This is the research paper of a project carried out by 8 member group including myself.

1. Controlling a Wheelchair According to the Head Movements
Dimutu Upeksha*, Chameera Wijebandara, Chamila Dilshan Wijayarathna, Sameera Nilupul
Wijayarathna,Milinda Fernando, Akila Pemasiri, Suneth Vijaindu Gamage, Madhushi Niluka Bandara
Department of Computer Science and Engineering, University of Moratuwa, Sri Lanka
dimuthu.upeksha2@gmail.com; chameerawijebandara@gmail.com; cdwijayarathna@gmail.com;
sameernilupul@gmail.com; milindasf@gmail.com; akilapemasiri@gmail.com; svijayindu@gmail.com;
madhushib@gmail.com;
+94 -719-084-638
1

2. Table of Contents
Abstract………………………………………………………………………………………………… 03
1.Introduction……………………………………………………………………………………….. 03
2.Materials and Method
2.1 Materials……………………………………………………………………………………….. 03
2.2 Method…………………………………………………………………………………………. 04
2.3 Implementation
2.3.1 Identifying the circuits embedded in the wheelchair……………………………………… 05
2.3.2 Identifying voice commands……………………………………………………………… 05
2.3.3 Controlling the movements of the wheelchair…………………………………………….. 05
2.3.4 Collision Avoider…………………………………………………………………………... 05
3. Results………………………………………………………………………………………………. 08
4. Discussion…………………………………………………………………………………………... 08
5. Conclusion………………………………………………………………………………………….. 09
6. Acknowledgement………………………………………………………………………………….. 09
7. Reference…………………………………………………………………………………………… 09
Appendices …………………………………………………………………………………………… 10
2

3. Abstract - The thirty-year long civil war in our set about designing a mobility control system
country left many young and able-bodied men that is capable of interfacing with head
and women with life-long disabilities. The movements of the officer as that is the only part
soldiers who fought for the defence of our of his body that can be moved. Based on a
motherland and civilians caught in the crossfire lightweight harness worn on the head and using
have suffered grievous injuries that led them to a solid-state gyroscope proving a tilt-sensor
be wheelchair-bound for mobility. However, for capability.
some of the most severely injured war heroes,
even a wheelchair cannot provide mobility. The The mobility control system was then interfaced
motivation for this engineering research project to a battery-powered motorized wheelchair by
is to invent a wheelchair mobility control system removing the joystick-based control unit of the
for an army officer who has lost all motor wheelchair and replacing it with our invention.
control functionality below his neck and is only The prototype was successfully tested for the
able to move his head. complete set of motions that the disabled officer
wishes to make. The next step in this
On a visit to the longterm-care facility in which engineering project is to further research on
this officer resides, his only request was to be providing safety and reliability features for the
able to move his wheelchair unassisted by mobility control unit.
another. Our team of engineering undergraduates
2. Materials and Method
2.1 Materials
1. Introduction - As a result of civil war Material Quantity
which existed for three decades there are LYEB103 Wheel 1
many soldiers who have been affected. So Chair(Electric Wheelchair)
as engineering students we thought that it is Arduino Mega 2560 1
our responsibility to utilize our knowledge development board
and capabilities in a way that is beneficial to Arduino Compatible Voice 1
the country. The motivation behind this recognition Module
ZCT245AN-TTL 2-axis tilt 1
project is to support for such an officer who
sensor
has not control below his neck. HC-SR04 Ultra Sonic 8
range finders
At the moment he is staying at Mihindu Seth
PIC 16f876 micro 1
Madura. What we noticed was that he controller
cannot do anything of his own. What he 10kΏ surface mounts 40 40
requested mainly was a way to move his (for R2R Ladder)
wheel chair of his own. 330Ώ surface mounts 10
7805 Regulator IC 1
So we came with up an idea to carry on a 4MHz crystal oscillator 1
project which results in making a wheel
chair which can be moved according to the
head movements of the officer’s head.
3

4. 2.2 Method
The overall logic behind this project can details refer the state diagram in the
be illustrated as follow. For further appendices
Power on the
wheelchair
Power In
Wheelchair
Check voice Check voice
Obstacle is Check voice
command is command is
detected command is
“Drive” “Stop”
“Manual”
Start Head Stop the movement Stop Head
Recoded user
Control towards the obstacle Control
guidance is played
Power off
Yes 4

5. 2.3 Implementation
2.3.1Identifying the circuits embedded in the  The commands used are “Drive”, “Stop”
wheelchair and “Manual”.
 The wheelchair selected was LYEB103. 2.3.4Controlling the movements of the
 Dissembled the control unit in the wheelchair
wheelchair.  According to the head movements
 We identified that a removal of one  According to the input signal (which
component will even lead the wheelchair to was generated by tilt sensor),
be locked, to the safety of the user. Arduino Uno board generates the
 Then we identified the inputs and outputs of relevant digital signal.
the joystick by the means of an oscilloscope  Those signals are converted to
(Figure (a)) signals which are analogous to the
 We supplied the analog voltages using two output signals of the joy stick and
variable resistors and deceived the circuit. given to the relevant ports of the
 As the prior step was successful we supplied control circuit.
digital values using digital to analog  According to the voice commands
converters (Figure (b))  The original control circuit uses
inputs from switches to control
power and speed.
2.3.2 Identifying the head movements.
 According to the input signal (which
 ZCT245AN-TTL 2-axis tilt sensor was used, was generated by Arduino
which is connected to the head of the user. compatible, voice recognition),
 This can measure the angles with reference Arduino Uno board generates the
to the x axis as well as the y axis of the relevant signal.
horizontal plane(Figure (c)).  The circuit is developed so that the
 Through a serial interface the angle tasks done by switches take place
measured by the tilt sensor is caught by an according to the generated signals.
Arduino Mega board which contains the
driving program of the wheelchair. 2.3.4 Collision Avoider
 8 sonar sensors were used to identify the
barriers which are along and on the sides
2.3.3 Identifying the voice commands of the moving wheelchair.
 By measuring the distance to the
 Arduino compatible, voice recognition
obstacles from the wheelchair, position
shield was mounted on the Arduino Uno
of the obstacle is found.
board.
 By using a clock interrupt the obstacles
 It can be configured to recognize 15 voice
are checked in regular intervals and the
commands.
movement of the wheelchair is
 It generates and sends a signal to the
terminated if there is an obstacle in the
Arduino Uno board according to the
relevant direction
command received.
5

6. 1 2 3 4
1. Vref
2. Vcc
3. X
4. Y
5. X
5 6 7 8 6. Y
7. GND
Figure (a) 8. Vref/ Not Connected
Figure (b)
6

7. Figure (c)
7

8. 3. Results - Some quantitative factors were 100
taken into the accountant to measure the 90
Accuracy Level (%)
accuracy and reliability of the system 80
 The response time of the tilt sensor – The 70
time taken between giving the input and 60
receiving the output (response time) was 50
40
measured
30
20
225 10
200 0
Response Time (ms)
175 1st 2nd 3rd 4th 5th
150
Testing Session
125
100
75
50
25 4. Discussion - The movement detecting
0 module of the developed system is very
1st 2nd 3rd 4th 5th flexible as it takes relative readings. That
Testing Session makes it enable for the user to operate the
system having any initial position. The user
can move not only to a limited number of
In the final result the response time is 10ms
(approximately) and at the time of stopping the directions, but he also can move to any
wheel chair, in first few test sessions a direction that he wishes. Also he can turn,
while moving forward or backward.
significant time was taken to slowdown and to
stop. But after few improvements it was reduced The voice recognition and responsible
and a sudden stop can be handled now. module works accurately in general
circumstances. But in some practical
 The accuracy level of the voice recognition
module – To measure the accuracy level of situations when there is a very large noise
this module may not be able to recognize the
the voice recognition module testing was
carried out by giving voice commands and commands and distinguish between the
getting the responses. commands.
The obstacle avoidance module of the
system works in proper manner. But it needs
a well maintenance because the dust
particles on the sensors may reduce its
detecting ability and the accuracy.
The main benefit of this system is this can
be easy interfaced to any kind of wheelchair
through an appropriate motor controller and
the voice commands can be tailored to the
each user according to their preference.
8

9. 5. Conclusion - The prototype we developed is Also we would like to convey our gratitude to
fully functional and it provides a very good Dr. Upali Khomban, President of Computer
solution to the identified problem. Science and Engineering Society, who gave us
necessary information to proceed in project. He
Even though in the short term we aimed always communicated with us, look at current
only for this particular army officer, he is
progress and motivate us to do the project. We
not the only one with this problem. There
are many army soldiers and officers in the also thank Lt. Col. Athula Samarasinghe who
same condition in Sri Lanka. motivated us to do this project.
Moreover considering about the whole We would also like to convey our special thanks
world it has been estimated that there are to “Brandix Apperal Ltd” who donated an
more than 50,000 individuals who don’t electric wheelchair for this task.
have use of their arms and legs
(quadriplegics). Majority of them are war In addition we like to thank Maj. Madugoda of
heroes of developing countries as Sri Lanka. “Mihindu Seth Madura” and official there, who
helped us to go there and do necessary tastings.
As this invention is a safe, reliable and
They also helped us in finding a suitable
economical for the purpose, these people
would be able to upgrade their living motored wheelchair for our project.
standard at least for some extent by using
At last but not in least we would like to thank all
this product.
who helped and motivated us.
So our idea is, that by expanding this
7. References –
product we would be able to support to the
 http://arduino.cc/en/Main/ArduinoBoard
social and economical development as a
Mega/
whole, while supporting to those suffering
lives.  Data sheet of Arduino Compatible
Voice recognition Module
6. Acknowledgement - First on foremost, we  Data sheet of ZCT245AN-TTL 2-axis
would like to thank, Head of the Department tilt sensor
of Computer Science and Engineering,  Data sheet of HC-SR04 Ultra Sonic
University of Moratuwa, Eng. Lt. Col. Dr. range finders
Chandana Gamage who guided us through  Data sheet of PIC 16f876 micro
doing this project, provided important controller
advices, helped us in difficult periods and  Data sheet of 10kΏ surface mounts 40
provided equipment for our project. His (for R2R Ladder)
willingness to motivate us contributed  Data sheet of 330Ώ surface mounts
tremendously to our project. Besides we  Data sheet of 7805 Regulator IC
would like to thank CSE staff who helped us
by giving advices and providing equipment
which we needed.
9

10. Appendices
The system architecture is depicted below.
Main sensor which
tracks the head Main Circuit
Module
movement Board of Wheel
converting Chair
Tilt Sensor digital signal
in to analog
US1
R2R
Ladder
US2
Arduino Board
(PIC) Voice Recognition
Model
US3
To activate the
US4 system by voice
MP3 module command
for voice
To avoid responses
Show the current
obstacles and
state
for safety
Of the wheelchair
10

11. The state chart of the system is depicted below
11