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

net/publication/350835886

Design and Implementation of EEG Based Home Appliance Control System

Conference Paper · April 2021

DOI: 10.1109/ICICT4SD50815.2021.9396982

CITATIONS READS

11 684

4 authors:

Tafhim Bin Nasir Md. Abdul Mukit Lalin

Islamic University of Technology American International University-Bangladesh
6 PUBLICATIONS 24 CITATIONS 2 PUBLICATIONS 23 CITATIONS

SEE PROFILE SEE PROFILE

Niaz Kabir Md Razaul Karim

American International University-Bangladesh Bangladesh Agricultural University
3 PUBLICATIONS 24 CITATIONS 49 PUBLICATIONS 191 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Tafhim Bin Nasir on 01 March 2022.

The user has requested enhancement of the downloaded file.

 2021 International Conference on Information and Communication Technology for Sustainable Development (ICICT4SD),
27-28 February, Dhaka

Design and Implementation of EEG Based Home

Appliance Control System
Tafhim Bin Nasir 1, Md. Abdul Mukit Lalin 2, Kabir Niaz 3, Md. Razaul Karim 4
Department of Electrical & Electronic Engineering (EEE)
American International University-Bangladesh (AIUB)
Dhaka, Bangladesh
[email protected], [email protected], [email protected] ,

[email protected]

Abstract— This paper represents the design and electrodes are placed into the brain in which a surgical
implementation of the Home Appliance control system using the operation is needed at the implementation time. Whereas, in
Human brain wave. Brain-computer interfacing (BCI) the Non-invasive type, the electrode is placed over the scalp
technology is followed for developing the system. Neurosky area, which means there is no need for surgical operation [5].
Mindwave Mobile 2 headset is used for collecting the desired Non-invasive is a technique with less amount of risk and very
brain data. This headset collects brain data and sends the data cost-effective for practical implementation [5].
wirelessly to the main PCB circuit board using a Bluetooth
module. The headset gives the output of attention level and Brain wave activity or electroencephalogram (EEG) is
number of eye blink based on EEG waves. According to the relatively small and always measured in microvolts [6]. EEG
attention level and eye blinks, the system controlled Light, Fan, signal frequency varies according to the human’s mental
TV, Buzzer. This system has two controlling modes with four condition. For this, the EEG signal is classified with different
different attention levels. In the first mode, the system can brand frequency. Infra-Low wave (<.5 Hz) [4], Delta wave
control light, fan, television, and buzzer. In this mode, the user (0.5-3Hz), theta wave (3-8 Hz), alpha wave (8-12 Hz), beta
can only power on or Power off these Appliances. Whereas in wave (12-38 Hz), gamma wave (38- 42 Hz) are main
the second mode, the system can control television volume and frequencies of human EEG wave [6]. Delta wave and theta
channel. This system is mainly designed to help Physically wave are usually generated at dreamless sleep and the most
Challenged People who can not move their body parts and are profound meditation [4]. Alpha wave is generated when the
incompetent in speaking. Simplicity in design architecture and
human brain is quietly following the thought of thinking or
cost-effectiveness makes this system very much applicable in
real-life usage.
calculating anything [4]. Beta brain wave shows dominance
when human is in regular activities such as busy in some task,
Keywords— Brain, Double Eye Blink, EEG, BCI, Neurosky problem-solving, judgment, focused on mental activity. Beta
brain wave has been divided into three bands. Low Beta (12-
Brain Wave, Attention Level, Home Automation 15 Hz), Midrange Beta (15-22 Hz), High Beta (22-38 Hz) are
the classification of beta waves [4].
I. INTRODUCTION
According to the world health organization, it is observed This paper is all about showing the new procedure of
that almost one billion or 15% of the world population designing and implementation of brain wave controlled Home
experience some form of physical disability. Moreover, this Appliance control system. Brain data is acquired using the
rate is higher for developing countries, which is very alarming non-invasive type technique. NeuroSky Mind wave Mobile 2
[1]. Out of this number, a significant portion is thoroughly headset is used in acquiring EEG data. The overall design of
unfit to do any physical movement. These physically disabled this system architecture is unique compare to other related
people face difficulties in meeting the demand of the present research works. This work would be beneficial for physically
world. For this, in the recent past, a vast amount of research disabled people in controlling different types of the domestic
work has been gone through to aid physically disabled people appliances.
so that they could survive this competitive modern world. II. LITERATURE REVIEW
Among these several research, Brain-Computer Interface
technology (BCI) is one of the major options to get out of these A lot of research work has gone through based on Brain-
difficulties. Computer Interface technology. Not only that, previously, a
significant amount of research work is accomplished based on
The principal methodology of Brain-Computer Interface the Home Appliance control system. In some paper, the
technology (BCI) is to use brain activity to control an external devices are controlled based on the voice of human [7]
device [2]. A human brain contains approximately 100 billion whereas, in some paper, it is seen that devices are controlled
neurons [3]. In our activity, behavior, and thinking, these using IOT technique [8]. Here, in this paper, the devices are
neuron masses are always communicating with each other, controlled by the brain wave. Brain data is also could be
which results in electrical pulses. Brain waves are produced acquired by different EEG headsets. According to the
by these electrical pulses generated in neurons [4]. Brain- selection of the EEG headset, the overall control algorithm
Computer Interface (BCI) technology uses these electrical varies. Here in this work, the system architecture is designed
pulses to control external devices and other things. keeping in mind of using the NeuroSky headset module.
Brain wave data could be acquired from the brain using
the electrode. With the usage of the electrode, Invasive and
Non-Invasive are the two varieties of brain wave techniques
present in the current time. In the Invasive technique, Sensor

978-0-7381-4418-4/21/$31.00 ©2021 IEEE

 III. SYSTEM DESIGN TGAM1 chip is consists of one processor and many filters.
These filters are used for filtering noise from the acquired
A. NeuroSky Headset
sample signal and separate different brain wave types. When
The brain data is acquired with NeuroSky Mindwave Mobile an EEG signal from the brain comes into the TGAM1 chip, it
2 headset. At the time of controlling the Home Appliance, is filtered by a low pass filter for separating all the wave
this headset has to be placed on the user's head. This headset signals with more than 100Hz frequency [13]. Extra wave
has an EEG electrode, which is placed inside the sensor arm. signal of frequency 50Hz or 60Hz is removed according to
This EEG electrode acquires all kinds of raw EEG data from the M pin configuration. By this, the range of signal
brain activity. Also, this headset has an ear clip-on left side. frequency comes between 0.3 Hz to 100 Hz [13]. This time
This ear clip is used as ground or references [9]. filter would divide the sample according to different brain
wave types. Finally, this signal is sampled with 512 Hz and
sent to ADC with the 12-bit resolution. It is possible to get
digital values for each brain wave type using this ADC [5].
For many digital values in a single brain wave, the processor
has calculated power Fast Fourier Transformation. With a
configuration pin, the processor gives digital FFT values for
each brain wave serially [13]. The baud rate is by default set
at 9600 with the expected output in this process [13]. The
NeuroSky headset follows the ThinkGear Communication
Fig. 1. NeuroSky MindWave headset module Protocol. It provides data values as ThinkGear Packets from
a ThinkGear module to an arbitrary receiver. By using some
The EEG electrode is placed on the Fp1 position according to hex code, the brain data can be extracted from this packet
the 10-20 system positioning of the electrode method [10]. [12].
Whereas the ear clip consists of grounding and the reference
B. Attention level calculation
sensor, it is placed in the A1 position. This module has to be
powered up with a AAA battery [9]. The EEG data acquired from the brain is sampled at a
sampling rate of 512 Hz. A Fast Fourier Transform (FFT)
technique is followed for transforming a segment with 512
sampling points to the frequency domain. Two hundred fifty-
six points are overlapped with the previous segment for
sliding the segment. Presuming, F(n) is the Fast Fourier
Transform the result of a segment where n =1,2,3,
4……...512; then the equation of Power Spectral Density
(PSD) is:
∗
P(n) = (1)
Fig. 2. Positioning Electrode According to the human head Here, the value of N is 512 and F*(n) represents the conjugate
function of F(n).
This headset can give the output of the human's brain's
attention and meditation level with the eSense meter [9]. The
By following the EEG signal waveband distribution, the
eSense meter gives the output of attention and meditation by
energy value of different brain waves is assumed. However,
calculating a wide spectrum of brain waves in frequency and
according to research [14], it is proved that α, β, δ, θ are the
time domain. It is seen that the attention level, which is used
four main types of brain wave which is most related to human
in this system, has shown more significance on the beta wave
mental states [14]. The features can be defined as follows:
of the human brain [11].
The NeuroSky headset uses a TGAM1 module for processing
=∑ (2)
brain signals. TGAM1 chip has an integrated configuration
pad, which is used to change the default settings. Baud rate =∑ (3)
and data content are the two types of default settings that can =∑ (4)
be changed. A configuration pad is placed on the backside of δ = ∑30.5 (5)
the TGAM1 chip. Gradually B0 and B1 are used for baud rate
and data content, whereas M pad is used to configure the Moreover, according to previous research, it is found that
notch filter frequency [12]. there exists an interrelation between α and β waves. The ratio
between this α and β wave is used to extract and find mental
attentiveness [14]. The equation for assessing the level of
mental attentiveness is:
!
R = !" (6)
#
Where, R indicates feature for determining whether user is
attentive.

Fig. 3. Image of showing a side view of NeuroSky EEG chip TGAM1 [12]
C. BlueSMiRF Module E. Flow chart
BlueSMiRF module is followed serial communication
method for sending and receiving data. In the time of sending
and receiving data, this module act as a data pipeline [15].
This module has TX and RX pin. Serial data that goes into
the module is received by RX pin, whereas data coming in
from Bluetooth is passed out by TX pin [15]. Two steps
process has to be followed for establishing data pipeline.

Fig. 4. BlueSMiRF module

Firstly, it is essential to connect something for sending and

receiving header data to the Bluetooth modem. Connecting
Arduino or any microcontroller with UART could be a
solution for this. Baud rate is needed to be configured so that
it can work at the same rate. The default baud rate is set in
this module is 115200 bps [15]. By configuring the baud rate,
any serial stream from 2400 to 115200bps can be passed
seamlessly from this module to other targets [16]. Finally, the
BlueSMiRF module is needed to establish a wireless
Fig. 6. Flow Chart
connection with another Bluetooth device [15].
D. Block Diagram F. Schematic view

Fig. 7. Schematic view

IV. HARDWARE IMPLEMENTATION

A. Overall Hardware Implementation and working
At the time of giving the command to the main system
Fig. 5. Block diagram
architecture, the user is needed to wear the NeuroSky headset
with proper placement. The power button of the headset is
needed to be ON mode while working. Then it sends data to
the main PCB via Bluetooth. BlueSMiRF module is
connected with both Arduino Nano. The receiver RX pin of
both Arduino is connected parallelly with the TX pin of the
BlueSMiRF module. In the meantime, the RX pin of the
Bluetooth module is only connected with the First Arduino
Nano TX pin. Four sets of attention level indicator LED are The second mode's algorithm and working principle are quite
placed parallelly with the BJT transistor. the same compared to the first control mode. The condition
for controlling television is given below:

TABLE III. CONDITION FOR TELEVISION CONTROL

Attention Number of LED Indicator Controlling
Level Eye Blink Television Menu
0-40 YELLOW Volume increase
41-60 2 BLUE Volume Decrease
61-80 GREEN Channel up
81-100 RED Channel down

V. RESULT AND ANALYSIS

A. Analytical Result
This system is designed and controlled based on EEG. So,
before implementing the overall system, it is imperative to
know whether the used headset gives the desired output or
Fig. 8. Hardware View
not. The headset gives the attention value based on EEG
Here, BJT transistors are used for switching purposes. All wave, which is ranged between 0 to 100. Also, the headset
four sets of LED blinks, according to the human attention could measure the level of blink strength. According to this
level measured by the NeuroSky headset. Each Arduino Nano level of blink strength, it counts the number of an eye blink.
is powered up with a 5V DC battery. First, Arduino Nano Before implanting the overall system, the output of attention
detects the attention level and controls attention level and eyeblink is analyzed in MATLAB, which helps to
indicator LEDs. It also sends the attention level data to the understand the output wave shape of EEG and create an
second Arduino Nano. Second Arduino Nano receives data algorithm.
sent by the first Arduino Nano and detects the number of
eyeblinks. According to the number of eyeblinks and
attention level, this Arduino controls the Home Appliances
used in this work. In the domestic load portion, the system
can control light, fan, television, and buzzer. Both light and
fan are controlled with relay, whereas television is controlled
with an IR transmitter.
B. Algorithm
This system has two modes of control. It can interchange the
modes by detecting the number of eye blinks. One control
mode is the load ON/OFF mode, where the system can
control the load's power. Then the other one is the television
volume and channel control mode. In this mode, the system Fig. 9. Measuring Attention level
can increase or decrease volume and change the television
channel. The condition for mode control is given below:

TABLE I. CONDITION FOR MODE CONTROL

Number of Eye Blink LED Indicator Control Mode
Green Load on/off
3 Blue Television Control
If the system control mode belongs to the Load on/off
portion, then it can control light, fan, television, and buzzer
either OFF to ON state or vice versa. For this, the user is
needed to reach a predefined level of attention, and then with
a double eye blink, the control could be done. Here, a double
eye blink is used for the triggering purpose. The condition for
load control is given below:

TABLE II. CONDITION FOR LOAD CONTROL

Attention Number of Eye LED Indicator Types of Load Fig. 10. Measuring number of Eyeblinks
Level Blink control
(ON/OFF) B. Practical Testing Result
0-40 YELLOW Light
41-60 BLUE Fan After implementing the overall system, some practical testing
61-80 2 GREEN Television is done to check the overall standard of the system. This
81-100 RED Buzzer system is successfully able to control light, fan and
Television. In the controlling Television portion, the practical Currently, this paper represents the course of action to control
test is successful in increasing or decreasing volume and a light, television and fan. However, in future, this system has
changing TV channels. the capability of upgrading the considerable number of a load.
As this system is designed mainly focusing on physically
disable people so it might be pervasive that they are badly in
need of the help of others in some cases.
REFERENCES
[1] WORLDBANK, “DISABILITY INCLUSION”, May 15, 2020,
Accessed on: July 10, 2020. [Online]. Available:
https://www.worldbank.org/en/topic/disability
[2] Chaudhary, U.; Birbaumer, N.; Ramos-Murguialday, A. Brain–
computer interfaces for communication andrehabilitation.Nat. Rev.
Neurol.2016,12, 513–525
[3] K. Cherry, “Comparing the Amount of Neurons in Human and
Different Animal Brains,”Verywell Mind. Accessed on: July 15, 2020.
[Online]. Available: https://www.verywellmind.com/how-many-
Fig. 11. Practical test result of controlling TV and LED Bulb neurons-are-in-the-brain-2794889.
[4] What are Brainwaves?,” What are Brainwaves ? Types of Brain waves
C. Overall Accuracy | EEG sensor and brain wave – UK. Accessed on: December 3,
2020.[Online]. Available: https://brainworksneurotherapy.com/what-
A practical test is made, which is consists of four members. are-brainwaves.
According to that test, an overall accuracy analysis is made. [5] A. A. Ghodake and S. D. Shelke, "Brain controlled home automation
In result, it is seen that most of the time, the system is system," 2016 10th International Conference on Intelligent Systems
successfully able to control the Home Appliance. In number, and Control (ISCO), Coimbatore, 2016, pp. 1-4, doi:
10.1109/ISCO.2016.7727050.
the overall accuracy rate of the system is 90.67%. The overall
[6] Bhat, Deeksha Seetharama, “Assessing Performance of Detectors of
accuracy of the system is given in the following table: High Frequency Oscillations in EEG Signals,” Ph.D. dissertation,
University of Texas at El Paso, 2018.
TABLE IV. ACCURACY ANALYSIS [7] N. bt Aripin and M. B. Othman, "Voice control of home appliances
USER Total Percentage of Accuracy using Android," 2014 Electrical Power, Electronics, Communicatons,
number of Control and Informatics Seminar (EECCIS), Malang, 2014, pp. 142-
applied 146, doi: 10.1109/EECCIS.2014.7003735.
commands Light Fan TV Overall [8] S. Dey, A. Roy and S. Das, "Home automation using Internet of
Control Control Control Accuracy Thing," 2016 IEEE 7th Annual Ubiquitous Computing, Electronics &
Mobile Communication Conference (UEMCON), New York, NY,
2016, pp. 1-6, doi: 10.1109/UEMCON.2016.7777826.
1 60 90% 92% 90%
2 60 94% 93% 96% [9] NeuroSky, “NeuroSky Store, NeuroSky Mindwave Mobile 2”
3 60 96% 87% 89% Accessed on: September 21, 2020. [Online]. Available:
4 60 88% 90% 94% 90.67% https://store.neurosky.com/pages/mindwave
5 60 91% 89% 90% [10] “Neuro Science for Kids, 10-20 system of electrode Placements”
6 60 92% 84% 87% Retrieved From: https://faculty.washington.edu/chudler/1020.html
Accuracy 91.83% 89.17% 91% [11] NeuroSky, "What is eSense," NeuroSky, September 21, 2020.
[Online]. Available: NeuroSky http://support.neurosky.com/kb/
science/what-is-esense/ [Accessed September 21, 2020]
VI. CONCLUSION [12] NeuroSky, “TGAM1 Spec Sheet” March 24, 2010, Accessed on:
Sept.16, 2020. [Online]. Available:
This paper represents the way to design and implement an https://cdn.sparkfun.com/datasheets/Sensors/Biometric/tgam1.pdf
EEG based Home Appliance control system. The main focus [13] “Understanding NeuroSky EEG Chip in Detail” Accessed on:
in designing this system is to make is very much simple in September 21, 2020. [Online]. Available:
https://www.engineersgarage.com/tutorials/understanding-neurosky-
use and cost-effective in implementation. In doing this, EEG eeg-chip-in-detail-part-2-13/
data has been acquired with the help of a non-invasive brain [14] Liu, N.H., Chiang, C.Y. and Chu, H.C., 2013. Recognizing the degree
wave acquisition headset. EEG data is used to create the of human attention using EEG signals from mobile sensors. Sensors,
control range of different load, whereas Eye Blinks is used 13(8), pp.10273-10286
for triggering. The main focus of this paper is to use the brain [15] Using the BlueSMiRF. Accessed on: September 21, 2020. [Online].
signal to control different types of domestic home appliance. Available: https://learn.sparkfun.com/tutorials/using-the-
bluesmirf/firmware-overview.
This concept reduces the amount of physical movement to
[16] “SparkFun Bluetooth Modem - BlueSMiRF Silver,” WRL-12577 -
control those Home Appliances. This system eventually helps SparkFun Electronics. [Online]. Available:
people who are physically disabled or reluctant to do any https://www.sparkfun.com/products/12577. [Accessed: 05-Dec-2020].
physical movement in controlling the Home Appliance.

View publication stats