See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/311706147

A Smart System for Driver's Fatigue Detection, Remote Notification and

Semi-Automatic Parking of Vehicles to Prevent Road Accidents

Conference Paper · December 2016

DOI: 10.1109/MEDITEC.2016.7835371

CITATIONS READS

13 3,554

5 authors, including:

Alamgir Hossan Faisal Bin Kashem

University of Central Florida University of Dhaka
25 PUBLICATIONS   163 CITATIONS    6 PUBLICATIONS   33 CITATIONS   

SEE PROFILE SEE PROFILE

Md. Mehedi Hasan

Daffodil International University
59 PUBLICATIONS   318 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

BCI Assistive Devices View project

Development of a Tropical Ocean Vector Wind (OVW) climatology data product View project

All content following this page was uploaded by Faisal Bin Kashem on 31 July 2018.

The user has requested enhancement of the downloaded file.

 A Smart System for Driver’s Fatigue Detection,
Remote Notification and Semi-Automatic Parking
of Vehicles to Prevent Road Accidents
1
Alamgir Hossan, 2Faisal Bin Kashem, 3Md. Mehedi Hasan, 4Sabkiun Naher, 5Md. Ismail Rahman
Department of Applied Physics, Electronics & Communication Engineering
University of Chittagong
Chittagong-4331, Bangladesh
1ah.apece@cu.ac.bd, 2fbk835@gmail.com, 3mehedi93hasan@gmail.com, 4sabikaislam@yahoo.com,5ismilerahman79@gmail.com
Abstract— Drowsy driving is one of the main reasons of road
accidents. Different techniques have been reported in literature
to detect driver’s drowsiness, but almost all the prevailing
systems only alert the driver if drowsiness is detected.
Consequently, the drowsy driver continues driving, with a high
risk of devastating accident. In this paper, we proposed and
verified an EEG based system which not only alerts the driver by
alarm, but also puts the vehicle in semiautomatic parking mode
by controlling fuel supply if drowsiness is detected. At the same
time, it reports nearby police station by SMS which contains
necessary information to take essential steps locating the vehicle.
Stored EEG signals, obtained with wireless wearable headsets
from numerous subjects in different conditions by different
research groups, were used in this work. Power spectrum
analyses were carried out in MATLAB to determine the
dominant frequency components in the brain signals. The slow
wave to fast wave ratios of EEG activities were assessed for a
number of epochs to determine driver’s drowsiness. GPS and
GSM modules were used with Arduino MEGA for tracking,
remote notification and servomotor control. Performance of the
proposed system was evaluated by stored data which confirmed
its feasibility and reliability.
Keywords— drowsiness detection; BCI; road accident
prevention; semiautomatic parking; neuro headset; EEG
I. INTRODUCTION
Road accident has been one of the main reasons of deaths,
injuries and a huge amount of economic loss worldwide. It kills
almost 1.3 million people, wounds or makes disabled other 50
million and causes about USD $518 billion economic loss
throughout the world every year [1-2]. Unfortunately, low and
middle-income countries are the main victims of these
fatalities, about 90% of total deaths and injuries occur there [3].
In Bangladesh, over 12000 people die and other 35000 people
get severely injured annually [4]. The economic loss is
estimated to be about 2-3% of national GDP of this country [5].
According to the study, these figures of toll would be doubled
by the year of 2030, if no effective steps are taken [2]. In most
road crashes, young and working people who are usually wage
earner of their family and main driver of the national economy,
die or get permanently disabled. So, a road accident is not
merely an accident, it usually abolishes the dreams and overall
social and financial structure of a family killing or making
978-1-5090-5421-3/16/$31.00 ©2016 IEEE
disabled their breadwinners. The victims become a heavy
burden to the society resulting a severe development challenge.
There are a number of reasons behind the road accidents,
most of which are related to drivers [6-7]. Driver’s fatigue or
drowsiness has been proved to be the vital factor of road
accidents [8]. When drivers get drowsy, they loss vigilance,
abilities of perception and overall control over the vehicle [9-
10]. Consequently, accidents become inevitable. There are
various factors that can result in drowsiness. Long monotonous
driving or over driving, lack of sufficient sleep, drinking
excessive alcohol, taking sedative drugs, driving time etc. are
mentionable [11]. Whatever the cause, drowsiness of a driver
makes him unfit for driving with a high risk of tragic accidents.
Therefore, for safe driving, avoiding drowsiness is essential.
There have been numerous works for driver’s drowsiness
detection and inattention monitoring [12]. Mainly three types
of techniques have been used in these works. First types of
techniques generally utilize visual features such as facial
expressions, eye blink, gaze, head movements of drivers to
detect their drowsiness [13]. Computer vision and image
processing techniques are usually used in these techniques
[14]. The second types of techniques use vehicle based
parameters such as the movement of steering wheel, the rate of
change of speed, the pressure on the brake, the angle of the
wheels etc. to determine the level of driver’s drowsiness using
sensor based techniques [15].
Though these two types of techniques use non-intrusive
sensors and take the benefits of sophisticated image processing
and computer vision technologies, these techniques have some
serious limitations. Since the physical parameters what these
techniques are based on, may be different for different drivers
and different driving conditions, these techniques sometimes
incorrectly detect drowsiness which can be misleading. For
example, if a driver talks while driving, it could be misdirected
as yawning. Different lighting condition, wearing sunglasses
may affect the eye state analysis. For detecting actual
drowsiness, more effective techniques have been a crucial
need. The third types of techniques which are based on
physiological analysis of drivers can do this work more
accurately since physiological conditions are directly related to
the driver’s physical and mental fatigue. Different
physiological indexes such as electrocardiogram (ECG),
electroencephalogram (EEG) and elecrooculogram (EOG) are
usually used in these techniques. However, the EEG is
considered the ‘gold standard’ to detect driver’s fatigue. The
EEG based techniques which determine the electrical activities
of brain have been so promising in detecting driver’s actual
fatigue [16]. Further with the advent of portable and wireless
EEG signal acquisition systems, implementing EEG based real-
time drowsiness detection techniques in any environment has
been easily feasible [17].
There have been a number of real vehicle and vehicle
simulator based researches for detecting driver’s drowsiness
using EEG [18-22], which prove the better accuracy and
reliability of these methods over the visual feature based and
the other physiological signal based methods [23]. Many of
these methods proposed buzzer based alarm system inside the
vehicle to alert the driver if he/she is detected drowsy. These
systems are useful enough but most of the time the problem
remain unsolved because the drowsy driver continues driving
even after drowsiness is detected and alarm is rung.
Consequently, the risk of accidents still vestiges.
In this paper, we proposed and implemented an EEG based
system that continuously monitor driver’s brain signal to detect
drowsiness. If drowsiness is detected once, our system
activates a microcontroller module which does the following
actions simultaneously.
i. It enables the buzzer to awaken the driver.
ii. It puts the vehicle on semi-automatic parking mode by
controlling the oil flow through the engine and switch
the parking light on so that the driver can park the
vehicle in a suitable and safe place.
iii. It reports the nearby police station the details
including vehicle’s current location, license, driver’s
license, conductor’s mobile number etc. using GSM
and GPS technologies so that the concerned authority
can take necessary steps to prevent road accidents.
The remainder of this paper is organized as follows. In
Section II, we described our proposed method and materials
with a brief description of algorithm development and coding
techniques in a step by step way. Results and performance
verification of the designed system is summarized and
discussed in Section III. A future research direction is given in
section IV. We conclude this paper discussing the significance,
limitations of this work.
classified into five different bands (delta, theta, alpha, beta,
gamma) based on their frequency and amplitude of oscillation
[24]. The brain wave pattern of people changes in frequency
and magnitude with what they do, think or feel. Drowsiness or
mental fatigue is associated with a particular frequency and
amplitude ranges of brainwaves. The fastest brain waves are
gamma (γ) which have frequency above 22 Hz. These waves
are related to very excited and disordered or restless mental
condition [25] and not ‘normal’ for driving. Beta (β) brain
waves are associated with high state of wakefulness and have
frequency range from 13 Hz to 22 Hz. When a person
experiences beta waves as dominant, he remains highly alert,
more focused and vigilant [26]. The alpha (α) waves which
have the frequency 8 Hz to 13 Hz, are linked with the most
relaxed and stable brain states. A person while having
dominant alpha waves remains balanced, calm with complete
awareness of surroundings. The brain signals of frequency
from 4 Hz to 8 Hz are known as theta (θ) band. Consciousness
or vigilance is seen to be significantly declined and drowsiness
or light sleep is observed when theta waves are dominant [27].
The delta (δ) brain waves, having frequency of 0.1 Hz to 4 Hz,
are the slowest brain waves and are experienced during deep
sleep. During fatigue or drowsiness, delta and theta waves
become dominantly stable, alpha waves decreases marginally
and beta waves decline significantly [28]. Different algorithms
such as θ/β, α/β, (θ+α)/(α+β), (θ+α)/β and (θ+ δ)/(α+β) have
been used to determine the drowsiness [29-30]. Among them,
(θ+ δ)/(α+β) and (θ+α)/β ratios are observed to rise more
sharply during drowsiness [31].

II. METHOD AND MATERIALS

In this section, we describe the complete procedure and
interconnection of the modules of the proposed system in a
step by step way. The overall functional block diagram is
illustrated in Fig. 1.
A. Brain Signals and Their Acquisition
Electrical activity of human brain which is measured by
electroencephalogram (EEG), directly corresponds to the
physiological and mental status of human. These activities
exhibit rhythmic patterns called brain waves which are Fig. 1. Block diagram of the proposed sytem
 There have been available a number of cheap, reliable and
user-friendly portable EEG signal acquisition systems, such as
Emotive EPOC, Muse, Xwave headset etc. which can
effectively be used to capture the brain signals of the drivers in
real time. For this, the drivers have to wear the neuroheadset
during driving. Since the headsets are light enough like
ordinary music headsets, it would not be much uncomfortable
or annoying to them. Different wireless EEG headsets use
different number of electrodes on the different position of the
subject’s scalp. For example, the Emotive EPOC uses 14
different EEG channels with two other references placed
according to international 10-20 electrode positioning system
to continuously acquire brain signals at a rate of 128 samples
per second and 14-bit resolution. The module continuously
sends the obtained data to signal processing module by
Bluetooth technology [32].
The datasets we used for this work were collected from
EEG Motor Movement/Imagery Dataset (eegmmidb) stored in
PhysioNet [33]. There were over 1500 short EEG recording of
109 subjects in different physical and mental conditions.
BCI2000 system was used to record EEG signal using 64
channels and 160 Hz sampling frequency [34].
B. Signal Processing
The signal processing unit is an embedded signal
processing chip. In this work, we used MATLAB
8.5.0.197613 (R2015a) under Windows 64-bit platform. We
loaded a large number of signals in workspace, filtered with a
6th order Butterworth filter to remove different artifacts and
noises. Then, we performed spectral analysis of those signals
by Fast Fourier Transform(fft) algorithm to resolve the signals
and to determine the dominant frequency component for each
epoch. Finally, we calculated (θ+δ)/(α+β) ratio of brain signals
for a number of subjects over a certain period of time.
Fig. 2. Flow chart of drowsiness detection algorithm
C. Microcontroller Unit
The signal processing unit continuously acquire and
process the brain wave signals from the driver’s neuro headset
to determine the current brain state of the driver. If the brain
states correspond to alert or vigilant, (θ+δ)/(α+β) ratio is less
than threshold level or the fast waves are dominant, it keeps
going and the rest of the system remain idle. But when
drowsiness is detected, i.e. slower brain waves become
dominant over faster brain waves or (θ+δ)/(α+β) ratio exceeds
certain threshold value, the signal processing unit immediately
enables the microcontroller module. When enabled, the
microcontroller unit performs a number of works
simultaneously as depicted in algorithm of Fig. 3. It triggers
the buzzer system to alert the driver, puts the vehicle in
semiautomatic-parking mode by controlling the fuel supply as
described in the following section, activates GPS module to
know the current location and sends SMS containing location,
driver’s license no., contact number, vehicle no etc. to the
nearest police station using GSM module and existing
network. In this system, we used Arduino MEGA-2560 as
microcontroller device and Arduino IDE was used as
programming platform.
Fig. 3. Flow chart of controlling and reporting system
D. Alarming Unit
The buzzer system is an immediate safety for the vehicle
until police locate the vehicle. It is connected with the Arduino
as output device. When the buzzer rings, it awakes the driver
and the driver is then bound to park the vehicle to the nearest
suitable place if it is made mandatory traffic rule. In addition,
once drowsiness is detected, the fuel supply to the engine of
the vehicle is controlled by semiautomatic parking mechanism
so that the driver cannot drive further with acceleration even if
he wants. The buzzer also alerts the passengers, in case of
public transports, about the drowsiness of the driver so that
they can ask the driver to park the vehicle immediately. For
this unit we used SparkFun RedBot Buzzer for its good sound
quality.
E. Semi Auto-Parking Mechanism
This sub module is enabled immediately after the
drowsiness is detected. We designed vehicles speed control
system by controlling the fuel supply to the engine as shown in
the Fig. 4. Usually, driver supplies fuel to the engine by
pressing accelerator through a pipe line. In our semiauto-
parking system we used a valve attached to the pipe line which
could rotate clockwise or anti-clockwise. The movement of
valve was controlled by a servo motor. When drowsy condition
of driver was detected, the microcontroller sent a signal to the
servo motor and then it moved in such a direction that the valve
connected to it blocked the pipe line. In this way fuel supply to
the engine was controlled and it would make the driver to park
the vehicle in a nearest suitable place. Besides microcontroller
turned on the parking light to avoid collision with the vehicles
from behind.
Fig. 4. Semi Auto-Parking Mechanism
F. Tracking System
Microcontroller system activates GPS module when signal
processing module transform drowsy signal into command. In
response, the GPS module collects the current location of the
vehicle and feeds the location as input message to GSM device.
The module remains idle unless drowsiness is detected and it is
triggered.
G. Remote Notification System
Global System for Mobile(GSM) system plays a major role
in our system. When drowsiness is detected and it is enabled by
microcontroller unit, the GSM module communicated to a
remote server to know the contact number of a nearest police
station. Then it incorporated the location from the GPS module
with other necessary information including driver’s license,
vehicle no, driver’s contact number etc. into a text message to
send to the nearest police station. As a result, law enforcing
authority would be able to locate the vehicle within very short
time and replace the driver or take other necessary steps to
prevent likely crash. Then the system could be resumed to
normal driving mode using a code which could be applied to
microcontroller by law enforcing authority.
III. RESULTS AND DISCUSSION
We analyzed a number of EEG samples obtained from
different subjects having different physiological states. A 20-
sec window was used to filter the data by a 6th order
Butterworth filter and find the spectral components by fft
algorithm. A typical epoch from electrode Fp1, for example, is
shown in Fig. 5 with magnitudes of its spectral components. It
was a brain signal from an awake subject. Fig. 6 depicts
another epoch from a subject who was drowsy.
Fig. 5. Brain signal & its spectral components of awake state
Fig. 6. Brain signal & its spectral components of drowsy state
For a number of epochs obtained from a subject we
computed the slow brain waves to fast waves, namely,
(θ+δ)/(α+β) ratios over a period of time at one second time
interval. We repeated it many times until we went through a
particular moment when the subject was drowsy, the ratios
were observed to cross the threshold line as shown in Fig. 7 at
seventh second for the given data. After crossing the threshold
line, the alarming system awoke the driver like all other
prevailing system to alert driver. Since the driver became
vigilant temporarily, it results in dominating fast wave again.
However, as we claimed, only the alarming system is not
sufficient because the driver became drowsy again later and in
our experiment, the value at 15th second of Fig. 7 justifies this.
Fig. 7. Slow wave to fast wave ratios of a subject
For this reason, the rest of the system was designed to start
working as soon as the ratio cross the threshold value very first
time and the system was seen to work as expected. The speed
of servo motor was controlled accordingly in a model of
vehicle’s fuel supply during the experiment. The buzzer was
rung immediately after the ratio crossed the fixed threshold
value and an SMS containing necessary information was sent
to the desired number as shown in Fig. 8. We used a fixed
number for verification purpose. However, cell number of
nearest police station could be collected communicating with a
remote server utilizing existing GSM network.
Fig. 8. Sample notfication SMS
IV. FUTURE WORK
We have continued working to develop a fully automatic
vehicle parking system using digital image processing
techniques. This would take the full control to park the vehicle
at a nearest possible place as soon as driver’s drowsiness is
detected once forestalling drowsy driver from driving.
CONCLUSION
In this paper, we developed a driver’s drowsiness
detection, alarm and a novel semi-automatic parking system to
prevent road crashes. In addition, we extended the work to
report the nearby police station of the occurrence. We couldn’t
use real time data during the work because of our current
limitation to afford a real time EEG data acquisition system.
But performance of the proposed system was successfully
verified using the reliable stored benchmark data obtained
from relevant subjects. The data were suitable for our system
as these were collected by portable BCI2000 system. We used
slow waves to fast waves ratio for detecting drowsiness as it is
more reliable to detect drowsiness than any single wave.
Because single wave like alpha, beta, delta etc. fluctuates very
quickly where change of different signals is needed
simultaneously for a small change in slow wave to fast wave
ratio. We processed the signals in MATLAB platform and
implemented a model of the rest of the system in real time
which worked successfully. If this proposed system is
implemented with the legislation of proper traffic rules, it
would hopefully reduce the rate of devastating road crashes.
REFERENCES
[1] Association for Safe International Road Travel, Internet:
http://asirt.org/initiatives/informing-road-users/road-safety-facts/road-
crash-statistics [22 May, 2016].
[2] Global Status Report on Road Safety, World Health Organization
(WHO), 2015. Available:
http://www.who.int/violence_injury_prevention/road_safety_status/2015
/en/ [23 May, 2016]
[3] “World report on road traffic injury prevention,” WHO, Geneva, 2004.
Available:
http://www.who.int/violence_injury_prevention/publications/road_traffi
c/world_report/en/ [23 May, 2016]
[4] Md. Mazharul Hoque, Zobair Ibn Awal, Sabreena Anowar, Shahnewaz
Hasanat-E Rabbi, S. M. Sohel Mahmud, “Road Safety Research Needs
and the Role of Accident Research Centre (ARC)”, Proceedings of the
National Symposium on Engineering and Technological Education
(NSET), BUET, pp. 95-100, 29 December, 2007.
[5] Hoque M. M.,Mahmud S. M. S., Paul S., "The Cost of Road Traffic
Accidents in Bangladesh", 10th Pacific Regional Science Conference
Organization (PRSCO) Summer Institute 2008, Bangladesh Regional
Science Association (BRSA), pp. 88, 15-17 May, 2008.
[6] Jonathan Mosedale, Andrew Purdy, Eddie Clarkson, “Contributory
factors to road accidents”, Available: http://www.mvs.net/safety.html
[23 May, 2016]
[7] “National Motor Vehicle Crash Causation Survey”, U.S. Department
ofTransportation, National Highway Traffic Safety Administration, July,
2008. Available : www-nrd.nhtsa.dot.gov/Pubs/811059.PDF [23 May,
2016]
[8] “Driver Fatigue and Road Accidents”, The Royal Society for the
Prevention of Accidents (RoSPA), Available:
http://www.rospa.com/road-safety/advice/drivers/fatigue/road-
accidents/, June, 2011.[23 May, 2016]
[9] Jennie Connor, Gary Whitlock, Robyn Norton, Rod Jackson, “The role
of driver sleepiness in car crashes: a systematic review of
epidemiological studies,” Accident Analysis & Prevention, vol. 33, no. 1,
pp. 31-41, January 2001.
[10] Al-Khateeb, G.G., “Analysis of Accident Data and Evaluation of
Leading Causes for Traffic Accidents in Jordan,” Jordan J Civ Eng, vol.
4, no. 2, pp.76-94, 2010.
[11] L. P. Kostyniuk, F. M. Streff, and J. Zakarajsek, “Identifying unsafe
driver actions that lead to fatal car-truck crashes,” AAA Found. Traffic
Safety, 2002.
[12] S. Kaplan, M. A. Guvensan, A. G. Yavuz and Y. Karalurt, "Driver
Behavior Analysis for Safe Driving: A Survey," IEEE Transactions on
Intelligent Transportation Systems, vol. 16, no. 6, pp. 3017-3032, Dec.
2015.
[13] L. M. Bergasa, J. Nuevo, M. A. Sotelo, R. Barea and M. E. Lopez,
"Real-time system for monitoring driver vigilance," IEEE Transactions
on Intelligent Transportation Systems, vol. 7, no. 1, pp. 63-77, March
2006.
[14] A. B. Albu, B. Widsten, T. Wang, J. Lan, and J. Mah, “A computer
vision-based system for real-time detection of sleep onset in fatigued
drivers,” Proc. IEEE Intell. Veh. Symp., Eindhoven, The Netherlands,
Jun. 4–6, pp. 25–30, 2008.
[15] A. Sahyadehas, K. Sundaraj, and M. Murugappan, “Detecting driver
drowsiness based on sensors: A review,” Sensors, vol. 12, no. 12, pp. 16
937–16 953, Dec. 2012.
[16] M. Simon et al., “EEG alpha spindle measures as indicators of driver
fatigue under real traffic conditions,” Clinical Neurophys., vol. 122, no.
6, pp. 1168–1178, 2011.
[17] M. Li, Z. Cheng, and J.-F. Fang, “An EEG-based method for detecting
drowsy driving state,” in Proc. 7th Int. Conf. FSKD, pp. 2164–2167,
2010.
[18] G. R. Kumar, S. V. P. Raju, D. Kumar, “Classification of EEG signals
for drowsiness detection in brain and computer interface,” Comput. Sci.
Telecommun., vol. 4, no. 36, pp. 1512–1232, 2012.
[19] U. Svensson, “Electrooculogram analysis and development of a system
for defining stages of drowsiness,” M.S. thesis, Dept. Biomed. Eng.,
Linköping Univ., Linköping, Sweden, 2004.
[20] M. V. M. Yeo, X. P. Li, K. Shen, E. P.W. Smith, “Can SVM be used for
automatic EEG detection of drowsiness during car driving?” Safety Sci.,
vol. 47, no. 1, pp. 115–124, Jan. 2009.
View publication stats
[21] K. Q. Shen, X. P. Li, C. J. Ong, S. Y. Shao, E. P. V. Wilder-Smith,
“EEG-based mental fatigue measurement using multiclass support
vector machines with confidence estimate,” Clin. Neurophys., vol. 119,
no. 7, pp. 1524–1533, Jul. 2008.
[22] J. Hyun, S. Gih, K. Ko, S. Kwang, “A smart health monitoring chair for
nonintrusive measurement of biological signals,” IEEE Trans. Inf.
Technol. Biomed., vol. 16, no. 1, pp. 150–158, Jan. 2012.
[23] A. Hashemi, V. Saba, S. N. Resalat, “Real time driver’s drowsiness
detection by processing the EEG signals stimulated with external
flickering light,” Basic Clinical Neurosci., vol. 5, no. 1, pp. 22–27,
Winter 2014.
[24] PL Nunez, Ramesh Srinivasan. Electric fields of the brain: the
neurophysics of EEG, Oxford University Press, USA, 2006.
[25] G Buzsaki, Rhythms of the Brain, Oxford University Press, 2006.
[26] KR Popper, JC Eccles, The self and its brain, Springer Science &
Business Media, 2012.
[27] Gillberg, M., G. Kecklund, T. Åkerstedt, "Sleepiness and performance
of professional drivers in a truck simulator—comparisons between day
and night driving," Journal of Sleep Research, vol. 5, no. 1, pp. 12-15,
1996.
[28] Stern, John M. Atlas of EEG patterns, Lippincott Williams & Wilkins,
2005, pp. 27-55.
[29] Lal, S. K., Craig, A., Boord, P., Kirkup, L., Nguyen, H, “Development
of an algorithm for an EEG-based driver fatigue countermeasure,”
Journal of Safety Research, vol. 34, no. 3, pp. 321-328, 2003.
[30] Eoh, H. J., Chung, M. K., & Kim, S. H, “Electroencephalographic study
of drowsiness in simulated driving with sleep deprivation,” International
Journal of Industrial Ergonomics, vol. 35, no. 4, pp. 307-320, 2005.
[31] Jap, B. T., Lal, S., Fischer, P., & Bekiaris, E., “Using EEG spectral
components to assess algorithms for detecting fatigue,” Expert Systems
with Applications, vol. 36, no. 2, pp. 2352-2359, 2009.
[32] Stytsenko, K., Jablonskis, E., Prahm, C, “Evaluation of consumer EEG
device Emotiv EPOC,” MEi: CogSci Conference, Ljubljana, August,
2011.
[33] Goldberger, Ary L., et al. "Physiobank, physiotoolkit, and physionet
components of a new research resource for complex physiologic
signals." Circulation 101, no.23, pp. e215-e220, 2000.
[34] G. Schalk, D. J. McFarland, T. Hinterberger, N. Birbaumer and J. R.
Wolpaw, "BCI2000: a general-purpose brain-computer interface (BCI)
system," in IEEE Transactions on Biomedical Engineering, vol. 51, no.
6, pp. 1034-1043, June 2004.