Adaptive Technology:
Making the Information Society
Accessible to the Disabled
Wayne Patterson
Dean in Residence, Council of
Graduate Schools
The
human race is in the midst of a profound transition in the way in which we
produce material to satisfy our needs. The nineteenth century saw another
profound transition through the industrial revolution, but today we are in the
midst of the “knowledge revolution.”
The trends
are clear and irreversible: fewer and fewer people are needed to ensure the
production, distribution, and marketing of the goods and services that all of
us require; but more and more people are need to produce, distribute and market
the information that is becoming our
most important commodity.
One
consequence of the change taking place in society is that it may provide new
hope and opportunities for those among us who, through physical disability, are
unable or less able to carry out some of the physical tasks that we currently
value in our economy.
In
earlier generations, a physical disability disqualified individuals from being
able to participate fully in society. However, as we will see, there are
technological advances, both in hardware and in software, that allow persons
with even very severe disabilities to work successfully in the information
industry.
In
particular, we will look at the field that is now called either adaptive technology or assistive technology. This area of
technological development provides means for individuals who are paraplegic,
visually impaired or completely sightless, hearing impaired, or with cerebral
palsy, muscular dystrophy or similar disorders to have access to the full range
of services and opportunities that are available in this increasingly
cybernetic society.
The
specific technological approaches we will examine will include:
a. Keyboard
technologies (for persons with limited or no hand and finger movement)
b. Pointing
devices replacing the “mouse” (for persons ranging from limited movement to
being fully paraplegic)
c. Video
display technologies (for the visually impaired)
d. Speech
synthesis (for the completely sightless)
e. Speech
recognition (again for the fully paraplegic)
f.
Braille devices
g. Speech
to text translation (for the hearing impaired)
In
addition to the discussion of specific technologies (with some demonstration),
issues of cost, availability, use in the African context, and legal issues
surrounding persons with physical disabilities will be discussed.
The paper concludes with a discussion of possible future technological advances in this field, looking at both opportunities and problems in such development.
We are
all so familiar with keyboard design and that the strange looking word QWERTY
has entered into our vocabulary. Have you ever wondered what led to the choice
of positioning the letters Q, W, E, R, T, Y etc. across one row of a standard
keyboard?
Strangely
enough, the design was chosen in response to a sense that human’s physical
capability could exceed the machine’s capacity. In the days of manual
typewriters, when QWERTY was first introduced, designers felt that keys that
would often be struck consecutively should be placed as far apart on the
keyboard as possible. The reason was, that the keystroke caused a metal key to
swing forward and making imprint on paper; if two consecutive keys were in
close proximity physically, there was a greater chance that the keys could
become entangled and have to be physically pride apart. And of course, the faster
the typist, the greater the chance that this entanglement could occur.
This
keyboard design been persisted through the development of electric typewriters
- where essentially spacing didn’t make any difference anymore - to the era of
keyboards for computing devices.
Indeed,
it has often been pointed out that the QWERTY design in fact is far less
efficient than other arrangements of keys, for example in the Dvorak design.
Of
course, these designs were made without consideration of the challenges that might
be faced by persons with limited dexterity, such as might be the case in a
person with a muscular disorder, or arthritis, or indeed with missing hands or
fingers.
Current
designers have given thought to such problems, and new software and hardware approaches
have been developed. For example, even with standard keyboards, techniques for
modifying the keyboard response are available:
Sticky
Keys. If a typist uses only one finger, the combination
strokes necessary, for example in using the Ctrl key, are not possible. A
sticky key solution allows the user to set a switch so that two consecutive
keystrokes takes the place of simultaneously depressing two keys.
Key
Repeat Rate. This defines the length of time that he needs to
be held down before it repeats the keystroke.
Acceptance
Delay. This is the amount of time that the key needs to be held
down before it is processed by the computer.
Mouse
Keys. Setting this option allows a pointing device such as a
mouse to be moved around using other keys on the keyboard, for example the
numeric keypad keys.
All of
these features are available on PC operating systems, including DOS, Macintosh,
and Windows from 3.1 to 98 and NT. And all of these features are free.
Keyboards
also come in different sizes and shades for persons with physical disabilities.
For example, small keyboards are available that may be more useful for
one-handed users, and such keyboards may more easily fit in between the arms of
the wheelchair.
On the other hand, some users may adapt more readily to a large keyboard. The user with muscular control problems may adapt more readily to keyboard where there is a larger size key and a larger target area. Many such keyboards have a built-in are as the letters are slightly recessed below the surface of the keyboard.
If you
have watched a court reporter, you may be familiar with the concept of the cord
keyboard. Such keyboards have only a few keys, but are designed for users to
depress several keys simultaneously.
Another
approach for a one-handed user is the one handed keyboard. This keyboard has
the full range of keys, but they are so designed to be reachable by the fingers
of one hand. An example of a Maltron left-handed keyboard, generously donated
for this presentation by Applied Learning of Wayne, Pennsylvania, will be
demonstrated here.
With
the advent of graphical computer interfaces such as the Macintosh OS and
Windows, the use of the pointing device has become central to computing. Of
course, the pointing device with which we are most familiar is the ubiquitous
mouse.
The
pointing device, or mouse, that is standard with most computer systems today is
a difficult tool to operate for someone with a physical disability affecting
the hand or fingers. The requirement to move it over a large area while holding
it steady for accuracy and then pressing the buttons, demands extensive
accuracy and fine motor control. Many people have considerable problems in its
application when they are first introduced to the device. However, it is the
most popular and most extensively used type of pointer.
Portable
or laptop computers and notebook computers have popularized other variations of
the mouse, usually for reasons of space limitation. The particular laptop I am
presently using has a touchpad. A touchpad is a device which is fixed, and
which operates by sensing the position of one’s finger on the pad to determine
position or direction. Other laptops may use trackballs, a fingerpoint, a
J-mouse (using certain key combinations as cursor directives), or other
variations using aspects of the keyboard. Computer games have also popularized
the joystick for the same purpose.
The
trackball is a device that in some ways turns the mouse upside down. The
rolling surface of the pointing device is face up to the user, and is
manipulated by the user; as opposed to the mouse, where the device is
manipulated, causing the spherical surface on the bottom to register position
and direction.
Like
the touchpad, the fingerpoint has been used with many portable computers. It is
the size of a pencil eraser and protrudes from the center of a keyboard. The
mouse buttons are placed under the spacebar for easy access. This device
requires a high level of control to get fine accuracy and you need to use quite
a firm touch. It eliminates the requirement for an external mouse making it an
effective alternative. You would traditionally expect the user of a fingerpoint
to combine this with keyboard short cuts. Having the mouse in the center of the
keyboard saves a lot of space on the desk for other essentials. For this reason
you will find fingerpoints mainly on space saving or small keyboards.
All of
these devices are similar in that they require the user to use a finger or a
hand to indicate direction, which is translated into a directional signal for
the cursor on a screen display.
For
persons without the capability of pointing using a hand or finger control
problem the number of other techniques have been developed.
The
pointer on the screen can be controlled via the movement of the head. There are
two alternatives available at the moment: (a) The Headmaster which is operated
by a headset in combination with a suck-puff switch that is attached to its
frame. (b) The Head mouse which is operated by a small circle of reflective
material which can be placed anywhere on the body and tracked by a sensor box.
The mouse buttons can be operated via 2 infrared switch boxes, allowing the
switch or switches to be located in any comfortable position. Coupled with an
on-screen keyboard head pointers can be an effective method of hands-free
computing.
A
suck-puff switch is operated by a suck-puff tube, which is a device consisting
of one or more switches operated by a puff of breath, usable by persons only
able to operate switches by breath control.
There
are many approaches available to assist persons with visual impairments. Many
people who are visually impaired still have some capacity for sight. And in
many cases, these persons can be assisted by a modification of the computer’s
video display.
Perhaps
the simplest of these modifications, readily available in all word processing
software, is the zoom function, or alternatively, changing the size of the font
in use. For example:
This sentence is displayed in 10 point
size.
This sentence
is displayed in 18 point size.
This
sentence is displayed in 22 point size.
Often,
improved contrast will also aid the visually impaired user. Windows provides
this option as part of its control panel: by pressing Start, then Settings,
then Control Panel, one will see displayed an icon called “Accessibility.” This
is a control for various features designed to assist the disabled.
Once
you have double-clicked the Accessibility icon, choose the Display tab in order
to see the setting for High Contrast in the monitor. This setting allows for
the menu items in Windows to be displayed with high contrast, with white on
black or black on white, or with customized approaches.
For
persons who are completely without sight, having the ability to produce speech
from text is a valuable asset. This is called speech synthesis, or
text-to-speech production.
Speech
synthesis is a difficult field for researchers. It is by and large not a
difficult problem (easier in some languages than others) to analyze a portion
of text into its spoken components. For example, the phrase:
is
easily divided into
It is also
not difficult to cause a computer to produce sounds corresponding to the
individual components of speech. It is, however, quite a difficult problem to
combine all of the audio components together in a way that closely mimics a
human’s inflection, breath control, intensity, and other characteristics not
clearly indicated by text. Thus speech-to-text translation, while quite
comprehensible to most human beings, is seldom of the quality and pleasing
nature of a well-spoken human being.
The
Bell Laboratories are an international leader in this field of research, and
their Web site provides a good introduction to the techniques they use for
speech production:
http://www.bell-labs.com/project/tts/
At this
Web site, you may input text and have spoken back to you-in many different
voices and languages. Consider this example:
I
am pleased to be speaking to you here at the
Atlas/Afgrad
Conference on Disabilities, here in
Dar
es Salaam, Tanzania.
Speech
recognition is the problem of recording human speech, and translating the
sounds into the words and text spoken. This has also been an area of
considerable research, and one that had only very modest results until very
recently. But the software improvements in just the past two or three years have
brought this technology to the fore.
How
well do these systems work? There are many people who use voice recognition
systems successfully and productively. Having said that it is important to
realize that they are not magic wands. To get the most out of voice recognition
it is important to:
(1)
choose the right system
(2)
choose an appropriate supplier
(3)
have realistic expectations
(4)
take the time and make the effort to learn new skills and a new system
(5)
get adequate training and support
We all
speak differently. What voice recognition systems do is to store “voice models”
for each user. These systems have a vocabulary of words that occur frequently
in speech; they match your spoken words to words in the vocabulary. At first,
this is not completely accurate and you will need to correct mistakes. However,
as you make corrections the systems learn and become more accurate.
To
initiate this process of adaptation some systems require you to do an initial
word match, where you have to read to the computer before starting to use the
software.
Each of
the voice recognition systems has a built-in vocabulary that is referenced
during speech input. Although words are sometimes not recognized the inserted
word is always correctly spelled. You can add words to the standard vocabulary
simply by saying the word and then spelling it.
The
systems vary in the number of words they keep in memory. A typical vocabulary
is 30,000 words. Those working in highly technical occupations may choose to
have a larger vocabulary or to purchase a specialized vocabulary (e.g. medical,
legal and so on).
There
are a number of continuous speech recognition products that allow you to
dictate using natural speech - you can pause as and when you like. However,
these systems work best when speech is at a predictable rate and some
additional effort is made to speak (but not stress) each syllable spoken.
Continuous
speech systems look at groups of words and use statistical models of your
speech to improve accuracy. These systems are capable of higher levels of
accuracy than the older discrete speech products. Experience shows that people
who have speech impairments are likely to have more success using discrete
speech products.
All of
the voice recognition systems mentioned will allow you to dictate words into a
word processor and some have additional features which allow you to control
other software applications by voice. For example you could say “computer, go
to Excel” to start your spreadsheet and “file, send mail” to send an electronic
mail. This is known as “command and control” and is particularly useful to
people who have difficulty in using a mouse and keyboard. For a system to be
deemed “hands-free” it would expected to be able to start programs, modify
settings, make corrections and have a method of moving the mouse pointer all by
voice.
If you
are happy with some keyboard use and only want to make use of voice recognition
for writing text then you may find that one of the dictation only systems will
meet your needs.
The
majority of voice recognition systems have facilities for you to add your own
macros. Macros allow you to insert standard pieces of text quickly and can
allow you to link together a number of operations. For example “sign-me-off”
could insert “yours sincerely,” 6 blank lines, your name and title. Macros are
a powerful way of speeding up your dictation and can make your system easier to
use.
Recognition
accuracy can be improved by making sure that you use a good microphone and
soundcard. Most of the systems detailed here are supplied with a headset
microphone. These keep the microphone a fixed distance from your mouth while
allowing you to move around. There are other choices in microphones, for
example: clip-on microphones, radio microphones as well as microphones fixed on
a stand. Some of these would be more suitable if you have difficulty taking a headset
microphone on and off.
The
soundcard in your computer passes the microphone signal on to the speech
recognition software. Soundcards vary in quality and it is worth checking that
your soundcard is compatible and of reasonable quality.
At the
dawn of the computer era, speech recognition was one of the highly touted
objectives of computer capability. For many years, because of both hardware and
software limitations, as well as the complexity of natural languages, effective
speech recognition has remained an elusive goal.
For
persons who are completely incapable of hand, arm, or finger movement, being
able to interact with a computer system via speech would clearly be highly
desirable.
Recent
software products have finally brought speech recognition capability to a
usable and cost-effective level. Indeed, almost this entire article was
dictated directly into Microsoft Word using Dragon speech recognition
software.
Among
the very difficult problems in speech recognition are the adaptation of
software to different voices, the difficulty in determining word brakes, and
the change of inflection by a single speaker at different times of the day or
under different physical conditions.
The use
of this particular package (Dragon Point
and Speak) involves an initial training session lasting about 45 minutes
that is meant to build up speech files based on the users diction and
inflection. The training session I used involves reading the selection of
standardized text, in one case the chapter from Arthur C. Clarke’s novel 3001.
My
experience to date has shown that this software performs above the 90 percent
level, to the point where I’m currently using this package for other document
preparation. And, incidentally, this package sells for about $50.
Braille
devices are available for those who are completely without sight. Braille
devices of course predate the computer era, but have been adapted to these
environments.
Unfortunately, among all of the devices and software approaches examined here, will devices tend to be the most expensive. Braille printers, which will receive as input standard text and display in Braille, tend to cost upwards of $1000.
There
are also software systems designed for sighted persons to prepare Braille. One
example is the Duxbury Braille Translator. Also, many standard installations of
Windows have a Braille font as part of the system. Here is an example of the
latter.
There are also software systems designed for
sighted persons to prepare Braille. One example is the Duxbury Braille
Translator. Also, many standard installations of Windows have a Braille font as
part of the system. Here is an example of the latter.
I have
chosen to differentiate this category from the category for speech recognition,
because the set of techniques involved here are somewhat different. First of
all, as was indicated above, the development of speech recognition software is
highly dependent on the vocal characteristics of any individual. It would be highly
unlikely that any of the current speech recognition systems would be capable of
translating speech emanating from someone for whom the system was not trained.
However,
in this category of speech-to-text, I refer primarily to techniques capable of
rapidly translating any spoken words. These techniques are used widely now in
television broadcasting, and are usually referred to as “closed captioning” in
this environment.
It is
possible to download at no cost a closed captioning software system developed
at the United States Department of Education, and the appropriate Web site is
quoted below. However, the system is one that requires some considerable
technical expertise to load and run.
In
order to provide interested parties with a list of resource areas to seek out,
here are the Web site addresses for numerous organizations involved in the
computer technology development for adaptive and assistive technology.
In
addition to each Web site, I have given a brief description of the purpose of
each site.
|
Web
Site |
Purpose |
|
Site
for demonstration of Bell Laboratories Text-to-Speech Software. |
|
|
Home
page for a private charity in England that does considerable research in
adaptive technology. |
|
|
Home
page for the Applied Learning Corporation, US supplier for Maltron keyboards
and other adaptive and assistive technology. |
|
|
Home
page for Aptech, and a voice recognition and voice synthesis system called
Keystone. |
|
|
Home
page for the Dragon speech recognition software. |
|
|
Site
for demonstration software for Braille translation. |
The
products described in this paper have been selected for discussion based on an
expectation that some, if not all of these technologies, may be available
generally in the economies of African countries. In addition, I have taken as a
given that in many of your agencies or enterprises, that the Windows operating
system environment running on an Intel-based personal computer is available to
the enterprise. I realize that this may not always be the case, but I have considered
this to be a baseline.
Also,
many of the resources cited here --- and the hundreds more that aren’t --- can
be researched, examined, downloaded and purchased over the World Wide Web on
the Internet. I acknowledge also the understanding that Web access is far more
restricted and more expensive to many people and organizations in Africa than
it is in Western Europe or North America.
In any
case, here are some prices of the items discussed earlier, along with
appropriate currency conversions in terms of the currencies of many of the
participants at this meeting.
|
Country |
United States |
Cameroon |
Cape Verde |
Ghana |
Guinea |
Malawi |
Senegal |
Tanzania |
Togo |
Uganda |
|
Currency |
Dollar |
Franc |
Escudo |
Cedi |
Franc |
Kwacha |
Franc |
Shilling |
Franc |
Shilling |
|
Currency |
Dollar |
Franc |
Escudo |
Cedi |
Franc |
Kwacha |
Franc |
Shilling |
Franc |
Shilling |
Code |
USD |
CFA |
CVE |
GHC |
GNF |
MWK |
CFA |
TZS |
CFA |
UGS |
|
Windows Enhancements |
$0 |
0 |
0 |
0 |
0 |
0 |
0 |
0 |
0 |
0 |
|
Maltron Keyboards |
$695 |
433,096 |
72,558 |
1,854,260 |
952,880 |
30,552 |
433,096 |
558,780 |
433,096 |
1,035,550 |
|
Bell Laboratories Text-to-Speech |
$500 |
311,580 |
52,200 |
1,334,000 |
685,525 |
21,980 |
311,580 |
402,000 |
311,580 |
745,000 |
|
Dragon Point and Speak |
$50 |
31,158 |
5,220 |
133,400 |
68,553 |
2,198 |
31,158 |
40,200 |
31,158 |
74,500 |
|
Braille Printers |
$1,000 |
623,160 |
104,400 |
2,668,000 |
1,371,050 |
43,960 |
623,160 |
804,000 |
623,160 |
1,490,000 |
|
Duxbury Braille Translator |
$595 |
370,780 |
62,118 |
1,587,460 |
815,775 |
26,156 |
370,780 |
478,380 |
370,780 |
886,550 |
This
has been an admittedly cursory and incomplete look at computer technologies
that have been developed specifically in response to the needs of persons with
one or another type of disability, ranging from motor skills, to vision and
hearing.
In
response to the opening paragraphs of this paper, it is interesting to note
that in earlier times, no accommodations were made for the development of the
human potential of persons who may have had limited sight, no physical movement
of various limbs, or other disabilities. However, with the shift of the economy
to an “information society,” our response has been to provide an entrée to this
society for persons who are disabled.
I would
like to conclude with a story of a former student of mine. At the time, I was
head of the Department of Computer Science at the University of New Orleans, in
Louisiana, USA. One day a young man came to my office --- in a wheelchair ---
to tell me his story. He had come to the university a number of years before,
straight out of high school, and had been dismissed after a year or so without
having made any grade above “D.” After this academic dismissal, he had gone to
work as a laborer in New Orleans, and in fact had done very well in that career
over quite a few years.
Unfortunately,
about a year before he came to see me, he had fallen from a bridge under
construction. And he had been rehabilitating for that period of time.
Furthermore, the Department of Labor had agreed to fund his retraining ---
although they were surprised to learn that his preference was to go back to
university to get a computer science degree.
(Incidentally,
one reason the student came to me was that he first tried to go to Tulane
University, a better-known institution. Unfortunately for Tulane, their
admissions office was on a third-storey with no elevator access.)
A few
years later, my student graduated with a Master’s degree in Computer Science.
In the interim, I had hired him to run all of our laboratories --- a
$10,000,000 plus physical plant. Since that time, he has gone on to be one of
the most successful computer system directors and consultants in the South.
He’s still in the wheelchair, at least part of the time, and he sometimes tires
easily --- but the dozens of clients who have used his services would all
concur that his talents outstrip anyone in the area.
From
this one isolated case, through all of the examples given of ways in which
persons with disabilities can function successfully in this information age, I
hope I have demonstrated that computer technology has a significant role to play
in providing the greatest of opportunities for our fellow human beings who may
not have the good fortune that some of us possess in having full use of all of
our senses, our limbs, and our bodily functions.
I look
forward to discussions with all of you throughout this conference in ways in
which we might bring these technologies to more and more persons throughout the
African continent.
Return to "Some Other Readings"