You are on page 1of 5

2017 IEEE 3rd International Conference on Sensing, Signal Processing and Security (ICSSS)

Driver Assistant for the Detection of Drowsiness


and Alcohol Effect
Prajnyajit Mohanty Pallem Siddharth Kunja Bihari Swain Rakesh Kumar Patnaik
Department of EIE Department of EIE Department of EIE Department of EIE
National Institute of Science National Institute of Science National Institute of Science National Institute of Science
and Technology and Technology and Technology and Technology
Berhampur, Odisha, India Berhampur, Odisha, India Berhampur, Odisha, India Berhampur, Odisha, India
prajnyajitmohanty@gmail.com pallemsiddharth@gmail.com kunja.swain@gmail.com rkshpatnaik@gmail.com

Abstract—Drowsiness is becoming a severe issue in case of epileptic seizures, movement disorders, or migraine variants
traffic accidents. Normally, Sleeping can be identified from [2,10] or testing for the depth of anaesthesia. It is even
several factors like eye blink level, yawning, gripping force on the possible to see changes in EEG signals when the eyes of a
wheel and so on. But all these measuring techniques will check
person are either opened or closed or the change in the state of
only the physical activities of the human. In some cases, people
will mentally sleep with eyes open for a few seconds. This will drowsiness. Power spectra of EEG signals commonly lie in
make very big accidents in driving. The key work of the project is distinct frequency band from (0.5-100) Hz (Table I).
analysing the brain signals. Human brain consists of millions of
interconnected neurons. This neuron pattern will change TABLE I. Frequency ranges of EEG signals
according to the human thoughts. If a person is mentally sleeping
with eyes open then the attention level brain signal will get Brain Wave Type Frequency in Hz Mental States and Conditions
changed compared to the normal condition. This project work Delta 0.5 to 3 Deep dreamless sleep
uses a forehead single dry electrode which can collect Theta 4 to 7 Fantasy, Imaginary dream
Electroencephalogram (EEG) based brain signals of different Alpha 8 to 12 Relaxed but not drowsy
frequency and amplitude. Then using data aquisition system and Low Beta 13 to 15 Formerly SMR
LABVIEW software the collected signals are analysed. Mid range Beta 16 to 20 Thinking, aware of self
Depending on mental stages diffrent EEG signals are generated. surroundings
The EEG signals are analysed in LABVIEW to reveal the High Beta 21 to 30 Alertness, agitation
features which describes the human stage. Gamma 30 to 100 Motor function, higher mental
activity
Keywords —EEG; LABVIEW; Electrode; Alcohol; Biomedical

I. INTRODUCTION
Drowsiness in drivers causes major road accidents in day The presence of alcohol in body affects the normal range
to day life. According to the report of National Sleep of brain signals .In this study a set of EEG signals is used to
Foundation (NSF) above 51% of drivers drive their vehicle in detect whether a person is drunk or not [6]. In reality the
a drowsy or sleepy state and about 17% had actually fallen signals received from the human brain contains the
asleep [1,2]. From earlier research it is found that there are information about the person’s current status. There is a
many number of methods to detect the drowsiness condition, change in certain range of frequency received from drunken
out of which some techniques are very popular like change in people. The effect of alcohol causes a frequency shift of the
position of driver which is detected by the camera, measuring brain signals and this can be detected much more clearly than
physiological parameters i.e. eye blinking, EEG activities. using only breathalyzers. However an alcohol sensor can be
used to cross check the alcohol effect.
The purpose of our work is to detect different mental
conditions and effect of alcohol on the brain signals based on II. SYSTEM DESCRIPTION
EEG signals [3,4]. The drowsiness and alcoholic state can be
detected from the frequency plot of different brain signals. For the desired model various types of Electrodes
configuration can be used like 250 number of electrode system ,
EEG is well known for noninvasive measurement of 10-20 electrode cap and single forehead electrode (SFE) .For
different brain waves with high resolution in mill range [5]. medical diagnosis mainly 250 number of electrode system and
These brain-waves measured as electrical activity on the scalp 10-20 electrode cap are involved for better accuracy in
provides various information about a person.EEG can be measurement. However for detection of mental states SFE is
measured from the backside of the cortex or from forehead popularly used [4]. The benefits of using SFE include reliability
depending upon the necessity .In the previous studies EEG has and flexibility during movement. The advantage of SFE over
been used for observing and analyzing the emotions including other SFE system is that it avoids motional artifacts
testing for brain death or coma, distinguishing between

978-1-5090-4929-5©2017 IEEE
279
2017 IEEE 3rd International Conference on Sensing, Signal Processing and Security (ICSSS)

and reduces the awkwardness in wearing. In addition to the consists of a non-invasive sensor that won’t cause any pain to
system a separate alcohol sensor is interfaced to detect the the user who wears the headset. The led indicates blue for
presence of alcohol. Eye blink detection can be used as an normal condition and red for low power supply.
alternate way to indicate the drowsiness condition [7].
Data acquisition (DAQ) is the process of measuring an
electrical or physical parameters such as voltage, current,
temperature, pressure, or sound with a computer. A DAQ
system consists of sensors, DAQ measurement hardware, and
a computer with programmable software. Compared to earlier
measurement systems, PC-based DAQ systems accomplishes
the processing power, productivity, display, and connectivity
capabilities of industry-based computers providing a more
powerful, flexible, and cost-effective measurement system.
DAQ hardware acts as the interface between a computer and
signals coming from the SFE. It primarily functions as a
device that converts incoming analog signals to digital form so
that a computer can interpret them. The three key components
of a DAQ device used for measuring a signal are the signal
conditioning circuitry, analog-to-digital converter (ADC), and
computer bus.

An alcohol sensor is used for cross verification of adequate


amount of alcohol in the body. The sensor has high sensitivity
Fig.1.System block diagram to alcohol gas and resistant to the interference of gasoline,
smoke and vapor. It is suitable for detecting alcohol at various
concentrations. Low cost, long life span and simple drive
Fig.1 shows the block diagram of the proposed work. The circuit makes it popular for its use in various applications. The
raw signal is at first captured by a SFE from forehead of the alcohol gas sensor consists of total 6-pins including A, H, B
person. In order to read brain activity the EEG SFE used is a and the other three pins are A, H, B (Fig.2). Out of the total 6-
fully integrated, single chip EEG biosensor. The SFE safely pins we use only 4 pins. The two pins A, H are used for the
measures brainwave signals and monitors the attention levels heating purpose and the other two pins are used for the ground
of individuals as they interact with brain applications [8,9]. and power. There is a heating system inside the sensor, which
is made up of Aluminum oxide, tin dioxide.
The raw brainwaves captured from the user are then
processed digitally, by analyzing those data the stream of
input is passed to the system. The SFE is interfaced with the
computer via data acquisition (DAQ). The signals captured by
SFE are transmitted through wired medium to the DAQ card.
The desired signal conditioning i.e. amplification, filtration,
and linearization is done in the DAQ [10,11].

The input received from the DAQ is analysed using


LABVIEW program. LABVIEW is a graphical language
which uses data flow model. LABVIEW programs are termed
as virtual instruments. The desired result is shown in the front
panel.

A. HARDWARE DESCRIPTION

The SFE used here i.e. SFE is one single embedded SFE
system which runs on an external battery source. This is a dry
Fig.2.Pin configuration of an alcohol sensor
SFE that can measure brain waves from the scalp and thus can
easily be worn over hair.It comes with a power switch, a
sensor tip, flexible ear arm and a ground connection ear clip. It

978-1-5090-4929-5©2017 IEEE
280
2017 IEEE 3rd International Conference on Sensing, Signal Processing and Security (ICSSS)

The SUBVI CLOSE is used when the communication


between the SFE and computer is done so that the serial port
B.PROGRAM DESCRIPTION channel is closed for other programs to start using this
resource.
The data coming from SFE is resolved into stream of bytes to
reconstruct the various types of brainwave data. The data are
then interpreted and used for various types of brainwave
signals (including Attention, Meditation, and signal quality
data).

The raw signals generated from the brain are captured by


the SFE and then passed to the SUBVI STREAM through
INIT SUBVI, which is used for the communication purpose
between SFE and computer. The input ports associated with
INIT block are comport, timeout and buffer size.

The STREAM SUBVI samples the data in two frames i.e.


512 Hz (raw value) and 1 Hz (calculated variable length). The
data which is coming from the SFE is executed until a
synchronous byte is encountered. It should be ensured that the
data occurred after the synchronized byte should be a
synchronized data for further processing.

The data generated after the synchronized one i.e. PLENGTH


should be less than synchronized byte. The data obtained after the
synchronized byte is first stored in an array and then the lowest 8
Fig.3.Data flow inside a SUBVI bits of data are summed up and inverted. The inverted data is then
compared with the next data obtained.
For the simulation of signals the software used is
LABVIEW (Laboratory Virtual Instrument Engineering Another alternate way for indication of drowsiness can
Workbench).It provides an efficient and easy-to-use be done using eye blink detection. Constant eye blink shows
environment for code development especially when the user the awakened state of a person. If the duration between two
needs to interact with the program and visualize the results. successive eye blinks crosses the normal range i.e. (2-10sec)
Unlike text-based programming languages like C which indicates that the person is in a drowsy state. A buzzer is
follow a control flow execution model, the environment of attached to the system using Arduino, which is activated once
programming follows a dataflow execution model. the duration of two successive eye blink exceeds the normal
range (Fig.4). For cross verification of alcohol presence an
The basic SUBVI used here is INIT whose function is to alcohol sensor is interfaced with the system using Arduino
start the communication between the SFE and the computer. interfacing mechanism.
The VI receives packets of data from the SFE.

As shown in Fig.3 the SUBVI STREAM does the


complete job of unpacking the received data of the frames of
512 Hz and 1 Hz filling a structure that could be accessed later
to make calculations. In this SUBVI all the major signals are
computed.

The block PRECISE is a part of SUBVI STREAM which


is used to capture the raw EEG signal (512 Hz) and the
variable length data (1 Hz).It introduces some delay which
causes the signals to wait to pass to the AUTO block.

The block AUTO is used to do the recompilation of


information of the frames, but sequentially, first the raw EEG
and then later the variable length data.

Fig.4.Interfacing of alcohol sensor with Arduino module

978-1-5090-4929-5©2017 IEEE
281
2017 IEEE 3rd International Conference on Sensing, Signal Processing and Security (ICSSS)

III. RESULT AND DISCUSSION

The raw EEG signal of a normal person is recorded and


shown on the front panel (Fig.5). The hike in amplitude
signifies an eye blink.

Fig.6.EEG of Normal person (Red) Vs Drunken person (Blue) on time axis

So the alcoholic condition of the driver can be detected by


observing the alpha wave .It can also be verified from the
reading of alcohol sensor interfaced with the system.
Fig.5.Waveform of raw EEG signal

From a typical observation (Table II) we found the


different brain waves lies in the amplitude range of 10 to 90
(μV) whereas the standard amplitude range lies between 5 to
100 (μV).

TABLE II. Range of signals in amplitude

Types of Signals Amplitude(μV) Standard range(μV)


Delta 80 50 to 100
Theta 18 5 to 20
Alpha 64 50 to 100
Beta 24 5 to 30
Gamma 13 10 to 30

As shown in Fig.6 the raw EEG of a normal person and that


of a drunken person is observed to be varied in a large amplitude
range. The blue line signifies the EEG of a drunken person while
the red line indicates the EEG of a normal person.

The raw signal fetched lies in the frequency range of (0.5- Fig.7.Front Panel view for detection of eye blink
100 Hz). However the normal range of the waves differs in the
presence of alcohol in the body. From observation it is found It is observed that when there is no eye blinking occurred
that during alcoholic condition all the four waves except alpha for 15 seconds duration the buzzer starts ringing and in the
increased to a certain extent while alpha wave decreased. front panel a red coloured indicator starts glowing until an eye
blink happens which is noticed with green coloured indicator
on the LABVIEW front panel, shown in Fig.7.

978-1-5090-4929-5©2017 IEEE
282
2017 IEEE 3rd International Conference on Sensing, Signal Processing and Security (ICSSS)

IV. CONCLUSION [3] Y. Istefanapulos, A. AdemoXglu, E.M. Tzanakou,


Analysis of pattern visual evoked potentials by spline
In this paper, the raw EEG signal is detected using an EEG wavelets, IEEE Trans. Biomed. Eng. 44 (9) (1997).
SFE and is compared with the signal obtained from an [4] M. Teplan, “Fundamental of EEG Measurement,”
alcoholic person. The detection of eye blink signifies that Measurement Science Review, vol. 2, pp. 1-11, 2002.
whether a person is in drowsy state or not. If the person is not [5] M Benning, S Beyond, A Cochrane, D Uddenberg
blinking the eye then it indicates that the person is sleeping
“The Experimental Portable EEG/EMG Amplifier”
and hence a buzzer starts ringing. The eye blink can also be
detected from the raw EEG, the hike in the amplitude shows University of Victoria Faculty of Engineering
that the person is blinking the eye and hence there is a change ELEC 49A Report, August 2003.
in the amplitude of the signal. From the results it is possible to [6] R. R. Knipling and W. W. Wierwille, “Vehicle-
prevent drivers from causing accidents due to drunken driving. based drowsy driver detection: Current status and
Since this method deals with biological signals to identify the future prospects,” in Proc. IVHS Amer.4th Annu.
presence of alcohol in the body. The acquired EEG signal is Meeting, Apr. 1994, pp. 2–24.
processed and analyzed for alcoholic activity. On the basis of
[7] G.N. Keshava, Murthy and Zaved Ahmed Khan
analysis performed, the EEG wave of a normal person from
that of a drunken person based on the condition i.e., the alpha (2013), ‘Smart Alert System for Driver Drowsiness
activity is decreased. The above finding helps to make the Using EEG and Eyelid Movements’ Middle-East
driver aware of drowsiness and save multiple lives every year Journal of Scientific Research P.P NO 610- 619.
while driving. [8] T. P. Jung, S. Makeig, M. Stensmo, and T. J.
Sejnowski,“Estimating alertness from the EEG
ACKNOWLEDGEMENT power spectrum,” IEEE Trans. Biomed. Eng.,vol. 44,
pp. 60–69, 1997.
The authors wish to acknowledge the lab facility provided
by the Centre of Excellence NI lab, National Institute of [9] Norizam Sulaiman , Cheng Chee Hau, Amran
Science and Technology. Abdul Hadi, Mahfuzah Mustafa, Shawal
Jadin“Interpretation of Human Thoughts using EEG
REFERENCES signals and LabVIEW”, IEEE International
Conference on Control System, Computing and
[1] NSF. SLEEP FACTS AND STATS. NATIONAL Engineering, 28 - 30 November 2014.
SLEEP Foundation,Washington, DC. [10] Brain Wave Signal (EEG) of Neurosky, Inc. Page
[2] Jinu Jai, Geevarghese Titus, Dr. Purushothaman, 8, December 15, 2009.
“EEG Based Automatic Detection of Drowsy [11] I A.N. Akansu, Wavelets and filter banks, IEEE
State”, Spring, May 2013. Circuits Devices,November(1994)14–1.

978-1-5090-4929-5©2017 IEEE
283

You might also like