24-07-2012, 11:26 AM
Wheelchair for Physically Disabled People with Voice, Ultrasonic
and Infrared Sensor Control
Wheelchair for Physically Disabled People with Voice, Ultrasonic.pdf (Size: 2.26 MB / Downloads: 87)
Abstract.
This paper describes a wheelchair for physically disabled people developed within the UMIDAM I
Project. A dependent-user recognition voice system and ultrasonic and infrared sensor systems has been integrated
in this wheelchair. In this way we have obtained a wheelchair which can be driven with using voice commands
and with the possibility of avoiding obstacles and downstairs or hole detection. The wheelchair has also been
developed to allow autonomous driving (for example, following walls). The project, in which two prototypes have
been produced, has been carried out totally in the Electronics Department of the University of Alcalfi (Spain). It has
been financed by the ONCE 2. Electronic system configuration, a sensor system, a mechanical model, control (low
level control, control by voice commands), voice recognition and autonomous control are considered. The results
of the experiments carried out on the two prototypes are also given.
Keywords: wheelchair, mobile robot, physically disabled, control, autonomous guidance, joystick, PID controller,
ultrasonic and infrared sensors, micro-controller, fuzzy, speech recognition
1 Introduction
The number of people who need to move around with
the help of some artificial means, whether through
an illness or an accident, is continually increasing.
These means have to be increasingly sophisticated, taking
advantage of technological evolution, in order to
increase the quality of life for these people and facilitate
their integration into the working world. In
this way a contribution may be made to facilitating
movement and to making this increasingly simple and
vigorous, so that it becomes similar to that of people
who do not suffer deficiencies. Systems already
exist which respond to many of the needs of people
with different degrees of incapacity (Leifer, 1981;
Borenstein & Koren, 1985; Madarasz, 1986; Jin et al.,
1993). However, there are still important advances
to be made in this field. This justifies the numerous
research programmes which are being carried out
at the present time; the TIM MAN project (Miller
& Grant, 1994; Grant, 1994), the COACH project
(Gelin et al., 1993) and the SKIL guide system (Sabbe,
1993). The main reasons for their justification are as
follows:
a) The present high level of technology in the
electronic and robotic systems permits some of the
mobility problems suffered by certain people to
be resolved. Electronics solves the problems very
acceptably for the users. This is because the electronics
used is eminently suitable for coping with
the needs presented.
b) Unfortunately more and more people are appearing
with incapacities which prevent them from carrying
out normal activities. Most have serious problems
related to movement.
The type of artificial aid needed by a disabled person in
order to move about depends, to a large extent, on the
level of his incapacity. For example, in order to guide
a wheelchair, various situations can be distinguished:
a) If the user is capable of controlling his head or his
hands, the ideal solution is the use of a joystick.
b) Where there is a high level of incapacity, solutions
are basically centred on the use of other means, such
as the voice or eye movements. In this case, the
presence of safety sensors is justified with the object
of assisting the user to guide the chair (detection of
obstacles, nearness to certain places, the existence
of stairs, etc.).
c) Only in extreme cases it is suggested that there may
be a need for the chair to cover certain distances in
an autonomous manner, without the need for any intervention
on the part of the user (interest in this type
of wheelchair could be in the following of prefixed
204 Mazo et al.
routes in hospitals, recreation centres, etc.). In this
case the presence of external sensors is vital.
Another important requirement that a wheelchair has
to fulfil is that of responding rapidly and efficiently to
the commands of the user, independently of the method
used for giving these commands.
1.1 The UMIDAM Project
Although a great deal of research work has been carried
out on mobile robot applications in industry (Sitherama
& Elfos, 1991; Meng & Kak, 1993; Moravec, 1985;
Watanabe & Yuta, 1990; Maravall & Mazo, 1990;
Mazo et al., 1992; Goto & Stentz, 1987), in agriculture
(Eden et al., 1993), in security (Crowley, 1987),
in toys (Bradley, 1980; McAlister, 1980), not so much
work has been done towards aiding physically handicapped
people (Leifer, 1981; Borenstein & Karen,
1985; Madarasz, 1986; Jin et al., 1993), basically
considering people suffering from great incapacity
(tetraplegics), that is, people who do not have the possibility
of using a joystick.
The UMIDAM project, described in this article, is
orientated towards providing solutions to the need for
moving around of disabled people who have great incapacity
with regard to driving. These needs and their
possible solution were originally put forward by researchers
in close collaboration with personnel from
the ONCE Foundation. A summary of the initial objectives
is given below:
a) To design an electronic system which could be installed
in an of the commercially produced electric
wheelchairs.
b) To make guiding the wheelchair possible by means
of oral commands, as well as by using the classic
joystick. The electronic interface would have to
permit the chair to be further guided using other
means (eye movement, etc.).
c) To incorporate a sensor safety system in the
wheelchair which would permit obstacles and the
presence of stairs or holes in the ground or floor to
be detected°
d) The electronic system had to be open and modular,
in the sense that future additions could be made
which would not necessarily be related to guiding
the chair (the control of household electrical equipment,
opening doors, etc.).
e) Another aspect of great importance with regard to
the design was that the final prototype should be
economically priced for its later manufacture and
commercialization.
f) The environment originally considered for the use
of the wheelchair was homogeneous floors and
ground where different types of obstacles (static and
mobile) and the presence of pronounced differences
in levels (such as stairs) might be encountered.
The final prototypes obtained, after three years of
research and development, complied with all the characteristics
initially required. The first phase consisted
of the production of an electronic system for guiding
the chair by means of a joystick and voice control, including
the power interfaces for controlling the motors,
speech recognition, etc. (first prototype). Later, in the
second phase, a sensorial system based on ultrasonic
and infrared sensors was also included combined with
the strategy for assisting the guiding of the wheelchair
(second prototype). In the following we refer to the second
prototype, since all the options in the first prototype
are included.
All the electronic system and the philosophy for
functioning has been sufficiently refined to achieve the
following performances:
a) To guarantee easy, comfortable driving.
b) To respond to the speed requirements for a system
of this type (maximum speeds of up to 3 m/s).
c) To be easily adaptable to any type of commercial
wheelchair chassis.
d) To facilitate learning to handle the chair and obtaining
maximum efficiency.
e) To guarantee practically constant speeds, to a
large extent independently of the characteristics of
the surface over which the wheelchair is moving
(greater or lesser roughness of the floor or ground
and the slope of same) and the weight of the person
using it.
f) To make the system easily configurable, on the basis
of the needs of the user: activating or de-activating
of the various sensors, selection of different voice
patterns, selection of different speed margins,
human-machine interface which permits up-to-date
information on the state of tile wheelchair, etc.
g) To make it possible for the same wheelchair to be
used by various people without the need for recording
the voice patterns each time the wheelchair is to
be used. This has been achieved thanks to a memory
board which can be personal for each user.
h) To make the electronic system open to future additions.
Fig. 1. Frontal view of the wheelchair.
i) To make decisions over stopping and reducing speed
when obstacles of the presence of stairs is detected,
as a function of the degree of danger supposed for
the user in each case.
As previously mentioned, the UMIDAM is characterized
by a speech recognition system, which constitutes
the fundamental base for driving, and by a sensor
system, made up of ultrasonic and infrared sensors for
detecting obstacles, stairs or any pronounced irregularities
in the floor or ground and by being able to be
guided autonomously following walls. Figures 1 and 2
show the distribution of the different UMIDAM modules
taking these sub-systems into account.
Among the work carried out on similar problems,
the following projects can be found: Tin Man (Miller
& Grant, 1994; Grant, 1994), COACH (Gelin et al.,
1993) and the SKIL guiding system (Sabbe, 1993),
Two prototypes were developed for the Tin Man
project (Tin Man I and Tin Man II) which are characterized
by being equipped with a complete sensor
system (encoders, contact, IR proximity and
fluxgate compass) which permits obstacles to be
Wheelchair for Physically Disabled People 205
detected and avoided, going to pre-designated places
and manoeuvring through doorways and narrow or
crowded areas. Present efforts in this project are being
centred on the preparation of a user interface at
task level.
The development of a semi-autonomous wheelchair
has been approached in the COACH project. The result
is a wheelchair guided by a joystick with the help
of a sensor system formed by ultrasonic and infrared
sensors. This sensor system is particularly considered
for avoiding obstacles and following walls.
The SKIL guiding system is made up of ultrasonic
and infrared sensors which permit obstacles to be detected,
the opening of doors, etc. The originality of this
system lies in the infrared sensors employed: the emitters
are located on one side of the wheelchair and the
receivers on the other, measurement of distances being
made by triangulation.
Various common aspects exist in all the projects,
including the UMIDAM:
a) They are all based on commercial wheelchairs in an
attempt to make the systems developed as universal
as possible.
b) Adaptations are made to the sensor systems normally
employed in mobile robots used for more
general purposes.
c) Various operating modes are available in an attempt
to cover different demands and autonomies.
d) Whatever the case, emphasis is placed on the importance
of achieving the autonomy of a disabled
person before the autonomy of the wheelchair
itself.
2 UMIDAM Operating Modes
The operating modes have been defined taking into
account the fact that a great deal of the design considerations
in industrial mobile robots could be applied to
robots for helping the disabled. But there are several
differences between them:
1) Many mobile robots for industrial applications follow
pre-designated routes which may be marked
by ropes, paint, etc. to facilitate guiding. But the
routes to be followed by robots for helping the
disabled are dictated by the user according to
his possibilities and for this reason have to be
generic.
2) Domestic environments are very largely disordered,
so the sensor system has to detect all the obstacles,
206 Mazo et al.
Fig. 2. Side view of the wheelchair.
stairs, etc. For an industrial robot the environment
may be ordered most of the time the position of the
fixed obstacles could be in the robot memory.
3) In a robot for the disabled, a person will always
act directly over the system, so the control must be
designed to follow the user, in the easiest way.
It is precisely this third point which caused the
UMIDAM to be conceived with various operating
modes. In this way it adapts to the various degrees
of incapacity of the Users. Whatever the case, it must
be remembered that the initial intention was to allow
people who could not use a joystick to be able to guide
the wheelchair.
There are three driving modes for the UMIDAM:
Joystick control: Control is carried out using the joystick
in this driving mode. This mode is called "Manual
control".
With voice commands: In this driving mode, the
wheelchair is controlled by means of various voice
commands. Here the sensors act as a safety means
against any possible errors which the user may commit
when giving the commands or in the unexpected
presence of obstacles. The user has the possibility
of activating or de-activating the sensors, either the
ultrasonic or the infrared or both. This possibility of
configuring the sensor system allows the user to drive
while subjected or not to the action of the sensors, in
terms of the environment in which he is moving or
even in his handling of the chair. This driving mode
is called "Voice control".
Autonomous driving: This method of driving is considered
so that the wheelchair may follow walls
or forward movement around obstacles. This permits
the user to move around in certain environments,
such as hospitals, rehabilitation centres, etc.,
without the need for having to give voice commands
constantly to the wheelchair. This mode
is called "Autonomous control". In this driving
mode, linear speed is modified by means of voice
commands.
Figure 3 shows the selection process for each of the
driving modes. It can be noticed that selection is necessary
among the driving modes before connection to
the "Manual control" and "Voice/Autonomous control"
systems. This choice is made by means of a manual
switch (there is no sense in the user himself be able
to do so if he cannot use a joystick.) If the chosen
mode is "Voice/Autonomous" the system comes into
the "Voice control" mode. In this mode, every time the
work "track" is pronounced, the mode automatically
switches between "Voice control" and "Autonomous
control".
Wheelchair for Physically Disabled People 207
@
TRAC~CK
Fig. 3. Mosdeel ection.
3 Electronic System Configuration
The block diagram of the electronic system is shown
in Fig. 4. It has been conceived as an open and adaptable
modular system. In this way, an eventual addition
could be made just by adding a board with the desired
function.
The main blocks in the system are as follows:
1) Feedback control of the angular speed of each motive
wheel.
2) Power and Activation Motor Unit.
3) Speech Recognition Unit.
4) Ultrasonic and Infrared Systems.
5) Memory Board.
6) User Interface (Display and Keyboard).
The system has a flexible configuration and is easily
upgradable. More features can be added by merely
changing/adding specific boards. The different blocks
of the present system use 8-bit micro-controllers of the
MCS51 family as a CPU. All of the CPU's have a standard
Serial Interface Unit (SIU) that allows an appropriate
communications way via the different modules.
The basic system consists of the motors, a joystick,
the power box and the motor control board. The voice
control is provided by a board that provides recognition
features. The display has its own processor making the
connections clear and easy: it is connected to the entire
system using only three wires. The ultrasonic and the
infrared systems have another processor, which provides
obstacle, hole and stair detection. The software
needs few changes from one configuration to another.
The system has a parallel and serial bus, both in the
same physical space. The serial bus interconnects all
the cards. Communication among the different CPU's
uses the serial bus, although boards can extend the parallel
bus to its close boards.
The system software has a core that is the same for all
the boards. This core provides connection, networking
and intertask facilities for the user card software.
The multi-processor configuration provides the same
working level for all the boards. Connections between
one and another are at the same priority level. The driving
mode gives more priority to one board depending
on the mode and not on the board itself.