[Sound source localization] Sound source localization based on matlab generalized cross-correlation [including Matlab source code 548]

1. Introduction

Recently, the TDOA positioning accuracy in the spectrum monitoring system is optimized. The most important problem in TDOA is nothing more than the extraction of signal time difference. After reading the domestic and foreign literature, they unanimously recommend the "Generalized Cross Correlation (GCC) algorithm".
I checked the literature and tested the generalized correlation algorithm using simulated sine wave simulation in matlab, but it did not achieve the expected effect, which is puzzled. After browsing the comments on GCC on the Internet, many people at home and abroad complained that the expected simulation results could not be obtained. Finally, I think that in the generalized correlation algorithm formula, the main parameter that affects the calculation result is the phase information, which should be tested with a simulated modulation signal. After using Matlab to simulate FM/FSK/PSK, the effect is really good. Implementing this algorithm in a C++ program is just a matter of workload. The matlab simulation source code is shown here for your reference. Note that different versions of Matlab need to use different modulation functions.
You can use Generalized Cross-Correlation to estimate the time delay for a signal to reach two sensors. For example, the application background of using a microphone array to determine the location of the sound source. The problem can be described by the following model:
Insert picture description here

Second, the source code

clear all;
close all;
clc
load s.mat;
s1=s1_10db(:)';             %变为行向量
s2=s2_10db(:)';             %变为行向量
wnd=512;
inc=256;

[delay]=GCC_Method('standard',s1,s2,wnd,inc);
subplot(411)
plot(delay-wnd,'*')
ylim([0,12])
title('标准GCC')
xlabel('帧数')
ylabel('延时/点')

[delay]=GCC_Method('phat',s1,s2,wnd,inc);
subplot(412)
plot(delay-wnd,'*')
ylim([0,12])
title('Phat-GCC')
xlabel('帧数')
ylabel('延时/点')

[delay]=GCC_Method('scot',s1,s2,wnd,inc);
subplot(413)
plot(delay-wnd,'*')
ylim([0,12])
title('Scot-GCC')
xlabel('帧数')
ylabel('延时/点')

[delay]=GCC_Method('ml',s1,s2,wnd,inc);
 subplot(414)
plot(delay-wnd,'*')
ylim([0,12])
unction frameout=enframe(x,win,inc)

nx=length(x(:));            % 取数据长度
nwin=length(win);           % 取窗长
if (nwin == 1)              % 判断窗长是否为1，若为1，即表示没有设窗函数
   len = win;               % 是，帧长=win
else
   len = nwin;              % 否，帧长=窗长
end
if (nargin < 3)             % 如果只有两个参数，设帧inc=帧长
   inc = len;
end
nf = fix((nx-len+inc)/inc); % 计算帧数
frameout=zeros(nf,len);            % 初始化
indf= inc*(0:(nf-1)).';     % 设置每帧在x中的位移量位置
inds = (1:len);             % 每帧数据对应1:len
frameout(:) = x(indf(:,ones(1,len))+inds(ones(nf,1),:));   % 对数据分帧

Three, running results

Insert picture description here

Four, remarks