3rd International Conference on Advancements in Electronics and Power Engineering (ICAEPE'2013) January 8-9, 2013 Kuala Lumpur (Malaysia)
Thought Controlled Wheelchair Using EEG
Acquisition Device
Rajesh Kannan. Megalingam, Athul. Asokan Thulasi, Rithun. Raj Krishna, Manoj. Katta Venkata,
Ajithesh. Gupta B V, Tatikonda. Uday Dutt
Abstract— With the advancements in technology and healthcare
facilities, the number of senior citizens has increased and thus the
number of elderly who find it difficult to walk. Hence there is a need
for designing a wheelchair that is user friendly and involves fewer
complexities. In this context, we propose a thought controlled
wheelchair, which uses the captured signals from the brain and
process it to control the wheelchair. This wheelchair can also be used
by the physically challenged who depend on others for locomotion.
Rehabilitation centers at hospitals can also make use of this
wheelchair. In this paper, we explain the design and analysis of the
thought-controlled wheelchair. In addition, we present some of the
experiments that were carried out and the corresponding results in
this paper.
II. PROBLEM STATEMENT
There are nearly 7 billion people our world as in [6] out of
which there are a considerably large number of people who
has at least one affliction. One of the most distressing one that
anybody can have is paralysis of body parts. Paralysis can
make the affected to seriously depend on others even for
activities of daily living. This paper intends to unfold a
solution for this problem. A wheelchair is designed using BCI
technique so that a person with any extent of disability can
operate the wheel chair to attain self-independence at least in
activities of daily life.
Keywords—BCI, EEG, P300, Wheelchair.
III. RELATED WORKS
The paper cited in [2] presents how to control a cursor in
2D using EEG based BCI. The movements of the cursor on
the computer screen are controlled based on the EEG scalp
potentials. The EEG signal is mapped from the user head to
the 2D cursor control. The horizontal motion is based on the
mu beta rhythm and vertical motion is based on P300 potential
respectively. The user can move the cursor horizontally to the
right just by imagining his right hand motion and the same
way the left hand motion imagination is mapped to move the
cursor left. The user can move the cursor vertically upwards
by focusing on one of the three UP buttons on the monitor and
similarly downwards by focusing on the either of the three
DOWN buttons. This same technique can be deployed in
designing a BCI based wheelchair.
The study cited in [3] focus on the target selection of BCI
2D cursor using hybrid feature. As in most of the applications,
2D cursor control is used as a sequential control, which
involves cursor control and target selection. The paper focuses
on the latter, the target selection of the cursor using a hybrid
feature from P300 potential and motor imagery. It has got an
advantage of better performance compared to others,
especially in case of subjects with poor performance.
Paper [4] proposes a wheelchair control methodology in
which the motion of the chair is controlled based on the inputs
from user (P300 brain signals), data obtained from obstacle
avoidance sensors and by using localization algorithm. The
brain signal processing is not fast enough to control
wheelchair in real time so these signals are used to make long
term decisions like choosing a distant destinations and
complex obstacle avoidance.
I. INTRODUCTION
B
RAIN Computer
Interface (BCI) is a technique that
provides direct interface between the human brain and the
computer. BCI techniques are broadly classified into
invasive and non-invasive techniques. Non-invasive
techniques are becoming more popular and more research is
being done on this topic. There are various non-invasive BCI
techniques such as EEG, Electro-Oculography. EEG technique
deploys an electrode cap that is placed on the user’s scalp for
the acquisition of the EEG signal, which relates the scalp
potential differences to various complex actions. Classification
of the EEG signal has been made into several bands like alpha,
beta, delta, theta and mu suppression, each corresponding to
various states of being like relaxing, ranging over 8-14 Hz;
concentrating, ranging over 13-30 Hz; deep sleep, from 0-4
Hz; meditating from 4-8 Hz; moving your hands or legs or just
by imagining these motor actions respectively. As it is being
non-invasive in nature, it has an advantage over traditional
BMI, not being hazardous to health. With the advent of
technology the EEG acquisition devices are made more
compact, handy and wireless. Using the above mentioned
technique, a simple thought controlled wheelchair system has
been proposed in this paper. A section that briefly explains the
various blocks included in the system is also added in this
paper.
Rajesh Kannan. Megalingam is with Amrita Vishwa Vidyapeetham
University, Kerala, India (Ph: +91 476 2802305; e-mail:
[email protected] ).
Athul Asokan. Thulasi* ([email protected]), Rithun. Raj Krishna*
([email protected]), Manoj. Katta Venkata* ([email protected]),
Ajithesh. Gupta B V* ([email protected]), Uday. Dutt*
([email protected])
* All with Amrita Vishwa Vidyapeetham University, Kerala, India.
207
3rd International Conference on Advancements in Electronics and Power Engineering (ICAEPE'2013) January 8-9, 2013 Kuala Lumpur (Malaysia)
IV. WHEELCHAIR SYSTEM DESIGN
filters. The high pass filter, removes the noise in the signal.
Low pass filter extracts the signal frequencies of interest. As
DC power supplies are used, one common problem to
encounter is the 60 Hz power line signal. This 60 Hz power
line signal will distort the EEG scalp potentials. Integration of
a notch filter will filter out this undesirable power line signal.
This block has been simulated and its details is been explained
in the later sections of this paper. One of the disadvantages
that can be considered about this system is that users with
Slow Cortical Potentials (SCPs) cannot use this system. SCP
is a state where the motor reflex of the patient is very slow and
hence there is a delay between the mechanical input to the
system and thereby a delay in the response of the system.
Thought capture and Identification Module
Electrodes
non-invasively
placed on head
Interfacing circuitry
Amplification, Filtering, Digitizing
Signal Processing Unit
(Microprocessor)
Microcontroller for Generating PWM and Direction signals
Motor controller Module
PMDC
RIGHT
MOTOR
PMDC
LEFT
MOTOR
Fig.1Block Diagram of the Wheelchair System
The system is broadly divided into four blocks, thought
acquisition block, thought transmission block and thought
processing block and motor control block each of which aims
at acquisition of the EEG signal from user scalp and processing
it for controlling a wheelchair. EEG scalp potentials obtained
are amplified, digitized and transmitted to a processor and after
processing the output of the processed signals are used to
control the wheelchair. The four main blocks involved in the
wheelchair system are briefly discussed below.
Fig.3 Schematic Outlining various stages involved in thought
acquisition block
B. Thought Transmission block
This block focuses on the transmission of the acquired EEG
signal (thought) to a processor. It consists of 12-bit A/D
converter for digitizing the EEG signal. The ATMega644
microcontroller is used for UART transmission. This
microcontroller is having a 10-bit A/D converter peripheral,
which cannot be used because of lesser resolutions. That is the
reason we are using an external 12-bit A/D converter. The
FTDI USB RS232 is an opto-coupler used for electrical
isolation to prevent electrical hazards during the transmission
of the digitized EEG signal to a DSP as shown in Fig.3. It is
also used for converting the RS232 signals to USB signals so
that this setup can be interfaced to a processor.
Fig.2 The Wheelchair Setup Used for the Implementation
A. Thought Acquisition Block
This block primarily targets at the careful extraction of the
EEG signal from the user scalp. It is made up of different
blocks such as instrumentation amplifier, operational
amplifier, high pass, low pass and notch filters. The purpose of
the instrumentation amplifier is to extract the EEG signal. The
extracted EEG signal is passed through the operational
amplifier block for proper amplification, as shown in Fig.2. It
is then, passed through the high pass, low pass and notch
Fig.4 Schematic outlining various stages involved in thought
transmission block
C. Thought Processing Block
This block consists of processor which processes the signal
that is transmitted from the thought acquisition block to
processor through the thought transmission block.
208
3rd International Conference on Advancements in Electronics and Power Engineering (ICAEPE'2013) January 8-9, 2013 Kuala Lumpur (Malaysia)
D.Wheelchair Control System
The wheelchair controller has the functionality of
controlling the direction and speed of the wheelchair based on
the output bits obtained from the signal processing block.
Hercules motor driver is used in between the Arduino and the
Permanent Magnet Direct Current (PMDC) motors. This
module is used to control the motors in real time wheelchair.
Two motors connected to the rear wheels of the wheelchair
have to be controlled to define its motion. The PMDC motors
can be controlled using Pulse Width Modulation (PWM)
techniques that can be generated using the Arduino Board.
The power wheelchair can be directionally controlled using
suitable motor driver. Super Hercules 9V – 24V, 15A motor
drivers are used to control the motors. These motor drivers can
take up to 30A peak current load and can be operated up to 10
KHz PWM. The direction of the movement can be controlled
using a separate pin on the motor driver which is used as
direction pin. Hence the wheelchair can also be made to move
in the reverse direction with the help of the direction pin on
the motor controller. The control signals which include PWM
and direction signals are given by Arduino board to Hercules
motor driver. The motor driver acts as a switch between the
input 5V signal and required 24V output signals. The
wheelchair moves at considerably good speed when it is given
15 volts. So to avoid jerk or a sudden unwanted vibration
when the wheel chair starts or stops or changes direction, the
PWM signal is incremented or decremented according to the
needs in linear steps rather that a sudden transition.
The gain of instrumentation amplifier is given by G =
(49.4kohm/Rg) + 1. It is calibrated to have a gain of about 23
(using Rg = R1).
V. THOUGHT PROCESSOR-WHEELCHAIR CONTROL SYSTEM
INTERFACE
The gain of the operational amplifier, used for post
amplification purpose, is given by (-R2/R1). It is designed to
have a gain of around 65 (using appropriate R2 and R3). Thus
the overall circuit gain is about 1500, approximately. Two
First Order high pass filters are integrated; one each after
instrumentation amplifier and operational Amplifier (OPAMP). A first order low pass filter is integrated along with the
operational amplifier. We have used AD620 as an
instrumentation amplifier; it is used for pre amplification
purpose, with a nominal gain of about 23. This stage of
amplification is mainly aimed at the EEG signal extraction,
whose value ranges in micro volts, without being distorted.
The instrumentation amplifier provides high input impedance
to the EEG electrodes. The idea behind infinite input
impedance is to assure zero attenuation of EEG signal across
electrodes. The voltage follower along with a virtual ground
set up with CA3140 collectively known as virtual ground.
Virtual ground is coupled with the instrumentation amplifier
to provide a DC offset. Virtual grounds are used to cope up
with the problem of dual power supplies. The DC offset
provided by the virtual ground to the pre amplification stage is
treated as the reference with respect to which the single power
supply is seen as a dual power supply.
Fig.5 Diagrammatic representation
corresponding gesture
The thought processing block detects the brain signals for
focusing eye to the left, right, up or down and uses it to
calculate a cursor location (co-ordinate of a point) by using the
algorithm in [2]. This coordinate is the used to finds out the
difference between newly calculated cursor location and
present cursor location. The difference in the coordinate
values of these locations is then used to find the angle between
the line joining newly calculated location and present location
with x axis as reference axis. The thought processing block
outputs 3 bits of data, out of which bit 1 is set if the calculated
change in abscissa and ordinate are both zero (indicative of
stop gesture).If one the change in value is non-zero then bit 1
is cleared and the calculated angle is used to find the other two
bits as explained in Fig. 3. The 3 bits of data which defines the
identified gesture are shown in table 1.
VI. EXPERIMENT
The simulation of the EEG amplifier used in thought
acquisition module is done using Multisim Simulation tool.
The EEG Amplifier, shown in Fig.6 is designed based on the
circuit shown in [1]. It consists of an instrumentation
amplifier, operational amplifier and a voltage follower with a
virtual ground set up.
209
of
calculated
angle
SL NO.
1
TABLE I
THE OUTPUT OF THOUGHT PROCESSING BLOCK
IDENTIFIED GESTURE
BIT 1
BIT 1
BIT 3
Stop
0
X
X
2
Left
1
0
0
3
Right
1
0
1
4
Forward
1
1
0
5
Backward
1
0
0
and
3rd International Conference on Advancements in Electronics and Power Engineering (ICAEPE'2013) January 8-9, 2013 Kuala Lumpur (Malaysia)
1MΩ
50%
Key=A
5V
R3
2.2kΩ
1
VCC
1
U1
8
3
3
C1
11R2
6
30mVrms
60 Hz
0°
1uF
2
4
C2
VCC
8
12
1
5
U2
2
4
0
R10
R7
6
R9
1MΩ
8
1
5
U4
3
3
AD620AN
5
7
6.8nF
5V7
2.2kΩ
VCC
5V
VCC
4
1MΩ
2
7
VCC
5V
VCC
R6
1MΩ
R1
VCC
V1
6
10kΩ
CA3140E
9.8MΩ
R11
9
18
9.8MΩ
C6
540pF
15
16
6
2
4
8
CA3130E
17
VCC
5V
VCC
C3
440nF
7
0
R12
4.9MΩ
0
1
5
U5
R13
3
6
C4
VCC
VCC
5V
5V
VCC
270pF
7
C5
270pF
2
0
4
8
14
CA3130E
50kΩ
50%
Key=A
VCC
R4
15kΩ
7
1
5
U3
10
3
13
6
2
R8
15kΩ
4
8
CA3140E
0
Fig.6.Schematic of the circuit simulated in the Multisim.
CA3140 is used as an OP-AMP, for post amplification
purposes, with an approximate gain of around 65. This stage
of amplification is aimed at maximum gain.A high pass filter
is coupled in between pre and post amplification stages and
another is coupled after post amplification stage, both with a
cut-off 0.13 Hz. Each of these high pass filters prevents the
low frequency noise being carried to the later stage. EEG
signals of interest range to a maximum of 48 Hz. A low pass
filter with a bandwidth of 48 Hz is integrated along with the
OP-AMP to extract the frequencies of interest. The next part
of the simulation involves the Proteus simulation tool for
building and testing the microcontroller Part of the system as
show in Fig. 3. The circuit shown in the Fig.8 is the simulation
schematics of the microcontroller for ADC conversion and
UART transmission. ATmega644 microcontroller used. It is
an 8-bit Atmel Microcontroller with 64K Bytes in-System
Programmable Flash and 10-bit ADC. For simulation purpose,
we have used the internal A/D peripheral block instead of an
external 12-bit ADC. The input sine wave is given through
analog pin A1 and expected output is taken through digital pin
D1. The digital output is verified through the Oscilloscope,
which is connected to the microcontroller as shown in the
Fig.6.
Fig.7 Schematic of the circuit simulated for UART transmssion
Frequencies outside the pass band got a gain less than one
third of the maximum gain, 1500, of the amplifier. A
maximum gain of about 1500 is observed for central
frequencies of the pass band. Fig. 8a shows the plot without
the notch filter. The small lobe on the right side of the bode
plot in Fig. 8b, refers to the 60Hz notch of the notch filter.
The waveforms (1) and (2) shown in Fig.9a are the input
and output waveforms of the first stage of amplification in the
circuit shown earlier in Fig. 6.For an input signal (1) of a
frequency well within the pass band of the amplifier, the
output signal after first stage of amplification is observed to
have a gain of about 23. This output waveform is observed
with a dc offset of 2.5 volts. The (3) and (4) output waveforms
shown in Fig. 7b are the input and output waveforms of the
post amplification stage of the same circuit.
VII. RESULT
The bode plots of the amplifier setup are shown in Fig 8a
and 8b. These plots show the frequency response of the circuit
in Fig.6, with its lower cut-off, 4Hz and upper cut-off, 48Hz.
Frequencies in between lower and upper cut-off refer to the
pass band of the amplifier.
210
3rd International Conference on Advancements in Electronics and Power Engineering (ICAEPE'2013) January 8-9, 2013 Kuala Lumpur (Malaysia)
Fig.8a Frequency response of the amplifier, without notch filter
Fig.10 Output waveform of the UART transmission
ACKNOWLEDGMENT
We gratefully acknowledge the Almighty GOD who gave
us strength and health to successfully complete this venture.
The authors wish to thank Amrita Vishwa Vidyapeetham, in
particular the VLSI lab and Electronics lab for access in
carrying out this research work.
REFERENCES
[1]
Fig.8b Frequency response of the amplifier, with the notch filter
[2]
[3]
[4]
[5]
[6]
Fig.9a Output waveform after first stage of amplification
Fig.9b Output waveform after second stage of amplification
For the input waveform (3) within the pass band frequency
along with a DC offset, the post amplification stage is
observed to have a gain of around 65.
211
http://people.ece.cornell.edu/land/courses.ece4760/FinalProjects/
s2012/cwm55/cwm55mj294/index.html
Yuanqing Li, Chuanchu Wang, Haihong Zhang and Cuntai Guan, “An
EEG-based BCI System for 2D Cursor Control,” IEEE International
Joint Conference on Neural Networks, 2008.
Jinyi Long, Yuanqing Li*, Tianyou Yu, and Zhenghui Gu, “Target
Selection With Hybrid Feature for BCI-Based 2-D Cursor Control,”
IEEE Trans. Biomedical Engineering. January 2012.
Brice Rebsamen, Chee Leong Teo, Qiang Zeng, and Marcelo H.Ang Jr.,
Etienne Burdet, Cuntai Guan and Haihong Zhang, Christian
Laugier, “Controlling a wheelchair indoors using thought,” published by
IEEE Computer Society 2007.
www.atmel.com/Images/doc2593.pdf - United States
http://www.100people.org/statistics_100stats.php?section=statistic
3rd International Conference on Advancements in Electronics and Power Engineering (ICAEPE'2013) January 8-9, 2013 Kuala Lumpur (Malaysia)
Measurement of Elder Health Parameters and
the Gadget Designs for Continuous Monitoring
Rajesh Kannan Megalingam, Goutham Pocklassery, Galla Mourya, Ragavendra M Prabhu, Vivek Jayakrishnan
Vazhoth Kanhiroth, Rocky Katoch, Pranav Sreedharan Veliyara
Abstract—Improving the quality of life for the elderly persons
and giving them the proper care at the right time is the
responsibility of the younger generation. Due to lack of awareness
on proper elder care, unavoidable busy schedule etc. the elderly
population is seem to be quite ignored. A simple, compact and
user-friendlyelectronic gadget for continuous monitoring of elder
health parameters is the need of the hour. Day by day the menace
of weakening health and chances of skin related problems, bed
sores etc are becoming critical in case of bed ridden patients. This
paper analyses the old age diseases and the parameters to be
monitored. Moreover, a detailed study is provided on the design of
the gadget – placement of sensors, wired/wireless technology,
Android applications etc. Based on the analysis a simple,
comfortable, low cost gadget to monitor elder health and report
any critical condition is proposed.
II. PROBLEM DEFINITION
Projected increase in both the absolute and relative size
of the elderly population in countries all over the world is a
subject of growing concern. The proportion of elderly
persons in the population of India rose from 7.5% in 2001 to
8.5% in 2010. The Indian aged population is currently the
second largest in the world. As population ages, more
demand is placed on caring for the elderly. The absolute
number of the elderly (above 60 years old) population in
India is projected to increase from 77 million in 2001 to 137
million by 2021.
Many of the elderly have various degrees of disabilities.
They are often dependent on others for their activities of
daily living. Some of them remain bedridden due to various
causes. This means that a ‘Care Team’ (including family,
friends, nurses and other professionals) will likely be
working together.The healthcare provider would want a
record of body temperature, pulse, and respiration. In case
of bedridden elders we have to make sure that the position
of the elder is changed periodically so as to reduce the
possibility of bed sore. Clearly, in view of such a
demographic trend, medical assistance to the rising number
of dependent elderly is a major problem that many countries
are facing now. In this scenario, a constant and reliable
assistive technology, which can cater the needs of these
homebound elders, is the need of the hour.Assistive
technology devices are helpful products that improve a
person’s ability to live and function independently. These
types of assistive gadgets can also prove to be a boon to the
caregivers, since their workload can be drastically reduced.
Keywords— Elder care, home-bound patient, Assistive
technology.
I. INTRODUCTION
P
ROVIDING quality and timely heath care for the
elderly has always been the area of concern for the
younger generation. Employment, work stress and other
family issues have always convoluted this problem. Though
old age homes and elder care centers have emerged as a
possible solution to this problem, they are rather business
oriented and quiet expensive. Moreover elderly people does
not prefer custodial care and want to be at home where they
are not detached from their family, friends and society.
Analyzing the diseases of the elderly, we can see that
many of the diseases that haunt them are chronic in nature.
In many a case the detection and cure of these diseases
require continuous monitoring of the physical parameters
due to their special nature of occurrence. Another major
issue affecting bed-ridden elderly is the case of bed-sore and
unintentional fall as in [1]. Though there are many products
available in the market today for monitoring patients as in
[2], we can see that many of them only measure individual
parameters like heart rate, body temperature etc. In addition,
many of them do not address the problem of tilt or fall.
III. RELATED WORKS
Some of the elder care systems as mentioned in [2]
monitor activities of the elders in their home. They embed
a video system in the living environment of elders and
continuously monitor their activities at home. However,
this system doesn’t measure any of the vital parameters of
the elderly patient. Measuring the vital parameters is
inevitable if the elder person suffers from any sort of heart
ailments, which are very common in individuals aged above
60[11]. In [1] mobile devices like Caalyx (Complete
Ambient Assisted Living Experiment) which can measure
vital signs like ECG, pulse, Blood pressure, Movement and
Fall detection. However, the design we have proposed can
monitor vital parameters and fall detection along with tilt
monitoring for the bed-ridden patients to monitor any case
of bedsore. Some devices as in [1] monitor only fall
detection for the elderly patients based on the sensor
readings from accelerometers and microphones attached to
the body of the patients. The system proposed in [4] is
applicable to patients
Rajesh Kannan Megalingam is with Amrita Vishwa Vidyapeetham
University, Kerala, India (Ph: +91 476 2802305, [email protected]).
Emails: Goutham Pocklassery* ([email protected]), Galla Mourya*
([email protected]),
Ragavendra
M
Prabhu*
([email protected]),
Vivek
Jayakrishnan
Vazhoth
Kanhiroth*
([email protected]),
Rocky
Katoch*
([email protected]),
Pranav
Sreedharan
Veliyara*
([email protected]) All with Amrita Vishwa Vidyapeetham
University, Kerala, India .
212