Twelfth LACCEI Latin American and Caribbean Conference for Engineering and Technology (LACCEI’2014)

”Excellence in Engineering To Enhance a Country’s Productivity” July 22 - 24, 2014 Guayaquil, Ecuador.

A.Ds: Aid Device for Deaf Drivers

Junghee Lee
Vaughn College, Flushing, NY, USA, [email protected]

Faculty Mentor:
Dr. Amir Elzawawy
Vaughn College, Flushing, NY, USA, [email protected]
Dr. Hossein Rahemi
Vaughn College, Flushing, NY, USA, [email protected]

ABSTRACT
Worldwide, deaf people are allowed to drive. However, as many people wonder, deaf people cannot hear audible
cues such as a police siren, an ambulance needing the right of way, or even a honking horn. Deaf people say they
simply pay attention to visual cues, such as the flashing lights, but honking horns do not have visual cues. For the
safety of deaf people and to reduce the discrimination of society about deaf people driving, need of a device that
will alert a deaf driver and also, indicate where the sound is coming is noticed. To build a desired device, the
Arduino Uno board is to be used as a microcontroller to acquire the sound signal, and since MATLAB is
compatible with Arduino Uno board [1], MATLAB is used to analyze the data and write the program..
Keywords: Arduino Uno board, MATLAB, Audio to visual cues, Communication, Deaf drivers driving

1. INTRODUCTION
In order to prove if hearing is really required for driving, many professional researchers study and collect data.
However, the data is still not enough, and some people believe the lack of data regarding hearing and driving
safety may be due to the fact that hearing plays only a minor part in the overall driving task. Still, very few data
exist for each situation to determine and more studies are needed, but the well-designed study of deaf automobile
drivers by Coppin and Peck in California [2] found that deaf men had 70% more road crashes than non-deaf men.
In fact, Henderson and Burg (1974) [2] have outlined four, non-routine driving situations in which hearing could
be important, and the audibility of warning sound from outside the vehicle was one of them.
Over 90 countries allow deaf people to drive. Even though there are still countries denying it, U.S. has nearly
always believed deaf people have right to drive. In the past, a number of case reports of deaf drivers driving safely
for years supported a theory that hearing is not a necessary requirement for driving [3], but many non-deaf people
still wonder how deaf people can drive without being able to hear what is going on outside of the car. Some of
them consider deaf driver as a safety threat. Even though most insurance companies consider deaf people as safer
drivers than non-deaf people, hearing people say letting deaf people drive is not safe because deaf drivers cannot
hear audible cues such as a police siren, an ambulance needing the right of way, or even a honking horn. Because
of this kind of reason, deaf people get discriminated, and some are not even allowed to drive at all. However, this
should not be the reason to stop deaf people from driving a car. As a human being, all the deaf people have right
to drive [3], and non-deaf people shall not judge their abilities to drive. If a device can be ears for deaf drivers and
make non-deaf drivers feel safer about deaf people driving, this device should be developed and bring equality for
deaf people.

2. DESIGN
One proposal to solve the problem is, creating a device that will alert a driver and also will indicate where the
sound is coming from is decided. A.Ds, Aid Device for Deaf Drivers, will mainly have three parts in order to
convert audible cues to visible cues and vibration alert.

2.1. INPUT
First, six small microphones will be installed to the outside of the car body to receive the audible cues. In order to
indicate where the sound is coming from, each microphone shall be installed around the car about every 60°.
Microphone sensor such as the one shown below in Figure 1 is to be used. This type of microphone sensor is
Arduino board compatible.

Figure 1. Arduino board compatible microphone sensor

2.2. MICROCONTROLLER
Second part is a microcontroller. When the input sound signal is received, the signal is should be analyzed and
filtered in order to trigger the output signal. The Arduino Uno board is used to process the signal.

Figure 2. Arduino Uno board

This microcontroller board has 14 digital input/output pins, 6 analog inputs, a 16 MHz ceramic resonator, a USB
connection, a power jack, an ICSP header, and a reset button. It can be powered by a computer with a USB cable
or a AC-to-DC adapter or battery to get started. Therefore, in a car, the device using this board, A.Ds, can be
simply powered by the cigarette lighter receptacle with a USB cable and a USB car adapter. The Arduino board
can be originally programmed with the Arduino software, but in this project, MATLAB was used instead.
Arduino board will be used to connect six microphone sensors to vibration device and LED lights. Arduino board
also will be programmed to sense car-honking sound through the traffic noise and convert this input sound into
output signal, which will activate the vibration device and LED light on the panel.

2.3. OUTPUT
To alert a deaf driver, output signal is important and necessary. Among many options, the vibration motor and the
LED light are chosen. The vibration motor will be added to the driver’s car seat to convey the instantaneous alert
signal. Additionally, LED panel indicator is designed, and will have six LED lights, and each LED light
represents each sensor. When the sound is received, LED light will flash to indicate where the sound is coming.

Figure 3. Output: Vibration motor and LED light

Following Figure 4 shows complete design of the device.

Figure 4. Conceptual design of three parts of microphone sensors, vibrating device and LED panel
indicator

3. ANALYSIS AND PROGRAMMING: MATLAB SCRIPT

After the design was completed, the sound samples of honking sound was recorded and collected. In this project,
instead of live input, recorded sample was used. Through the analysis and programming, success of verifying if
the sound is a honking sound or not is desired.
By using MATLAB, testing the sample, reading a WAVE file, and plotting it was possible. First, in order to load
the sample WAVE file, wavread(‘’) commend was used. After loading the sound, the sound was plotted as Figure
6. The size of the sample was computed, and with its sample rate, which is 44100 Hz, time increment per sample
was calculated. Next, Fast Fourier Transform algorithm function [4], FFT, was used in order to create the
frequency spectrum [5]. In MATLAB, discrete Fourier transform (DST) is used to implement FFT as described
below.

Where y is the FFT output for x vector (sound signal in time domain) with length n, j and p are the indices
associated with the vector in a double for loop. ω is a s complex nth root of unity [6].

Figure 6. Plot of a second sample from 37 to 41 sec

The FFT allows you to efficiently estimate component frequencies in data from a discrete set of values sampled at
a fixed rate. Following MATLAB script is used to get the frequency spectrum of the sample.
% Selecting a portion (37s to 41s)

b1=b(44100*37:44100*41);
size(b1)
soundsc(b1,44100);
pause(3)

% FFT algorithm

n=176401;
fs=44100;
ts2=1/fs;
tmax2=(n-1)*ts2;
t2=0:ts2:tmax2;
f2=-fs/2:fs/(n-1):fs/2;
z2=fftshift(fft(b1));

% Plotting the Frequency Spectrum

figure(2)
plot(f2,abs(z2))
title('Frequency Spectrum');
xlabel('Frequency Hz');
ylabel('Amplitude');
Plot of the frequency spectrum of a sample after FFT is shown below.

Figure 7. Frequency spectrum of a sample

After realizing that each frequency shows difference in sound, the frequency spectrum of a sample was zoomed
and studied closely.

Figure 8. Maximum amplitude at f = 2619 Hz

From the research, frequency of the car honk was found to be around 400 to 500 Hz [7]. However, unlike the data
from the research, from close observation, maximum amplitude of the frequency spectrum came out to be 2619
Hz. Therefore, instead of the research data, following the experimental data seemed like a right choice.
After FFT, the sample sound was filtered based on this experimental data. For filtering, bandpass filter [8] was
chosen.

Figure 9. Bandpass filter specifications

After the filter was applied, following graph was obtained.

 Figure 10. Frequency Spectrum after Bandpass Filter

The filtered sound was checked, and it definitely had high noted beep sounds. It was clearly different from the
original sound, and proved that this data can be used for the A.Ds device to filter out other noise and verify if
received sound is a honking sound or not. Therefore, x data and y data were obtained from Figure 10, and found
maximum amplitude as y max value and found x value for this y max. With those two values, the code that will
verify whether the sound is a honking sound or any other noise was programmed. Also, the code that will connect
to the Arduino Uno board and send out the output signal was written. However, in order to reduce the time to
receive the outcome signal, the code connecting the Arduino Uno board to the computer was replaced at the
beginning of the whole script. Following MATLAB script shows the whole process.

4. MATLAB SCRIPT
% Connect Arduino board :D !!
a=arduino('/dev/tty.usbmodemfa131')
% specifies output pin mode for pin 10
a.pinMode(10,'output');
% Once the Arduino board is connected, second sample should not have above
% command
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%
% Reading wav file
b=wavread('/Users/Brielle/Desktop/traffichonk.wav');
% Selecting a portion (37s to 41s)
b1=b(44100*37:44100*41);
size(b1)
soundsc(b1,44100);
pause(3)
% FFT algorithm
n=176401;
fs=44100;
ts2=1/fs;
tmax2=(n-1)*ts2;
t2=0:ts2:tmax2;
f2=-fs/2:fs/(n-1):fs/2;
z2=fftshift(fft(b1));
% After bandpass toolbox is used, alert saying "The variable 'Hbp' has been
% exported to the command window.' will show up.
% Plotting the sample b1
figure(1)
plot(t2,b1)
title('Plot of a Sound Sample');
xlabel('time(s)');
ylabel('Sound data');
% Plotting the Frequency Spectrum
figure(2)
plot(f2,abs(z2))
title('Frequency Spectrum');
xlabel('Frequency Hz');
ylabel('Amplitude');
% Checking the sound
b3=filter(Hbp2,b1);
z4=fftshift(fft(b3));
soundsc(b3,44100)
% Plotting the frequency spectrum after filtering
figure(3)
plot(f2,abs(z4))
title('Frequency Spectrum after Filtering');
xlabel('Frequency(Hz)');
ylabel('Amplitude');
h = gcf; %current figure handle
axesObjs = get(h, 'Children'); %axes handles
dataObjs = get(axesObjs, 'Children'); %handles to low-level graphics objects in axes
objTypes = get(dataObjs, 'Type'); %type of low-level graphics object
xdata = get(dataObjs, 'XData'); %data from low-level grahics objects
ydata = get(dataObjs, 'YData');
% MAX Amplitude ymax and x value
[ymax,xindex]=max(abs(z2))
xvalueforymax=abs(xdata(xindex))
% Varifying Honking sound 1=Honk 0=Noise
for i = xvalueforymax;
if le(2630,i)
p=0;
elseif le(i,2615)
p=0;
else
p=1;
end
for p=1;
if le(ymax,30)
p=0;
else
p=1;
end
end
end
i
p
% Output
a.digitalWrite(10,p);
pause(3);
a.digitalWrite(10,0);

5. APPLICATION
Even deaf people have right to drive. They cannot hear, but they still can see and feel. The main purpose of this
device is to decrease the gap of sensing alert while driving between deaf people and non-deaf people, so the
discrimination against deaf people driving can decrease.
A study has found that the number of crashes caused by blind spots [9] has risen by 50% in between 2009 and
2011. The blind spot causes accidents to all the drivers. However, sometimes before you get into a car accident,
honking from the other car alerts drivers and prevent the car accidents. Reacting to that alerting sound, drivers
will rapidly take an action to correct their course. However, since deaf people cannot hear the honking sound, it is
not applicable for them. Therefore, instead of that alerting sound, by using this A.Ds device and giving a vibration
alert, the same result is expected for deaf drivers.

6. CONCLUSION
The primary goal of this project was to make Aid Device for Deaf Drivers that will alert them when the car
honking sound was detected near a car. Because each sound forms its own sound wave, the frequency of sound
waves is also unique. Therefore, frequencies of the sounds were essential to be studied, and through the process of
research and practice, it is possible to distinguish between the car honking sound and other sounds by using the
difference of frequencies and the amplitudes of frequencies from sounds.
MATLAB is used for analyzing and filtering in the process. After the frequency spectrums were plotted, the
useful band of frequencies was identified. Other than that band of frequencies, they were considered as noises and
got eliminated by bandpass filtering. The maximum amplitude was used to find the useful band of frequencies,
and the presence of this maximum amplitude in the waves triggered the output signal. The output signal was
transmitted to the Arduino board and lit the LED light.
Proving the possibility of creating the Aid Device for Deaf Drivers by succeeding in getting output signal out
through the Arduino board, the hope of reducing the discrimination of society toward deaf drivers increased.
However, the time MATLAB took to communicate with the Arduino board delayed the output LED light signal.
For spontaneous output signal, generating the Arduino code from MATLAB code could be practiced to upload
and save the programming code on the Arduino board directly. To make a practical device, microphone sensors
are to be added as an input and develop code for receiving the live input signal instead of reading wave sound of
samples. Until this device comes true in real life and help deaf drivers, this project should be continued.

REFERENCES

[1] "Arduino Support from Simulink." MathWorks. N.p., n.d. Web. 14 Nov. 2013.
<http://www.mathworks.com/hardware-support/arduino-simulink.html>.
[2] Hearing Disorders and Commercial Motor Vehicle Drivers. Washington, D.C. 20590: Office of Motor
Carriers, Oct. 1992. PDF.
[3] World Federation of the Deaf, "WFD Statement on Deaf People's Right to Drive a Car or Other Vehicles."
March 31, 2009 <http://www.wfdeaf.org/wp-content/uploads/2011/04/WFD-Statement-on-Deaf-Peoples-
right-to-drive-a-car-or-other-vehicles-updated-31-March-20091.pdf>
[4] Didier. "FAST FOURIER TRANSFORM (FFT) (PART 1-11)." Arduinoos RSS. N.p., 05 Oct. 2010. Web. 01
Dec. 2013. <http://www.arduinoos.com/2010/10/fast-fourier-transform-fft/>.
[5] "Graphing a Frequency Spectrum with Matlab." YouTube. YouTube, 16 Nov. 2010. Web. 5 Nov. 2013.
<http://www.youtube.com/watch?v=dwzQnbeKnQg>.
[6] “Discrete-Time Signal Processing”, by Oppenheim, A. V. and R. W. Schafer, Prentice-Hall, 1989
[7] "Vehicle Horn." Wikipedia. Wikimedia Foundation, 09 July 2013. Web. 23 Oct. 2013.
<http://en.wikipedia.org/wiki/Vehicle_horn>.
[8] "Bandpass Filter." MathWorks. N.p., n.d. Web. 14 Nov. 2013. <http://www.mathworks.com/hardware-
support/arduino-simulink.html>.
[9] "Blind Spot Crashes Increase." The Telegraph. Ed. David Millward. N.p., 22 Sept. 2011. Web. 01 Oct. 2013.
<http://www.telegraph.co.uk/motoring/news/8779153/Blind-spot-crashes-increase.html>.

Authorization and Disclaimer

Authors authorize LACCEI to publish the paper in the conference proceedings. Neither LACCEI nor the editors
are responsible either for the content or for the implications of what is expressed in the paper.