I am not able to get the desired angle of the sound source using microphone sensors connected to arduino where the data is sent serial to matlab

So I have written the code to determine the direction of sound source using two microphone sensors connected to arduino where the data is sent serially to matlab. So the problem is when I set the delay of 0.57 I am getting the sampling frequency to be 1754 so I am able to play the sound recored butwhen I sample with this frequency and I place my sound source at 90 I am getting 90 and when I place it at zero I am getting zero but anything between 0 and 90 I am getting 30 I don't know what is the problem. So when I increase by sampling frequency the recored sound will be played at a faster rate and the angles will decrease so can you guys help me hoe to overcome this problem. I am using the method called TDOA

Code for arduino

const int microphonePin1 = A0;

const int microphonePin2 = A1;

void setup() {

pinMode(microphonePin1, INPUT);

pinMode(microphonePin2, INPUT);

Serial.begin(250000);

}

void loop() {

// Read the analog value from the microphone

int microphoneValue = analogRead(microphonePin1);

int microphoneValue1 = analogRead(microphonePin2);

// Send the microphone value over Serial

Serial.println(microphoneValue);

Serial.println(microphoneValue1);

delay(0.57);

}

Code for matlab

% Function to call the arduino

arduinoPort = "COM3";

s = serialport(arduinoPort,250000);

% We are opening the the data present in arduino

fopen(s);

% Parameters for real-time data acquisition

Fs=1754;

% Note we know that fs= num_samples/duration

num_samples =80000; % No of samples

Data = zeros(num_samples, 1); % Creating an empty array to store the voltage data coming from arduino

tic

% tic and toc is used to calculate the time taken to run the program

% Read voltage data in real-time and store it in Data

for i = 1:num_samples

Data(i) = str2double(readline(s)); % Read data from Arduino

end

toc

% Since we are using two microphones the data coming from arduino is a an

% array in order to split we use a reshape function two get a row of two

% data

z=reshape(Data,[2,length(Data)/2]);

signal1=z(1,:); % Signal1 is from mic1

signal2=z(2,:); % Signal2 is from mic2

% Close the serial port connection

clear s;

% Scale the voltage data to the audio range (-1 to 1)

scaledData = (signal1 - min(signal1)) / (max(signal1) - min(signal1)) * 2 - 1; % for microphone 1

scaledData1 = (signal2 - min(signal2)) / (max(signal2) - min(signal2)) * 2 - 1; % for microphone 2

figure(1);

plot(scaledData)

figure(2);

plot(scaledData1)

% Once we get the data we store it a .wav file

audiowrite('sab3.wav',scaledData,Fs);

audiowrite('sab4.wav',scaledData1,Fs);

[y,Fs1]=audioread('sab3.wav');

[l,Fs]=audioread('sab4.wav');

% % Play the audio

% sound(y, Fs);

% sound(l,Fs);

% Get delays

delay1 = getDelay(y,l,Fs); % Writing the function to call delay

tdoa=delay1;

% Define the microphone spacing and sound speed

micSpacing = 0.4; % Distance between the microphones in meters

soundSpeed = 348; % Speed of sound in meters per second

% Calculate the angle of arrival (AoA) based on the TDOA

angleRad = asin(tdoa * soundSpeed / micSpacing);

angleDeg = rad2deg(angleRad);

disp(angleDeg)

% Function to calculate the delay

function [delayTime,delaySamples] = getDelay(y,l,fs)

len = length(y);

y = y(1:len);

l = l(1:len);

[r,lags] = xcorr(y,l); % get the cross-correlation and view the lags

figure

stem(lags,r)

xlabel('Lags [samples]')

ylabel('Cross-correlation')

[~,maxlags] = max(abs(r)); % position of maximum lag

delaySamples =lags(maxlags);% number of samples difference

delayTime = delaySamples/fs; % convert to time

fprintf('\nSignal 2 lags behind Signal 1 by ~%.1f microseconds\n',1e6*abs(delayTime))

end

Answer 1

[Brahmadev]

Hi Abhilash,

I understand that you would like to read data from the serial port using MATLAB and calculate the angle of arrival from two microphones.

The implementation of AoA seems mathematically correct. Since you mentioned that you can hear the correct audio in both cases, here are a few things to consider:

1. Using "delayMicroseconds()" instead of "delay()" when a delay smaller than 1 milisecond is needed.
2. "fopen()" should not be used for "serialport" objects.
3. Minimize noise and reflections: Noise and reflections can affect the accuracy of TDOA-based sound source localization. Try to minimize background noise and avoid surfaces that can cause reflections or echoes. This will help improve the accuracy of the measurements.

Hope this helps!

[ABHILASH N]

Thank you Sir