Third International Conference on Broadband Communications, Information Technology & Biomedical Applications

Single Channel Wireless EEG: Proposed Application in Train Drivers

Surya Darma Ridwan*, Robert Thompson*, Budi Thomas Jap*, Sara Lal*, Peter Fischer+
* +
University of Technology, Sydney Signal Network Technology Pty Ltd
Broadway NSW 2007, Australia Lane Cove NSW 2066, Australia
Surya.D.Ridwan@student.uts.edu.au p.fischer@telstra.com
Robert.Thompson@student.uts.edu.au,
Budi.T.Jap@uts.edu.au,
Sara.Lal@uts.edu.au

Abstract activities as fatigue increases. Beta activity has also

been linked to task performance. As beta level
Electroencephalography (EEG) can be used as an decreases, task performance has also been found to
indicator of fatigue. Several studies have shown that decrease [5]. However, an EEG is a complex piece of
slow wave brain activities, delta (0-4 Hz) and theta (4- equipment that is normally designed for laboratory use.
8 Hz), increase as an individual becomes fatigued, In order to successfully develop a fatigue
while the fast brain activities, alpha (8-13 Hz) and beta countermeasure device, a simple, portable, and wireless
(13-35 Hz), decrease. However, an EEG is a complex EEG device is required.
piece of equipment that is generally used in laboratory The human EEG signal ranges in frequency from
based studies. In order to develop a fatigue 0.5-100 Hz, with amplitudes of 1-300 ȝV measured at
countermeasure device for train drivers using EEG, the surface of the skull [2, 11, 12]. These signal
there is a need for a simple and wireless EEG monitor. characteristics have inherent challenges to their
This paper explains the development of a single measurement. Firstly, 50 Hz noise is present within the
channel wireless EEG device. EEG frequency range. Significant electrical noise is
present at this frequency in most environments. The
very low frequency EEG signal makes it susceptible to
1. Introduction 50 Hz noise. The signal may further be corrupted by
impedance imbalance at the skin-electrode interface,
In the late 1800s Richard Caton (1842-1926) first electrode half-cell potential, and movement artifacts
reported the presence of biopotentials on the surface of [2].
the human skull [1]. Since then, electroencephalograms This technical paper proposes the design of a single
(EEG) (Hans Berger pioneered the field), have been channel wireless EEG suitable for fatigue detection on
used by medical practitioners to diagnose various train drivers.
neurological conditions [2].
EEG has been monitored in many driver studies 2. Hardware Development
conducted in the lab and field [3-7]. Subsequently, Lal
& Craig [8] proposed that EEG could be used as and A prototype of a wireless EEG system has been
indicator of fatigue and proposed driver fatigue developed in order to extract EEG signal from human
algorithms. Later, Jap et al. [9] proposed an EEG-based subjects, to be analysed for fatigue detection. The
fatigue algorithm for application in the train driving hardware module consists of transmitter and receiver
environment. EEG has been found to be one of the ends. While the receiver end is wholly digital, the
most reliable physiological indicators of fatigue [10]. transmitter end consists of analog and digital sections.
Lal & Craig [3] reported increases in slow wave brain A block diagram of the system is depicted in Figure 1.
activity with driver fatigue. Delta (0-4 Hz) and theta (4- The prototype is designed to operate with a 6-volt (V)
8 Hz) activities have been found to increase as one gets battery. A voltage divider and voltage follower are
fatigued [8]. Jap et al. [9] found significant decreasing used to obtain a +/-3 V supply.
trends of alpha (8-13 Hz), and beta (13-35 Hz)

978-0-7695-3453-4/08 $25.00 © 2008 IEEE 58

DOI 10.1109/BROADCOM.2008.69
Authorized licensed use limited to: University of Witwatersrand. Downloaded on February 23,2024 at 13:42:51 UTC from IEEE Xplore. Restrictions apply.
 Figure 1 Block diagram of the EEG system prototype

2.1. Front End subsequent to the instrumentation amplifier. The DC

restorator is implemented by using an op-amp in the
The subject is connected to the front end of the feedback loop of the AD620. The DC restorator has
transmitter through three EEG electrodes. The front two settings, namely the ‘monitor’, and the ‘quick
end is mainly the analogue electronics, which consists restore’. The quick restore setting can be used to speed
of an instrumentation amplifier followed by signal up the recovery of the output if it becomes saturated
amplification and conditioning stages (i.e. direct [2]. The model in Figure 3 was used to design the DC
current (DC) restorator, gain amplifier, high-pass and restorator.
low-pass filters, variable attenuation and signal level The Driven Right Leg (DRL) is incorporated in the
shifter). The front end schematic is shown in Figure 2. design to reduce common-mode noise, such as 50 Hz
The function of the amplifier is to amplify the interference unavoidably coupled into the subject and
difference of the signal from the left and right EEG electrodes. The name, Driven Right Leg, has been
electrodes, and simultaneously reject common mode preserved from its first use in electrocradiography
noise at both inputs. The AD620 instrumentation (ECG) equipment [14], although this electrode is not
amplifier (Analog Devices, USA) was the connected to the subject’s right leg for EEG recordings.
instrumentation amplifier chosen for this design as it As shown in Figure 2, the DRL is a feedback circuit
offers low power and high Common Mode Rejection similar to the DC restorator. It feeds the inverse of the
Ratio (CMRR). Without a high CMRR, common mode common-mode voltage back to the human subject,
voltages, which are typically 1V [13], will get which acts to reduce the common-mode noise present
amplified and prevent the EEG signal from being at the active electrodes [14].
recovered. The high-pass and low-pass filters are used to cut-
In the subsequent stages, LMC6484 (low-power off the signal with frequency components below 0.25
Operational Amplifier (op-amp), National Hz and above 95 Hz. The high pass filter is
Semiconductor, USA) Integrated Circuits (ICs) are implemented by using a 2nd-order Sallen-Key filter
used to implement all signal conditioning circuitry. The [15]. It attempts to eliminate the DC potential present
LMC6484 has low-power consumption making it at the electrode. The low pass filter is implemented by
suitable for portable battery powered applications. using two 2nd-order Sallen-Key filter stages, therefore it
The DC restorator is used to eliminate DC offset is effectively a 4th-order filter.
which would otherwise saturate the op-amps

59

Authorized licensed use limited to: University of Witwatersrand. Downloaded on February 23,2024 at 13:42:51 UTC from IEEE Xplore. Restrictions apply.
 Figure 2 Schematics of the EEG Front End

2.1. MSP430 and Wireless Modules

The first prototype of the digital section utilises the

eZ430-RF2500 development tool (Texas Instruments,
USA). It contains both the MSP430 microcontroller
and CC2500 wireless module (see Figure 1). The
microcontroller is suitable for low power operation as
it only consumes 0.7 microAmps (uA) while in stand-
by mode. In addition, it processes 16-million
instructions per second, and is equipped with two 10-
bit Analog-to-Digital (A-D) converters with maximum
Figure 3 DC Restorator Model sampling rates of 200 kilo samples per second.
The communication protocol running on CC2500 is
SimpliciTI, which is a dedicated low data rate protocol
for low power applications. The wireless transmission

60

Authorized licensed use limited to: University of Witwatersrand. Downloaded on February 23,2024 at 13:42:51 UTC from IEEE Xplore. Restrictions apply.
 utilises the 2.4 Giga Hertz (GHz) Industrial, Scientific signal (heart beat) by connecting 3 electrodes from the
and Medical (ISM) frequency [16]. device to the left wrist, right wrist and right leg of the
At the transmitter end, the microcontroller takes the subject. The subject was asked to remain seated and
output of the analog front end and performs the A-D stop moving in order to avoid movement artefacts.
conversion. A fatigue detection algorithm could be With the aid of the testing program, the device was
implemented in the microcontroller at the transmitter able to successfully capture the PQRS complexes of the
end to detect fatigue before the digitised signal is ECG signal from the subject, as shown in Figure 4.
transmitted to the CC2500 module of the receiver end. Subsequently, the prototype was used to capture the
At the receiver end, the same eZ430-RF2500 EEG signal from electrodes placed on the frontal lobe
development board, which is used at the transmitter of the head. This was performed by connecting a pair
end, is connected to a PC via a USB slot. of shielded electrodes to the pre-frontal brain sites (FP1
and FP2) and the DRL electrode was connected to the
3. Experiments and Results ear lobe of the subject. The subject was again asked to
remain seated and stay still to avoid movement
To aid the experiment, a testing program was built artefacts. Figures 5 and 6 depict the EEG signal during
(using LabView 8.2, National Instruments, USA) in eyes-opened and eyes-closure, respectively. In Figure
order to display the received signal at the receiver end. 5, it was noted that an eye blink would cause an artifact
The development prototype was tested on a human in the EEG waveform. In contrast, the EEG reading
subject. Firstly, the prototype was used to capture ECG with the eyes closed in Figure 6 showed fewer artifacts.

Figure 4 ECG reading from the EEG prototype

Figure 5 EEG reading from FP1 and FP2 with eyes opened – blink artifact

61

Authorized licensed use limited to: University of Witwatersrand. Downloaded on February 23,2024 at 13:42:51 UTC from IEEE Xplore. Restrictions apply.
 Figure 6 EEG reading from FP1 and FP2 with eyes closed

4. Discussion and Future Work wirelessly transmitting the signal to a computer. Future
work is required to implement a fatigue algorithm [8,
The area of safety in transportation has received 9] to detect fatigue from the EEG recording system
attention due to a high number of accidents caused by described. This simple EEG system will need to be
driver fatigue [17]. In the train driving environment, miniaturized to fit into a headband or a cap, as shown
human errors caused by fatigue contributes to about in Figure 7. Such a design implementation will make it
75% of accidents [18, 19], and research has shown that suitable for application in train drivers.
driving while fatigued is just as dangerous as driving Lastly, the next prototype will be designed to meet
with a blood alcohol concentration of 0.05-0.1% [20]. the safety requirements according to the IEC 60601-
Train drivers are required to work in a 24-hour 1:1988 (safety standard of medical electrical
irregular and rotating shifts, including weekends [21], equipment).
and irregular shifts are related to increased fatigue level
and have been correlated to higher risk of fatigue-
related accidents [22]. Therefore, there is a need for an
automated and non-intrusive fatigue countermeasure
device.
Electroencephalography has been shown to be a
reliable fatigue indicator [10]. However, the
complexity of an EEG system makes it difficult to
deploy it as a fatigue countermeasure device in the real
train driving environment. Simple and wireless EEG
device needs to be developed to enable the use of EEG
as a fatigue countermeasure device.
The current paper proposes the development of a
simple single-channel EEG device that transmits
wirelessly to a computer. Bipolar montage has been Figure 7 Proposed future EEG headband
used for the analog front-end, by amplifying the
difference of the EEG signals from the left and the 5. Acknowledgement
right active electrodes, and the DRL electrode has been
implemented to reduce the common-mode noise that The research was supported by an ARC Linkage
are present at the active electrodes. The eZ430-RF2500 grant Australia (LP0560886) and by SENSATION
development tool has been utilized to transmit EEG Integrated Project (FP6-507231) co-funded by the
signals wirelessly. Sixth Framework Programme of the European
The simulation result of the device shows that noise Commission under the Information Society
is present in the ECG and EEG recording (Figures 4 to Technologies priority.
6). Electrical noise, such as 50 Hz interference, has
been filtered by the high- and low-pass filters
6. References
implemented in the design. The noise that is captured
in the recording can be due to insufficient conducting
gel applied to the electrodes before being attached to [1] L. F. Haas, "Hans Berger (1873-1941), Richard
the skin, or subject’s involuntary movements. Caton (1842-1926), and electroencephalography.
The current EEG system has successfully performed (Neurological Stamp)," Journal of Neurology,
the function of acquiring human EEG signal and Neurosurgery and Psychiatry, vol. 74, p. 9, 2003.

62

Authorized licensed use limited to: University of Witwatersrand. Downloaded on February 23,2024 at 13:42:51 UTC from IEEE Xplore. Restrictions apply.
 [2] J. J. Carr and J. M. Brown, Introduction to <http://focus.ti.com/docs/toolsw/folders/print/simpl
Biomedical Equipment Technology, 4th ed. New iciti.html>
Jersey, USA: Prentice Hall, 2001. [17] D. F. Dinges, "An overview of sleepiness and
[3] S. K. L. Lal and A. Craig, "Driver Fatigue: accidents," Journal of Sleep Research, vol. 4, pp. 4-
Electroencephalography and Psychological 14, 1995.
Assessment," Psychophysiology, vol. 39, pp. 313- [18] G. J. S. Wilde and J. F. Stinson, "The monitoring of
321, 2002. vigilance in locomotive engineers," Accident
[4] T. Åkerstedt, G. Kecklund, and A. Knutsson, Analysis & Prevention, vol. 15, pp. 87-93, 1983.
"Manifest Sleepiness and the Spectral Content of [19] G. D. Edkins and C. M. Pollock, "The influence of
the EEG during Shift Work," Sleep, vol. 14, pp. sustained attention on Railway accidents," Accident
221-225, 1991. Analysis & Prevention, vol. 29, pp. 533-539, 1997.
[5] H. J. Eoh, M. K. Chung, and S.-H. Kim, [20] G. D. Roach, J. Dorrian, A. Fletcher, and D.
"Electroencephalographic study of drowsiness in Dawson, "Comparing the effects of fatigue and
simulated driving with sleep deprivation," alcohol consumption on locomotive engineers'
International Journal of Industrial Ergonomics, performance in a rail simulator," Journal of Human
vol. 35, pp. 307-320, 2005. Ergology, vol. 30, pp. 125-130, 2001.
[6] G. Kecklund and T. Åkerstedt, "Sleepiness in long [21] RailCorp, "Careers in Service Delivery," viewed 26
distance truck driving: an ambulatory EEG study of March 2008,
night driving," Ergonomics, vol. 36, pp. 1007- <http://www.railcorp.info/careers/careers_in_servic
1017, 1993. e_delivery>
[7] O. G. Okogbaa, R. L. Shell, and D. Filipusic, "On [22] T. Åkerstedt and P. M. Nilsson, "Sleep as
the investigation of the neurophysiological restitution: an introduction," Journal of Internal
correlates of knowledge worker mental fatigue Medicine, vol. 254, pp. 6-12, 2003.
using the EEG signal," Applied Ergonomics, vol.
25, pp. 355-365, 1994.
[8] S. K. L. Lal, A. Craig, P. Boord, L. Kirkup, and H.
Nguyen, "Development of an algorithm for an
EEG-based driver fatigue countermeasure," Journal
of Safety Research, vol. 34, pp. 321-328, 2003.
[9] B. T. Jap, S. Lal, P. Fischer, and E. Bekiaris,
"Using Spectral Analysis to Extract Frequency
Components from Electroencephalography:
Application for Fatigue Countermeasure in Train
Drivers," in The 2nd International Conference on
Wireless Broadband and Ultra Wideband
Communications, 2007. AusWireless 2007 Sydney,
Australia, 2007.
[10] P. Artaud, S. Planque, C. Lavergne, H. Cara, P. de
Lepine, C. Tarriere, and B. Gueguen, "An on-board
system for detecting lapses of alertness in car
driving," in 14th E.S.V. Conference, Session 2 -
Intelligent Vehicle Highway System and Human
Factors, Munich, Germany, 1994.
[11] A. J. Rowan and E. Tolunsky, Primer of EEG: With
a Mini-Atlas. USA: Elsevier Science, 2003.
[12] J. Malmivuo and R. Plonsey, Bioelectromagnetism:
Principles and Applications of Bioelectric and
Biomagnetic Fields: Oxford University Press,
USA, 1995.
[13] B. B. Winter and J. G. Webster, "Reduction of
Interference Due to Common Mode Voltage in
Biopotential Amplifiers," IEEE Transactions on
Bioimedical Engineering, vol. 30, pp. 58-62, 1983.
[14] B. B. Winter and J. G. Webster, "Driven-Right-Leg
Circuit Design," IEEE Transactions on
Bioimedical Engineering, vol. 30, pp. 62-66, 1983.
[15] D. J. Comer and D. T. Comer, Advanced Electronic
Circuit Design. Great Britain: Wiley, 2003.
[16] Texas Instrument, "SimpliciTI Network Protocol,"
viewed 07 May 2008,

63

Authorized licensed use limited to: University of Witwatersrand. Downloaded on February 23,2024 at 13:42:51 UTC from IEEE Xplore. Restrictions apply.