Research paper

Project Name:
Voice controlled wheelchair for physically disabled people

Author Ubaid Khan

Qualification Masters

Degree Electronics engineering

Abstract:
Speech signals are the most important means of communication
in human beings. Almost every conversation to interact is done by
means of voice signals. Sounds and various speech signals can be
converted into electrical form using a microphone. Physical
disability can occur due to multiple reasons like injuries from
accident, age related & health problems. Wheelchair is used to
provide a mode of transportation for such disabled people with
impairments in hands and legs. People with such issues like
paralytic people find it difficult to operate the wheelchair
manually or using a remote assembly. For such people the project
is designed to work on voice based commands so that the
paralytic or disabled person can give direction commands by just
speaking into the microphone given. The system also includes
directional buttons for wheelchair control using remote. The
system consist of an Atmega328 based circuit interfaced with an
voice recognition module that takes speech commands from the
user converts this speech into digital data which is then debugged
by the micro-controller to get directional commands. The entire
system consist of 2 circuits i.e. the transmitter circuit and a
receiver circuit. Transmitter circuit comprises of the voice
recognition module and the receiver circuit consist of the motor
and driver assembly. We use a NRF trans-receiver module for the
communication. A 16*2 LCD is used to display the command
which is given to the wheelchair.
Introduction:
The wheelchair is the most ubiquitous equipment used by people
with lower limb disability. It enables them some degree of
freedom in mobility and independence as opposed to those with
both upper and lower limb disabilities. Most of the wheelchairs
available in the market are manual in nature with some available
with motorized option. Anything beyond that is custom made
which is costly and not within the reach of most people. People
with severe lower and upper disabilities have to resort to costly
electronic controlled wheelchairs or be totally dependent on
another person to move them around in their manual
wheelchairs. Motorized wheelchairs controlled through joystick,
softball, finger, tablet, chin and head are readily available at a high
cost but most of t hem do not cater for those with upper limb
disability. The advances in speech recognition technology have
made it possible to control any electronics based device using
voice command. This technology is capitalized for voice
controlled wheelchair to assist those with both upper and lower
limb disabilities. A variety of voice controlled wheelchairs have
also been developed by other researchers [1 5]. The World Health
Organization (WHO) has estimated that of the 75 million people
who need assistive technologies wheelchair, only 5% to 15% of
those have access to one [6]. Figure such 1 shows the block
diagram of the voice controlled wheelchair. There are three main
parts in the wheelchair: (1) wireless communication that
incorporates voice recognition module, (2) microcontroller and
(3) motor controller. The voice recognition module converts the
voice from analogue to digital with built in digital signal
processing system that recognizes the voice commands. This
model is pre-programmed for each command and calibrated to
the user’s voice. The microcontroller processes the voice
recognition module output to control the left and right motors
which in turn controls the movement of the wheelchair.
Methodology:
The voice recognition module is the key feature of this project
that is used to setup the desired voice command and output. It
consists of three phases, which is voice customization, voice
capture and voice recognition. Voice customization is the process
of matching the desired voice recorded to the desired output
signal. Voice capture is the phase that records the desired
person’s voice command and saves the voice based on the
customization configuration. The voice recognition phase is the
fin al phase where when voice command has been recognized,
this module will send a specific signal to the microcontroller for
the necessary operation. Figure 2 shows the block diagram of the
voice recognition module. The voice instructions are recorded via
serial communication with the PC using Access Point
communication software with baud rate of 9600. After the
connection of the voice recognition module to the PC is
successfully implemented, the existin g voice instructions are
deleted by sending hex command AA 01. The recording is started
with the desired voice command in group 1 by sending the hex
command AA 11. After this command has been sent, user is
required to record a total of five voice commands in order to
complete the group recording. On completion of the voice
capturing phase, verification is required by sending the hex
command AA 21 to import group 1 into the voice recognition
module. The recorded voice command is verified again by
repeating t he five commands that has been recorded earlier.
Figure 3 shows the returning result of the voice verification after
each command has been verified successfully by the module. The
five voice commands used are: Forward, Backward, Turn Left,
Turn Right and S top . The HC05 Bluetooth module is used as the
wireless transmission medium between the microcontroller and
the voice recognition module in controlling the wheelchair
movement. The Cytron MD13S motor drive is used to control the
speed and direction of the DC brushed motor. The pulse width
modulation (PWM) pin is to control the speed and the DIR pin is
used to control the direction of the motor. The PWM pin is
connected to 5 V and the DIR is fed to the PWM signal of the
microcontroller. When the PWM signal is at 50% duty cycle, the
motor wil l stop running. If the PWM has less than 50% or more
than 50% duty cycle, the motor will turn clockwise or counter
clockwise based on the polarity connection between the motor
and MA/MB pins of the motor drive. Infrared (IR) sensors with
measuring ranges of 10 cm to 80 cm were fixed to both sides of
the wheelchair for obstacle detection. Two Hall Effect sensors
were also fixed on both sides of the wheelchair to detect the
wheel rotation speed since the DC brushed motor does not have a
speed control. These sensors will detect the rotation of the wheel
and feedback to the microcontroller to achieve speed
stabilization. Twelve magnets were fixed to each side of the wheel
before implementing the sensors as shown in figure 5. The
diameter of the wheel is 40.64 c magnets were fixed 10.64 cm
apart along the circumference of the wheel. In this set m and the
12 up, when the wheelchair moves approximately 10.64 cm from
the initial point, the Hall Effect sensor will trigger a signal to the
microcontroller as a feedback signal. In this step, the
microcontroller will start the timer to count another incoming
signal. This method will enable the microcontroller to distinguish
how fast or slow the wheel is turning thus enabling the speed
variation communication from Hall Effe ct speed sensor to
microcontroller, and from microcontroller to control the speed by
using PWM in the motor driver. Figure 5 . Positioning of magnets
on the wheels. An analog joystick module was also incorporated
as an additional control option for thos e with lower limb
disability only. Atmega 328p microcontroller is used to
communicate with the Bluetooth module, joystick control, Hall
Effect sensor, IR detectors and the motor controller. Arduino IDE
code compiler is used to complete the whole system int
erconnection and communication. An emergency button was also
incorporated using the joystick press key. This is to prevent any
voice command malfunction caused by noisy environment. Figure
6 . MY1016 scooter motor used for this Chained gear ratio setup.
prototype provides a rated torque of 250 W (0.9~0.97 N/m). It is
a 24 V high RPM motor with no load up to 3300 RPM. A 5to1
chained gear ratio system was applied between the motor and the
wheelchair to deliver five times the rated torque as shown in figu
re 6. Power supply to the DC brushed motor and the two motor
drives is provided by two parallel 12 V rechargeable lead acid
batteries that provide 14 A current. All the above mentioned
hardware was implemented onto a conventional foldable
wheelchair which is readily available in the market. The I2C OLED
display is implemented to provide a more user friendly interface.
The user is also able to observe the executed instructions on the
display. Tests and Results:
The main control part is the function of the prototypestarting
period. At first, the prototype will wait for the joystick key to
select user preferences. It has total of three user preferences,
which can be selected by using joystick. The first user preference,
second user preference, and third user preference has been set as
turn left, forward, and turn right respectively. After the user
preference selection, the program point will verify both control
input, which are joystick and voice. If joystick is detected, it will
provide manual control operation of the wheelchair, else if voice
is detected, the pointer will go to the voice command control for
further operation.

Components:
 Atmega Microcontroller
 Robotic Chassis
 NRF module
 DC Motors
 Resistors
 Capacitors
 Transistors
 Cables and Connectors
 Diodes
  PCB and Breadboards
 LED
 Transformer/Adapter
 Push Buttons
 Switch
 IC Sockets

Block Diagram:
Software Specifications
Programming Language: C
Advantages:
 It works on simple voice command.
 The size of this robot is small.
 It is user friendly.
 It reduces man power.
 Low power consumption.

Disadvantages:
 It only performs command like Backward, Left, Right, Stop.

Applications:
 We can use this system for those people who is physically
handicapped
 We can use this system for agriculture purpose.