五元十字阵模型,介绍传统的基于广义互相关相位变换加权(generalized cross correlation-phase transform,GCC-PHAT)时延估计的定位算法,并给出基于GCC-PHAT时延估计和反向传播(back propagation,BP)神经网络的定位算法、基于抛物线互相关时延估计和BP网络的定位算法,进而通过分析影响时延估计的主要因素,提出了基于互相关序列和BP网络的新定位算法,该算法将GCC-PHAT互相关序列最大值点的位置、最大值点及其左右各一点的相关值作为BP网络的输入,通过对BP网络进行训练来实现声源的三维定位。仿真实验表明:与传统的基于GCC-PHAT时延估计的定位算法相比,所提出的各个算法均具有较好的定位效果,后者均比前者的定位精度更高,而且提出的基于互相关序列和BP网络的新定位算法在低信噪比和高混响的条件下,也具有较好的定位效果。In the sound source localization algorithm with the application of microphone array,the basic idea of a commonly used algorithm is to calculate time delays between the signals received by microphones and then determine sound source location according to the array structure.Based on the analysis of traditional sound source localization algorithm,for the double five element cross array model,the traditional localization algorithm based on generalized cross correlation-phase transform time delay estimation was firstly introduced.Then a localization algorithm based on generalized cross correlation-phase transform(GCC-PHAT)time delay estimation and back propagation(BP)network was presented,and a localization algorithm based on parabola cross-correlation time delay estimation and BP network was also present.Furthermore,by analyzing the main factors of influencing the time delay estimation,a new localization algorithm based on cross-correlation sequence and BP network was proposed,took the location of the maximum point,maximum value and other two values in the left and right points near the maximum point in the GCC-PHAT cross-correlation sequence as the inputs of BP network,and realized three-dimensional localization of the sound source by training the BP network.Simulation and experimental results show that all proposed algorithms have better localization effect and every latter algorithm has higher localization accuracy than the former one in comparison with traditional localization algorithm based on GCC-PHAT time delay estimation.The new localization algorithm based on cross-correlation sequence and BP network also has good localization effect under the condition of low signal-tonoise ratio and high reverberation.

关 键 词：声源定位 时延估计 反向传播(BP)网络 相关序列 广义互相关相位变换加权 麦克风阵列 

分 类 号：TN912.3[电子电信—通信与信息系统]