






                         BRAILLE DEVICES AND TECHNIQUES TO

                                ALLOW MEDIA ACCESS





                                    MARCH 1992











                                    Prepared by

                   Daniel E. Hinton, Sr., Principal Investigator
                                        and
                                 Charles Connolly

                  SCIENCE APPLICATIONS INTERNATIONAL CORPORATION
                         3701 N. Fairfax Drive, Suite 1001
                               Arlington, VA  22203
                                  (703) 351-7755

1.0   SCENARIO

      Braille Devices and Techniques to Allow Media Access.

2.0   CATEGORY OF IMPAIRMENTS

      Persons with vision impairments.

3.0   TARGET AUDIENCE

      Consumers with Vision Impairments.  Persons with vision impairments will
benefit from enhanced access to media information services and computer systems. 
This scenario on advanced materials and technology for implementing Braille
provides a means to disseminate information to consumers with vision impairments. 
In particular, it provides a better understanding of the technology available to
produce Braille over the next three to five years.

      Policy makers, including national representatives, Government department
heads, and special interest organizations.  Policy makers can use this scenario
to better understand the issues related to media access for persons with vision
impairments.  In addition it provides a point of departure for policy makers to
understand how advanced technology legislative or regulatory funding priorities
within Government programs can accelerate Braille output device development.

      Researchers and Developers.  This group will benefit through a better
understanding of the needs of persons with vision impairments and specifically
their printed media communications needs.  Understanding media access
requirements will assist researchers and developers in designing Braille media
access functions into their future products to meet the needs of persons with
vision impairments.

      Manufacturers.  Manufacturers will benefit through a better understanding
of Braille device requirements, the potential market size and the existing Federal
Government requirements for media access for persons with vision impairments
which can be met by adding a Braille capability to their systems.

4.0   THE TECHNOLOGY

      Louis Braille published a dot system of Braille in 1829 based on a "cell" of
six dots.  He defined the alphabet, punctuation marks, numerals, and later a
system for music using the 63 possible dot arrangements.  Braille is read by
running a finger over a character and sensing the raised dot pattern.  Braille
output devices in use today include a stylus on a pocket-sized metal or plastic
slate (analogous to a pencil and clipboard), Braille writers like the Perkins
Brailler (analogous to typewriters), and computer Braille devices.  Printed Braille
can be stamped on both sides of a page in a process called interpointing.  This
process saves paper and reduces the size of Braille books.

      The nominal specifications for Braille dot, Braille cell and Braille page
dimensions are set by the National Library Service for the Blind and Physically
Handicapped (NLS).  The NLS certifies all Braille transcribers sponsored by the
Library of Congress based on these specifications:

      Braille dots:
             --    Height for paper Braille, 0.019 inches, uniform within tran-
                   scription;
             --    Base diameter, 0.057 inches;

      Braille cell:
             --    Center-to-center distance between dots, 0.092 inches;

      Corresponding dots of adjacent Braille cells:
             --    Horizontal separation, 0.245 inches;
             --    Vertical (down page) separation, 0.400 inches;

      Braille page:
             --    Standard size, 11.5 inches wide by 11 inches high.
             --    Minimum margin for binding side, 1 inch;
             --    Minimum margin for other sides, 0.5 inch;
             --    Minimum weight of paper, 80-pound;
             --    Paper must be thick enough so that, at worst, 10% of dots
                   break the paper surface, but thin enough to permit uniform
                   dots of proper height.

      The Perkins Brailler, made by the Howe Press of the Perkins School for the
Blind, Watertown, Massachusetts, is a machine for embossing characters on paper. 
It is widely regarded as the standard for quality within the industry and has
been used for over 100 years.  It is capable of embossing 25 lines of 40
characters each, which is the page layout implied by the NLS standards.  The
term "Perkins" is used almost generically to refer to all Braille machines.

      It should be noted that even though 11x11.5 inch paper is standard, many
rely on 8.5x11 inch paper because it works well with a slate and stylus.

      For paperless Braille, approximately 20 grams of force at 0.010 inches
displacement, and 0.020-0.030 inches displacement without opposing force, may be
a useful guideline for acceptable feel.  Different technologies have different
force-displacement characteristics which Braille-literate people must evaluate on
a case-by-case basis.

      The Braille Authority of North America is the committee that sets standards
for Braille code in the U.S., and all sanctioned Braille code is based on a 6-dot
Braille cell:  2 columns of 3 dots each.  Figure 1 shows the Braille alphabet.
Nemeth Code, which is the standard Braille notation for mathematics, Computer
Braille Code, and Textbook Braille 


    1      2      3       4      5      6      7      8      9     10a
 .
. .
. .        b
 .
 .
. .                c
 
. .
. .                         d
 
. 
. .                                 e
 .
. 
. .                                         f
 
 .
. .                                                 g
 
 
. .                                                         h
 .
 
. .                                                                 i
. 
 .
. .                                                                         j
. 
 
. .k
 .
. .
 .        l
 .
 .
 .                m
 
. .
 .                         n
 
. 
 .                                 o
 .
. 
 .                                         p
 
 .
 .                                                 q
 
 
 .                                                         r
 .
 
 .                                                                 s
. 
 .
 .                                                                         t
. 
 
 .u
 .
. .
         v
 .
 .
                 w
. 
 
.                          x
 
. .
                                  y
 
. 
                                          z
 .
. 
 
                        Figure 1. Grade 1 Braille Alphabet

are all based on 6-dot Braille cells as is the literary Braille used for mainstream
text translation.

      Some paperless Braille cells are produced in the U.S. with 8 dots per cell--
2 columns of 4 dots each--but these are for compatibility in the European market. 
In Europe, the extra two dots are used to represent upper case letters and
computer characters: control characters and extended ASCII characters.  Eight-
dot Braille cells could be adopted as the standard for computers in the U.S., but
that seems unlikely for five reasons:

      1.     Six-dot Braille has a long and successful history in the U.S., and
             the cost of replacing Braille printers and paperless Braille displays
             in a short time would be prohibitive.

      2.     More lines of 6-dot Braille fit on a page than 8-dot Braille, and
             Braille already takes up several pages per printed page.  An
             embossed Braille document takes about 15 to 20 times the volume of
             the same document printed, making it even less likely that 8-dot
             cells would be adopted for Braille books.  This is based on the fact
             that it takes approximately three Braille pages per standard print
             page and a Braille page is 11x11.5 inches vs 8.5x11 inches for a
             standard text page.  Also, a Braille page is 3 to 6 times thicker than
             a standard text page.

      3.     No single standard has emerged for special computer characters in
             8-dot Braille.

      4.     One third more dots per cell would make an 8-dot Braille output
             device considerably more expensive than a 6-dot output device; the
             cost per dot generally dominates the total cost of Braille displays.

      5.     The "War of the Dots," which ended in 1918 with the choice of
             modified French Braille notation over American Braille and New York
             Point, has made Braille experts extremely cautious about making
             major changes in Braille notation.

5.0   STATEMENT OF THE PROBLEM

      Persons with visual impairments have limited real-time access to computer
information because existing Braille output devices are expensive and can only
display 20-80 characters at a time.  In the U.S., voice synthesis devices are used
by more visually impaired Americans than paperless Braille devices due to their
lower cost.  Paperless Braille displays are more common in Europe, where the
Government generally pays for displays.  Affordable paperless Braille is needed
because voice synthesis does not allow the user to quickly review material as it
appears on the monitor or printed page, including its format and structure.  With
the advent of large CD-ROMs with database libraries containing millions of print
characters, and the increasing availability of information accessible by computer,
persons with visual impairments need Braille displays that allow them equal
access to the text displayed for sighted persons on the computer monitor.  The
best Braille displays now available limit persons with severe vision impairments
to a single line of 20, 40 or 80 Braille characters.  This makes it difficult to scan
through text files and look for headings or jump from paragraph to paragraph. 
There is an urgent need for larger Braille displays to allow persons with vision
impairments text access capability equivalent to that of sighted persons.

      Several factors influence the demand for Braille displays: the rate of Braille
literacy is low among persons with vision impairments in the U.S., perhaps 20
percent. This is because, in part, visual impairments often set in with advancing
age when it is more difficult to learn Braille.  Most legally blind Americans are
elderly.  Also, age can adversely affect hearing, so there are older Braille-literate
Americans who cannot use voice synthesis technology.  The segment of the
population with deaf-blindness with little or no residual hearing, regardless of
age, also cannot benefit from voice synthesis technology.  Some people cannot use
Braille because they have reduced tactile sensitivity, as with diabetes, age, and
occupations that callous hands.  Overall, the largest demand for paperless Braille
in conjunction with computers comes from people who can use voice synthesis
technology but, because of the need to study, review and edit text, need to use
paperless Braille.

      Many people with vision impairments want the capability to produce
computer-driven Braille displays containing 3 or 4 horizontal lines of 80 Braille
characters each.  Others want 3 or 4 lines of 40-42 Braille characters.  Many
persons with vision impairments would be satisfied with a refreshable Braille
display that simulates the 25-line Perkins Brailler page.  However, size, weight,
power, reliability and cost per unit will determine the maximum Braille page size.

      Researchers should focus their attention on identifying fresh approaches
to producing the dots required to form the Braille characters within the space
limitations imposed by the Braille specifications listed in Section 4.0.

      According to Noel Runyon, an engineer at Personal Data Systems and a
Braille user, the critical factors that are easiest to overlook in the design of a
full-page Braille display include:

      1.     Speed.  Most reading is skimming, not sequential, cover-to-cover
             reading.  Also, people can learn to read Braille as fast as sighted
             people read print.

      2.     Navigation.  If display updates cannot occur in the blink of an eye,
             it is important to be able to "point" to a part of the display,
             evaluate it, and go to another page without waiting for the entire
             display to update, because everyone needs to flip through pages. 
             Single characters must be individually addressable, and readers need
             a feel for where they are on a page.

      3.     Cursor location is critical on a computer display.

      4.     Application-specific devices are too restrictive to meet the broader
             communication needs of Braille users.  For example, sequential output
             devices are awkward for most types of reading, whether their output
             is Braille or speech.

      5.     Humble things like dust can render laboratory successes almost
             useless in real homes and offices.

      6.     Graphics capability is a major justification for the use of a full-page
             display rather than a smaller display.

      7.     Battery power is highly desirable.

      8.     Noise is an important factor, especially in offices, libraries, and
             other public places.

      9.     Cost can make the difference between a device that is evolutionary
             and a device that is revolutionary.

      10.    Elderly and pre-employment-age people have generally received the
             least attention when developing new Braille technology, so they
             tended to be left out of the decisions that led to existing devices.

6.0   THE DEPARTMENT OF EDUCATION'S PRESENT COMMITMENT AND
      INVESTMENT

      The primary reason that Braille media access is a priority is because
approximately 100,000 Americans with vision impairments use Braille for written
communication.  According to the 1988 National Health Interview Survey, 600,000
Americans between the ages of 18 and 69 have blindness or visual impairments
severe enough to limit their employment opportunities, and that number rises
sharply with age.  This is an indication of the size of the population who could
potentially benefit from Braille literacy.  Although the number of visually
impaired people under 18 is relatively small, they can learn Braille most easily
and use it for the rest of their lives, thus they can gain the most from Braille
literacy.

      The Department of Education, and its predecessor, the Department of Health
Education and Welfare (HEW), have funded Braille device research and develop-
ment over the past 20 years.  With the advent of personal computers in 1975,
HEW began to fund research and development of computer Braille output devices
such as the TeleBrailler, and MicroBrailler.  Currently, the development of Braille
capability is a stated research priority of the Department of Education as follows:

            The Electronic Industries Foundation (EIF) Rehabilitation Engineering
             Center's Technology Needs Assessment Paper, "An Inexpensive
             Refreshable Braille Display," points out a need for a "low-cost,
             reliable paperless Braille display mechanism."  That report follows up
             on the recommendations of the National Workshop on Rehabilitation
             Technology, sponsored by EIF and the National Institute on Disabili-
             ties and Rehabilitation Research (NIDRR).  The Workshop recom-
             mended making "information processing technology for access to
             print graphics, including computer access" the top technology
             priority for visual impairments.

            Several of the funding criteria of the Department of Education's
             National Institute on Disability and Rehabilitation Research (NIDRR)
             are directed at the high unemployment and underemployment rate of
             persons with vision impairments and severely visually impaired
             populations.  Most severely visually impaired Americans are unem-
             ployed.  Larger and more affordable Braille displays would improve
             the educational outlook of blind individuals, promote Braille literacy,
             and improve employment opportunities and job retention among the
             Braille literate.  Another stated priority, advanced training for the
             blind and visually impaired at the pre- and post-doctoral levels, and
             in research, would benefit greatly from improved Braille display
             technology.

            The Panel of Experts for the Department of Education program
             sponsoring this study consists of experts from industry and
             Government, including members of the sensory-impaired community. 
             Their consensus opinion was that developing a larger Braille display
             is the highest priority for persons with visual impairments.

            One of the Department of Education's 1991 Small Business Innovative
             Research (SBIR) Program Research Topics is to develop or adapt
             communication devices for young children who are blind or deaf-
             blind.  An affordable Braille display could be used for games that
             would help young children develop the skills needed to read and
             write Braille.  A Braille display would also be of some use for tactile
             graphics, though an evenly spaced array of dots based on the same
             technology might be better.

            The Department of Education's NIDRR Program Directory, FY89, lists
             the Smith-Kettlewell Rehabilitation Engineering Center, among many
             other tasks, as testing, developing, and/or evaluating a Braille
             display technology.

7.0   ACCESS TO COMMUNICATIONS MEDIA

      Many federal, state, and local laws which influence access for persons with
visual impairments.  The most important single law related to access for persons
who are vision impaired is Public Law 101-336, enacted July 26, 1990.  Better
known as the Americans with Disabilities Act (ADA), this law has broad
implications for all disabled Americans and establishes the objective of providing
access to persons with disabilities to physical and electronic facilities and media.

      The other law that impacts technology for persons with visual impairments
is Public Law 100-407-AUG. 19, 1988 titled "Technology-Related Assistance for
Individuals with Disabilities Act of 1988."  Also known as the Tech Act, this law
established a comprehensive program to provide for technology access to persons
with disabilities.  The law defines assistive technology devices:

"Assistive technology devices means any item, piece of equipment, or
product system, whether acquired commercially off the shelf, modified, or
customized, that is used to increase, maintain, or improve functional
capabilities of individuals with disabilities."

      Braille technology clearly meets this definition for persons with vision
impairments and should be exploited to increase the ability  of persons with
vision impairments to obtain access to printed media.  Within the findings and
purpose of this law, Braille technology can provide persons with vision
impairments with opportunities to:

            exert greater control over their own lives by making literacy
             possible;

            participate in and contribute more fully to activities in their home,
             school, and work environments, and in their communities;

            interact with nondisabled individuals; and 

            otherwise benefit from opportunities that are taken for granted by
             individuals who do not have disabilities.

8.0   POTENTIAL ACCESS IMPROVEMENTS WITH ADVANCED BRAILLE
      TECHNOLOGY

      Table 1 shows a sampling of the Braille technology currently available.  The
base price of adding paperless Braille to a computer is now about $5000.  This
high cost forces many persons with visual impairments in the U.S. to use voice
synthesizers which costs about $1000.  Braille embossers starting at approximately
$1700 for the Braille Blazer, cost about three times that cost of text printers
used by the sighted population.

      Advanced Braille technology offers persons with visual impairments the
potential for dramatic improvements in access to books and periodicals stored in
computer-readable form or scanned.  For example, at least 800 titles are already
available on CD-ROM and that number will probably increase rapidly in the years
to come.  Another important access improvement would be to computer-based
telecommunications, including databases, electronic mail systems, computer bulletin
board systems and mail order systems, all of which generally consider a computer
screen as a single unit.  One-line paperless Braille displays have been a cost
compromise when compared to the speed and agility that a full-screen display
could offer.

      It is often desirable to skim text for relevant information, whether that text
is a computer's display, magazine or newspaper article, or book.  When skimming,
the field of the display needs to be as large as possible.  The only practical
alternative is Braille paper output, but relying on a Braille paper printer (priced
for individual use) is slow and paper-intensive.  A multiple-line paperless Braille
display offers tremendous improvements in skimming speed and effectiveness over
existing Braille printers and single-line displays.  It also would have a great
impact on the ability of persons with vision impairments to do research and
academic study, which often requires reading and rereading information.

9.0   ADVANCED BRAILLE TECHNOLOGIES

      There are two major approaches to producing paperless Braille.  The
simplest approach is to apply constant power to keep each dot raised or lowered,
but many of the technologies used to move dots require a substantial amount of
power (50 to 100 milliwatts per cell).  An analysis of the power available to a full
page Braille display provides insight into the power that can be allocated to each
Braille dot/cell.  In older houses, standard electrical outlet can provide about
1200 watts of power.  About 250 watts of that must be allocated to the computer
controlling the display, leaving 950 watts for the Braille display.  Assuming the
display's power supply is 50% efficient, that leaves only 475 watts of power in
the form the display can use.  An 80-cell display with 6 dots per cell can allocate
almost 1 watt per dot; 8 dots per cell lowers that to about 0.75 watts per dot. 
A standard Braille                  Table 1.  A Sampling of Existing Braille Products
         Note:  Prices range from 1989-1991 so they may not be comparable.
   Brand Name    Manufac-
turer                          Price   System            DescriptionHARD COPYPerkins
BraillerPerkins
School for
the Blind$395-
$730NoneBraille Writers, Manual and
ElectricMountbattenHumanWare
Inc.$2595-
$3170NoneBraille Writer, ElectronicIndex
Braille Em-
bossersHumanWare
Inc.$2895-
$16,900IBMBraille EmbosserBraillo 90Braillo Nor-
way AS$5795IBMBraille EmbosserBraillo 200Braillo Nor-
way AS$39,995IBMBraille EmbosserBraillo 400
SBraillo Nor-
way AS$78,995IBMBraille EmbosserRomeo
BraillerEnabling
Technologies
Company$2695-
$3450AllBraille EmbosserMarathon
BraillerEnabling
Technologies
Company$11,500AllBraille EmbosserTED-600
Text
Embossing
DeviceEnabling
Technologies
Company$37,500AllBraille EmbosserBraille Bl-
azerBlazie Engi-
neering$1695AllBraille EmbosserATC/Resus
214 PrinterAmerican
Thermoform
Corporation$15,995AllBraille EmbosserVersapoint-
40 Braille
EmbosserTelesensory
Corporation$3795AllBraille Embosser/TranslatorOhtsuki
BT-5000
Braille/Prin
t PrinterAmerican
Thermoform
Corporation$5180IBM
AppleBraille Embosser/PrinterDuran
Dots-40Arts
Computer
Products
Inc.$710-
$1510IBMAdapter to Convert Brother
HR-40 Daisy Wheel Printer
for Braille PrintingStereo Copy
Developing
MachineMatsumoto
Kosan Com-
pany$6250NoneBraille CopierThermoform
Duplicators
for BrailleAmerican
Thermoform
Corporation$1750-
$2895NoneBraille CopiersPlate
Embossing
Device
PED-30Enabling
Technologies
Company$62,500NoneBraille Plate Embosser for
Printing Houses              TACTILE READING SYSTEMOptacon IITelesensory
Corp.$3495-
$3995AllPortable Tactile Reading
SystemInTouchTelesensory
Corp.$395MacOptacon II Accessory Soft-
ware for Mouse AccessOptacon PCTelesensory
Corp.$395IBMOptacon II Accessory Soft-
ware/Hardware for Mouse
Access                      ONE-LINE BRAILLE DISPLAYSBraille
Display
ProcessorTelesensory
Corp.$3695IBM
ApplePaperless Braille
20 CellsBraille
Display
Processor
BDP 21Telesensory
Corp.$3695IBMPaperless Braille/Translator
20 CellsBraille
Display
Processor
BDP 20Telesensory
Corp.$3695ApplePaperless Braille/Translator
20 CellsBraille
Interface
TerminalTelesensory
Corp.$3995IBMPaperless Braille
20 CellsNavigatorTelesensory
Corp.$3,995-
$14,995IBMPaperless Braille
20,40,80 CellsVersaBraille
II+Telesensory
Corp.$5995IBMPortable Paperless Braille
20 CellsKeyBrailleHumanWare
Inc.$5025-
$7025ToshibaPaperless Braille
20,40 CellsAlvaHumanWare
Inc.$8,995-
$14,495IBMPaperless Braille
40,80 CellsBraillex
IB80Index Inc.$14,495IBMPaperless Braille; 80 CellsNew Ability
BraillerDensttron
Corp.$2995IBMPaperless Braille (Soft
Braille)
40 Cells                     BRAILLE NOTES/COMPUTERSNotexIndex Inc.$5800-
$7900IBMPortable Braille Notetaking
Device/Computer with 20- or
40-Cell Paperless BraillePersonal
TouchBlazie Engi-
neering$5500AllPortable Braille Notetaking
Device/Computer with 20-
Cell Paperless BrailleBraille 'n
SpeakBlazie Engi-
neering$905AllPortable Braille Notetaking
Device/TranslatorSpeakSysBlazie Engi-
neering$149IBMBraille 'n Speak InterfacePocketBra-
illeAmerican
Printing
House for
the Blind$905AllPortable Braille Notetaking
Device/Word ProcessorEureka A4Robotron
Access
Products
Inc.$2595IBMPortable Talking Computer
with Braille KeyboardNomadSyntha
Voice Com-
puters Inc.$2295AllPortable Talking Computer
with Braille Keyboard OptionFOR THE DEAF-BLIND POPULATIONAFB
Tellatouch,
MS 170American
Foundation
for the
Blind$595NoneTypewriter Keyboard
Controlling a Paperless
Braille Cell for 1-Way 1-on-1
CommunicationDiaLogosFinnish
Central
Association
of the
Visually
HandicappedNoneBraille Keyboard with Six
Paperless Braille Cells
Connected to a Typewriter
Keyboard with 1-Line Dis-
play (TDD) for 1-on-1 or
ASCII or TDD Modem Commu-
nicationInfoTouchEnabling
Technologies
Company$4000-
$4900NoneBraille or Typewriter Key-
board Connected to a Romeo
Brailler and a Typewriter
Keyboard with 1-Line Dis-
play (Superprint TDD) for
1-on-1 or ASCII or TDD
Modem CommunicationTeleBrailleTelesensory
Corp.$5500NoneBraille Keyboard and 20-Cell
Paperless Braille Display
Connected to a Typewriter
Keyboard with 1-Line Dis-
play (Superphone TDD) for
1-on-1 or ASCII or TDD
Modem Communication
page with 6 dots per cell could allocate just under 0.08 watts per dot; 8 dots per
cell lowers that to 0.06 watts per dot.  An 80-cell by 25-line Braille display,
which could provide full text access to an IBM-compatible personal computer
screen, could allocate just under 0.04 and 0.03 watts per dot, for 6- and 8-dot
cells, respectively.  Without any blank lines, a page of Braille text could be
expected to have an average of 2 dots raised per cell, so, if only raised dots
require power, that would mean the typical power available per dot would be
about 3 times the minimum values for a 6-dot cell (4 times the minimum values
given for an 8-dot cell).  Unless the display is being used for graphics, it would
be unrealistic to expect all dots to be raised at once.  On the other hand, it
would be unwise to design a display so that raising all the dots would blow a
fuse or trip a circuit breaker in the user's home or office.  A compromise may
be necessary for very large displays but it is desirable to have the capability
to raise or lower all dots simultaneously.

      Applying continuous power to the actuators is impractical for many Braille
display actuator technologies because even a one-line display would require more
power than a wall socket can provide; far more than a portable battery system
could tolerate.  Therefore, many paperless Braille displays raise or lower dots
and then lock them into position until another page is displayed.  Historically,
the locking and unlocking mechanisms have required little or no power except
while displaying a new page.  In practical operation, these locking mechanisms
reduce average power consumption by several orders of magnitude.  The problem
with locking mechanisms has been that they increase mechanical complexity, which
tightens the manufacturing tolerances.  Reliable actuators for Braille cells are
available today but most of them require a locking mechanism to avoid excessive
power requirements.  Even if power constraints could be ignored, some actuators'
locking mechanisms double as a way of ensuring that dots are raised to a
uniform height, which is a requirement for Braille.

      Designs that employ locking mechanisms update the display dot by dot, cell
by cell or in small groups of cells.  This minimizes the peak power consumption
by decreasing the display update rate.  Alternatively, a storage device could
slowly accumulate energy from the power source and release it all at once; which
is how a portable camera flash works.  With the more energy-intensive actuator
technologies, a tradeoff is necessary between display size and refresh rate of the
display.  The inherent size and weight of most Braille display technologies
usually justifies slowing the display update time moderately.  Portable Braille
displays are almost certain to require tradeoffs in power vs refresh rate because
both average and peak power capabilities of batteries are strictly limited by
acceptable battery size, weight, and frequency of replacement or recharge.

      According to a 1990 Smith-Kettlewell study, a one-line Braille display that
slides up and down a "page" provides some of the advantages of a full-page
display.  The study, by TiNi Alloys, Oakland, California, and Smith-Kettlewell, San
Francisco, suggests that a six inch long virtual page can even create the illusion
of a full size page of Braille.  This work may lead to an alternative approach to
providing the feel of a full page Braille device in a simpler and more reliable
format.

      Solenoid electromagnetic actuator technology has been most often tried for
producing Braille.  Tight packing is needed for displays of useful size, even with
coil assemblies and components fabricated with truly miniature solenoids. 
Historically, solenoids have been power-intensive (requiring locking mechanisms)
and prone to failure with dirt from normal use (i.e., grease, skin cells, pollen,
and even volcanic ash).  This leads to reliability problems because cleaning 6000
solenoids regularly for a full page display would not be a realistic option. 
Covering the solenoids with a protective plastic membrane keeps the solenoids
clean, but slightly moist fingers skip across plastic so a plastic surface is
undesirable.  Power requirements and interference between neighboring solenoids
are also problems that must be overcome.  Developments in superconducting
materials, and in motor and solenoid miniaturization, may help to solve the
problems associated with large electromagnetic Braille display fabrication.

      Metec (Stuttgart, Germany), EHG (Nordstetten, Germany), and Tiflotel
(Calolziocorte, Italy) have each produced electromagnetic (solenoid) Braille cells
that are scalable to multi-line or full-page displays.  But the technology was less
than successful because of a combination of reliability and repair problems. 
Power requirements may have also been a factor.  Clarke and Smith International
(Surrey, England), has produced small quantities of electromagnetic Braille cells,
but they were limited to two-line displays.  Novanik (Karlstad, Sweden) was
working on a 42-cell 29-line electromagnetic display as of 1987, but its status is
unknown.  Generally, companies seem to have given up on using electromagnetic
actuators for Braille.  Smith-Kettlewell's proprietary design, described later in
this document, is the exception.

      Piezoelectric benders, sold in the U.S. are used in all mass-produced
refreshable Braille displays with more than a few characters.  Called bimorphs,
the benders can be made with any of several materials.  Lead zirocnate and lead
titanate ceramics seem to be the most popular for Braille cells but other
piezoelectric materials include single crystals such as Rochelle salt and ceramics
such as barium titanate.  Piezoelectric materials flex in the presence of an
electromotive force.  The piezoelectric Braille cells made by Telesensory in
Mountain View, CA, are considered by many visually impaired people to have the
best feel of any Braille cell available in the U.S.  The Tieman cell, another
popular piezoelectric Braille cell, is probably made by Kogyosha (Tokyo) working
with Braille Equipment Europe in the Netherlands.  It is used in the Alva and
Braillex displays, and possibly the KeyBraille and Notex displays.  Metex, and
possibly other electromagnetic Braille cell manufacturers, have switched to making
piezoelectric Braille cells.  The current state of the art in piezoelectric benders
limits refreshable displays with horizontal benders to one or two lines, and only
single-line displays are commercially available.  The reason for the size limitation
is that piezoelectric benders bend very little per unit length, so they have to
be much more than an inch long to obtain the bending motion necessary to lift
the Braille dots into place.

      Elinfa, in France, developed vertical piezoelectric benders for Braille cells,
reducing the area required per dot by a factor of five.  In theory, the Elinfa
cell used in the Personal Touch, could be used to produce a very large display
but, to date, producing Braille cells with horizontal or vertical piezoelectric
benders has been expensive and labor-intensive.  New manufacturing technologies
may be needed to overcome this problem.  Ceramic piezoelectrics tend to be
brittle, and single-crystal piezoelectrics tend to have other undesirable
properties.  Single-crystal Rochell salt, for example, has the strongest piezoelec-
tric effect known; but its dielectric properties have led to the use of ceramic
piezoelectrics instead.  The retail price of piezoelectric displays is  $20-25 per
dot, which would easily put the retail price of a full-page piezoelectric Braille
display over $100,000 per unit.  This is far outside the price range of the typical
user.

      Five factors have enabled piezoelectric displays to dominate the paperless
Braille display market.  First, although driving piezoelectric cells requires on the
order of two hundred volts direct current (VDC), the average current is low
enough that they have very low net power requirements.  The Telesensory
Navigator's power supply would only allow a maximum of 0.023 watts per dot. 
Power consumption is so low that all dots can be raised continuously, thus
eliminating the complexity and reliability problems associated with locking
mechanisms.  Piezoelectrics are actually their own locking mechanisms, requiring
no power to stay in position except to cancel leakage currents.  Also, low power
consumption allows the option of portable battery power for small displays. 
Second, each dot has few moving parts, the bender and the dot shaft, in the
case of telesensory's Braille cells and there is no friction-based locking
mechanism.  The result is that piezoelectric displays are relatively immune from
dirt and wear, though dirt can cause dots to stick, requiring ultrasonic cleaning. 
Piezoelectric displays are reliable due to a minimum number of moving parts and
minimal friction. Third, piezoelectric displays can provide fast display updates
because they are energy-efficient.  Fourth, piezoelectric displays are very quiet. 
They make just enough noise to let the user know an update has occurred. 
Finally, the dots can be closely packed and therefore come very close to the
standard Braille dimensions.

      The next generation of full page Braille displays must be able to provide
refreshable Braille for significantly less than $20-25 per dot.  It is very unlikely
that a full-page Braille display could be sold for much more than the $15,000
price of existing 80-cell displays.  This means a 25-line, 40-character display with
6 dots per cell, would have to be produced at a cost of $2.50 per dot.  A
significantly lower price, perhaps $1 per dot, would open up a much larger
market for full-page Braille displays and serve many more persons with vision
impairments.

      Tactiles is working on a machine with very low cost self-locking dots
(<$.10) and a travelling "printhead" similar to a dot matrix printer. The target
price is under $4000.

      The reliability needed for every dot in a Braille display is significant.  For
example, if every dot in a full-line Braille display (480 dots) worked 99% of the
time the display would be error-free once in 125 displays.  If the dots were
99.99% reliable, the 80-cell display would be error-free only 95% of the time and
the full-page display would be error-free about 55% of the time.  At 99.9999%
reliability, the one-line display would have an error every 2000 lines, but the
full-page display would still have an error once in about 165 pages.  This makes
the design of a reliable full-page display difficult.  Moving parts tend to make
a device unreliable, but a Braille display must have 6000 independently moving
parts, each with reliability much greater than 99.99%.  Telesensory's Braille cells
have been around for a long time and are extremely reliable, but there have been
considerable differences in reliability among the manufacturers.  Also, character-
istics of the user such as sweaty hands and a tendency to eat potato chips,
affect reliability.

      A major technology shift is required to design a full page Braille display
to meet the media access needs of persons with vision impairments.  This new
technology shift would incorporate advanced materials and computer control
technologies.  Advanced materials and manufacturing technology may make it
possible to implement several lines of Braille; perhaps a full 40-character by 25-
line page of Braille output.

      An example of a technology improvement that could facilitate the
implementation of full-page paperless Braille is large array controllers for liquid
crystal displays (LCDs).  These LCD controllers can control 64 high-voltage lines,
on the order of 120-180 volts direct current (VDC) from a single chip and could
be used to control 10 piezoelectric Braille cells.  They may also be useful for
switching electrorheological fluids, which will be described later.  Other LCD
controllers are available that could control 20 or more elements.

      Before discussing the newer technologies, a review of earlier attempts at
a full-page paperless Braille display is useful to prevent repeating mistakes.

Historical Developments

      Thermostat metals were used in Braille Inc.'s prototype Rose Braille Display
Reader, a full-page display patented in 1981 but was never commercially
produced.  The Texas Instruments thermostat metals used are bimetallic strips
that bend when heated.  In the Rose Reader, the shaft of each Braille dot has
a grooved ring around it.  The shaft would be pushed up by a spring, but a
hook on the end of a bimetallic strip catches the groove on the ring and
restrains the dot.  When heat bends the bimetallic strip away from the ring, the
spring raises the dot.  A separate manually activated mechanism pushes all of the
dots down again and signals the machine to display the next page.  The unit
included a panel of 12 control buttons and a cassette drive for storing text.

      According to Leonard Rose, the one prototype that was built had some dots
that did not work because the device was handmade; possibly machine-made parts
would have been more reliable.  Unfortunately the only accurate way to measure
reliability very near 100% would be to manufacture parts.  This would require a
considerable investment.  According to Mr. Rose, putting the system into
production would cost about $750,000 with units eventually selling for as little
as $7500.  The thermostat metals used in the Rose Reader are less expensive than
piezoelectric elements, and can be designed in modular units for easier repair. 
However, the number of moving parts per dot, the direct use of heat and
friction, and the use of a manual mechanical reset mechanism are all potential
sources of reliability problems.  In principle, replacing the manual reset lever
with an electric one is easy, but that is likely to affect reliability.

      The Rose Reader requires raising the temperature of the metal strips by
30 degrees Fahrenheit to overcome friction.  This requires significant power, so
the dots are raised one at a time, at a rate of 200 dots per second.  If all dots
are raised, the total time required for a full display is about 30 seconds,
although the average Braille page would take approximately 10 seconds.  Based
on an estimate from the patent information, if a page could be displayed within
one second, it would trip a 15 ampere circuit breaker, even with only one third
of the dots raised.  Each dot requires on the order of a watt-second of energy
to be actuated.  This is because the temperature changes required in thermostat
metal actuators that have to overcome friction make them relatively energy-
intensive.

      It is not clear whether sufficiently reliable mechanical locking mechanisms
are available to circumvent higher energy requirements, but there seems to be
a pattern.  Energy-intensive actuators with lower materials costs tend to require
locking mechanisms that cost as much to manufacture as more energy-efficient
actuators for a given level of reliability.  In the end, the prototypes with locking
mechanisms are generally too costly to manufacture or too unreliable to sell. 
When the development funding for these devices runs out, the designs are
shelved indefinitely.  What is needed is an actuator that is energy-efficient
enough to be used without a locking mechanism, yet costs less than $2.50 per dot
or less and fits in a standard Braille cell profile.

      From the late 70's to the mid-80's, the American Foundation for the Blind
(AFB) experimented with injection-molded arrays of 64 x 64 dots, manipulated one
row at a time by a single row of 64 solenoids, one row at a time.  Four of these
prototypes were built with the combined capability of producing a full page of
Braille or graphics.  Graphics capability turns out to be a mixed blessing because
evenly-spaced dots are incompatible with standard Braille dimensions but would
be a major advantage of a full-page display over a one-line display.  The system
had three major advantages: the pins were mechanically latched into position so
power consumption was moderately low (because the system was slow); the feel
of the display was good; and since the system was modular, a mechanism for
repair by replacement was provided.  There were two major problems:  the 64-
step display update was slow enough to offer no great advantage over paper
output, and the system was expensive.  Ultimately, the cost of the mechanical
system was its downfall, and the method was not recommended for further
development.

      Shape memory alloys were the technology used by TiNi Alloys, to develop
a 20-cell by 3-line prototype display of 8-dot cells.  Shape memory alloys are
nickel-titanium alloys that forcefully return to a preset shape when heated and
are usually alloys of nickel and tin, or of copper, zinc and aluminum.  In this
case, TiNi used a nickel-titanium (nitinol) allow in the form of a wire, one inch
long and 0.030 inches in diameter.  When an electric current heats the wire, it
shortens, pulling the shaft of a dot down against the force of a spring.  Each
dot has a small flexible piece of sheet metal with a hole big enough to let the
shaft of the dot pass through it if the metal is lying flat, but small enough that
it catches the dot if the piece of metal is angled.  The metal's resting position
is angled.  When a dot is lowered, its shaft catches on the piece of sheet metal
and pulls it down flat enough to let the shaft pass through.  When the wire
cools, it stops pulling the shaft of the dot down, and the dot tries to spring up
again, but that pushes the sheet metal back to its angled position, catching the
dot's shaft so it cannot move.  When a plate pushes all of the pieces of sheet
metal flat, the dots are then free to move and all raise, clearing the module. 
Stops were used to give the display adequate feel.  The display was built of 4-
cell units because modularity makes it easier to build and repair a display. 
Funded by the National Institute of Health, the project did not get past the
prototype stage, though it received an Excellence in Design award from Design
News, a respected journal for design engineers.

      The technology was capable of displaying a full Perkins Brailler page.  In
fact, the software was written for an 80-cell by 25-line display but there were
reliability problems with the detent and release mechanisms, which required tight
manufacturing tolerances.  With further development, the technology might have
become cost-competitive with other Braille cells, but the two-year grant ended
in 1990.  An important cost driver was making and attaching the special metal
wire, although better techniques are now available.  Power requirements were 50
watts instantaneous, or about 1 watt per dot.  The dots were given that power,
one module (32 dots) at a time, for 50 or 60 thousandths of a second, so the
energy requirement per dot is 0.050 to 0.095 watt-seconds.  That is much better
than thermostat metals, but not as efficient as piezoelectric technology.  Shape
metal alloys were a valuable experiment for paperless Braille, but their cost and
power performance seem unlikely to do much more than match piezoelectric
technology.

Current Developments

      For several years, Smith-Kettlewell has been working on a proprietary
electromagnetic Braille cell technology funded by NIDRR.  It is limited to displays
of 80 characters or less, but the cost is estimated to be $20 per cell, which is
below the cost of piezoelectric displays.  Few details are available, but the
technology has a fast refresh rate and has the potential to be used in portable
systems operated from battery power.  Smith-Kettlewell has passed the design on
to a developer, though no estimated date for production was given.

      During the past six months to a year, Blazie Engineering, Street, Maryland,
has been working on a pneumatic display that uses puffs of air to move tiny
bearings supporting Braille dots.  No product is anticipated until 1993.  The
device requires a spacing approximately 0.015 inches greater than the standard
distance between Braille cells.  The display is supposedly scalable to a full page. 
Preliminary cost estimates are as low as $5 per cell, which would be less than $1
per dot.  The feel of the display is said to be solid, and it is expected that the
present refresh rate can be increased.  Power requirements are predicted to be
low, and a 20-cell prototype has been built.  The display has two moving parts
per dot and can be cleaned by immersing it in liquid.  Performance and cost
predictions based on the prototype must be considered preliminary.

      Recent advances in sequential soft-copy Braille displays have been made
by Tactilics, Inc. and Densitron Corporation.  Sequential soft-copy Braille displays
are essentially belts that move across a "window" while Braille dots are raised
on their surface. Densitron has been selling prototypes of a 40-cell deformable
plastic disposable belt device for $2995.  Its lack of navigability would appear
to limit its use.

      Tactilics' belt is made of hard, molded nylon cell sections which they
indicate is long lasting and self-cleaning.  They claim its bi-directional control
makes it highly navigable and that it is a true realtime "monitor."  Also, when
battery powered, the unit may be used as a portable "book" and that its mixture
of high and low tech is a price breakthrough.  Two units will be introduced in
mid-1992: 1 50-cell for $1200 and an 85-cell model for $1500.

Future Developments

      What lies beyond the existing systems is impossible to predict with
certainty because, though completely new technologies are seldom discovered, old
ones are constantly revitalized by new computer capabilities, materials, and
manufacturing processes.  Sometimes older technologies suddenly become practical
due to material or other technology breakthroughs.  Some companies are unwilling
to discuss technologies they are considering for paperless Braille.  Blazie
Engineering suggested three:  magnetostriction, electrorheological (ER) fluids, and
polymer gels.

      Magnetostriction is the property of some alloys that cause them to
forcefully expand in a strong magnetic field.  "Giant" magnetostriction, an
expansion on the order of 0.15%, occurs in alloys of certain rare-earth elements. 
An alloy of iron with terbium and dysprosium is used in an actuator sold by
Edge Technologies in Ames, Iowa.  There are serious problems with using that
technology for Braille.  Rare-earth elements are not really rare, but they are
expensive to purify.  The alloys used must be in the form of a single crystal
which is presently expensive to refine and produce.  Finally, the effect of
magnetostriction is too small to be used directly without long pieces of the alloy,
which may be both voluminous and cost-prohibitive.

      Levers are being tried for converting some of the force from the actuators
to linear displacements that would be adequate for Braille.  A hard limit seems
to be that the cost of an individual actuator is still not competitive with
piezoelectric technology, and the cost of coils to produce the magnetic field,
exceeds the cost of the special alloys as the actuators gets smaller.  As with
piezoelectrics, benders can theoretically replace levers as a way of trading force
for increased movement, but the single crystals are brittle.  No one knows if
their tensile strength is such that the elements will break if used in benders. 
Power requirements are estimated at a maximum of 10 watts per dot, which is
high, but locking mechanisms may allow power management strategies with low
average power requirements.

      Other magnetostrictive materials exist, but it is not clear that any provide
enough of an effect to be useful for Braille cells.  Magnetostrictive ribbons,
which are used in sensors, provide extremely high efficiency, but they lose most
of that efficiency in strong magnetic fields, thus limiting their maximum
expansion.  Magnetostriction has its highest efficiency when the actuator is
moving back and forth rapidly, though that problem might be solvable by making
the dots move back down slowly, thus extracting a displacement as close to the
maximum as possible.  In summary, magnetostriction does not appear to be a cost-
competitive technology for Braille cells, though this may change with future
breakthroughs in materials technology.

      Fundamental research on electrorheological (ER) fluids is being conducted
at the University of Michigan (UM), Ann Arbor.  In 1988, a UM scientist made an
important breakthrough in electrorheological fluids development.  ER fluids
thicken when a strong electric field is applied to them, on the order of 2000
volts per millimeter.  Their consistency changes from liquid to something "more
like Velveeta cheese."  This allows hydraulic actuators to be constructed.  ER
fluids stop flowing while in a strong electric field, so, as a hydraulic fluid, they
can selectively apply pressure to actuators.  Per hydraulic switch, power
requirements are lower than piezoelectric technology but a pump is required to
supply the pressure for the hydraulics, and therefore their overall efficiency is
unclear. 

      The breakthrough at UM was to find an inexpensive ER fluid that does not
contain water.  Water content lowered the efficiency and predictability of
previous ER fluids making them impractical for actuators.  The fluids used at UM
are inexpensive but the particles suspended in them tend to separate from the
liquid. More expensive ER fluids do not have this problem.  Three problems are
likely to arise with ER fluid-based Braille cells: the use of a liquid, difficulty
with modularizing a system with fluid lines, and fluid pump power and noise. 
Without modules, a large Braille display could be very difficult to build and
repair.  There may be ways to modularize a hydraulic display.  Pump power and
noise may not turn out to be an issue, but the use of a liquid seems likely to
be a challenge to developers.  The need for intense electric fields could be
reduced by using narrow gaps.  Overall, ER fluids may be feasible for Braille cell
development in the immediate future.

      Polymer gels are another promising technology for full page Braille displays
is polymer gels.  Polymer gels collapse when exposed to intense light.  These gels
are being developed at the Massachusetts Institute of Technology's (MIT)
Department of Physics and Center for Materials Science and Engineering.  Under
the proper conditions, gels can be induced to reversibly release a large portion
of their liquid content.  This is called collapsing, because releasable liquid
content increases the volume of a gel by factors ranging up to 350 or more and
multiplying their length, width and height by a factor of 7.  In 1990, researchers
at MIT induced a light-absorbing gel to collapse by heating it with a visible laser
after having induced gels to collapse with exposure to ultraviolet rays, voltages
on the order of 5 volts, and changes in the surrounding liquid's temperature,
composition, pH, and salt content.  The visible light has the advantage of safety
and speed over ultraviolet radiation, as well as providing a controlled way to
induce small temperature changes through a sealed container.  The sealed
container is necessary because, to reabsorb the liquid, a collapsed gel must be
immersed in the liquid.  The liquid and gel have to be separable to exploit the
volume change of the gel, but gel reaction times below a second require gels
significantly thinner than a human hair.

      To manage fine fibers, researchers in Japan have formed gels into sponges
or bundles of fibers, but reaction times are still greater than one second.  The
leading light-sensitive gel researcher at MIT, Dr. Toyoichi Tanaka, estimates that
strands of gel one thousandth of a millimeter in diameter, about the diameter of
muscle fibers, would react as fast as muscle tissue.  Doubling the diameter of a
gel quadruples its reaction time, though; some gels react very slowly with modest
increases in fiber diameter.  So far, heating a gel with a laser is moderately
power-intensive.

      Minimal research and development could conceivably reduce the power
required by several orders of magnitude.  The choice of gel material, the
concentration and choice of light-absorbing material in the gel, and other factors
could significantly reduce power requirements.  According to Tanaka, red lasers
work better than the violet-blue laser that was used to estimate power efficiency. 
Diode lasers with power efficiencies of over 30% are now available that emit red
light.  Though diode lasers with lenses cost on the order of $50 a set, and that
is for lower power and in quantities of a thousand, if one or more lasers were
scanned over a Braille dot array with a mirror, gel technology might make it
possible to implement a reliable full-page Braille display with reasonable cost,
size, weight, and power.

      Tight temperature control would be a potential problem (i.e., temperatures
held within + one degree), but the MIT researchers have experimented with gels
at room temperature without temperature regulation, and in water.  The MIT
group uses a low-cost gel, which is also encouraging.  The temperature at which
a gel collapses can be controlled by the proportions of two liquids into which the
gel is immersed.  Even if some temperature regulation turned out to be
necessary, advances in solid-state Peltier effect heating/cooling might be
applicable, balancing the power requirements of laser(s) with those of a
temperature regulation system.  It is too early to predict whether the feel of a
gel-based Braille cell would be adequate, but the technology shows promise with
additional research and development.

      According to "Tactile Displays for the Visually Disabled--A Market Study,
July 1987," published by the Swedish Institute for the Handicapped, materials
that expand with moderate heating have been tested for application to Braille
displays, apparently without success.  That reference does not indicate what
material was tested, but it would not have been a polymer gel.  Gels could be
used for a phase change, but that phase change could not be accurately
described as a transition from a solid to a fluid.

      The piezoelectric materials currently used in Braille cells are not the only
ones available.  A study of recent alternatives would be worthwhile.  For
instance, A.V.X., in Myrtle Beach, South Carolina, started selling lead zirconate
titanate (PZT) in multiple layers around the beginning of 1991.  It appears to be
possible to get actuators made from this material for less than $20 each, in
quantity, making multilayer PZT marginally competitive with existing piezoelectric
displays.  Layering reduces voltage requirements, but it is unclear whether the
increased materials costs would be justified.  A tough plastic film, called
polyvinylidene difluoride (PVDF), sold by Atochemwith the trade name of Kynar,
may also be useful for Braille cells.  It can be used at very high voltages,
compared to ceramic piezoelectrics, but its efficiency is lower than that of
ceramics.  It is apparently better suited to vibrating Braille dots than static
ones, for reasons that include power efficiency, but the feel of vibrating Braille
dots is not as good as the feel of static dots.  Further study is needed to
determine whether these and other materials are appropriate for use in  Braille
cells.  In particular, their response at low frequencies must be taken into
account.

      Superconducting magnets will eventually facilitate the miniaturization of
solenoids because strong superconducting magnets can be fabricated in a small
package without the overheating, even when densely packed.  By definition,
electricity running through superconducting materials produces no heat, which
makes superconducting magnets incredibly energy-efficient.  So far, the "high
temperatures" required to use high-temperature superconductors are on the
order of 300 degrees below zero Fahrenheit, but they are slightly higher than
the temperature at which nitrogen, the principal component of air, liquefies. 
Liquid nitrogen, which is used to keep existing high-temperature superconductors
cold, has been touted as being cheaper than beer, but anything that cold must
be treated with extreme caution in devices developed for the general public. 
Liability issues could be enough to eliminate superconductive solenoids from
consideration for Braille displays.  Also, the known high-temperature supercon-
ductors are brittle and somewhat expensive, so even the discovery of supercon-
ductivity near room temperature, if that is possible, would not guarantee
applicability to Braille cells.  Superconducting materials applications are in the
fundamental research stage and it will be 5 to 10 years before applications are
marketed.

      An entirely different approach to paperless Braille uses electrodes to
indicate the presence of a Braille dot with a tiny electric shock below the
threshold of pain.  This approach has potential for low cost, high speed, and
small size, but experiments have not produced a design acceptable to the end
users.  Until technology is perfected, it cannot be considered a viable technology
for Braille displays.  Problems include reduced reading speed and very wide
variations in skin resistance, both among different people and with sweat and
other factors.

      An alternative approach to high reliability would be to use what are called
"smart" materials.  Smart materials combine sensors and actuators to react to
special situations.  In this case, smart materials might be able to provide high
reliability with imperfect locking mechanisms by verifying that a dot has been
raised or lowered.  If not, an actuator could be triggered several times, allowing
the reliability per trigger to be lower.  Alternatively, the actuator could be
vibrated to free any dirt that caused the failure.  The smart materials approach
is a compromise, attempting to avoid the high cost of piezoelectric technology
which does not need a locking mechanism against the high cost of an extremely
reliable locking mechanism.  The smart materials approach would give a reliability
boost to a mechanism that is already reliable.  It would not be adequate with a
mechanism that has a high failure rate.  This is because if a dot fails to work
after perhaps two or three tries, then the display's electronics would have to
indicate an error.  When a display operates properly, the error message should
not appear except in the case of a catastrophic failure.  The smart materials
approach might also be used to adjust the overall height of the dots on the
display.  The cost of this feature may be prohibitive.  At this time, shape memory
alloys would probably be the best choice for testing along with low-cost
piezoelectric-film sensors.

      Telesensory's Optacon II bears mentioning because the piezoelectric device
provides access to printed text and the technology might be adapted to Braille
in some way in the future.  Based on ceramic piezoelectric technology, it uses 100
vibrating rods (5 columns by 20 rows) to present the image of letters from a
small camera.  Many persons with visual impairments find the Optacon II useful
for reading print and interpreting graphics.  It provides instantaneous results
and offers great flexibility and portability.  Like Braille, learning to use the
Optacon II takes many hours of training and not everyone can become proficient
in its use.  The Optacon II is intended to supplement Braille, not replace it. 
Embossed Braille requires no special reading equipment or equipment costs, and
it was also found to be easier and faster to read than raised letters.  The
Optacon produces raised letters when used to read print.

      It is important to note that the skin's sensitivity to fixed dots differs from
its sensitivity to vibrating dots.  Vibrating dots are not used for Braille because
vibration temporarily reduces the skin's tactile sensitivity and actually reduces
the ability to read tactile information.  This is unfortunate because many
piezoelectrics move at least an order of magnitude more efficiently near their
resonant frequency.  Vibrating dots also generate a buzzing sound, but that is
a solvable secondary issue compared to the human factors problem.

      A detailed discussion of tactile graphics displays is beyond the scope of
this document, but they are closely related to Braille displays.  The big
difference is that they generally use an array of evenly-spaced dots instead of
the standard Braille spacing.  These displays offer the advantage of graphics
capability, but, in general, they cannot produce Braille with standard spacing.

      This document has concentrated on paperless Braille technology.  However
embossed Braille technology is also important.  The primary technology for
embossing Braille is solenoids, and this seems likely to continue for many years. 
The only commercial alternative has been the use of molten plastic containing
magnetic material.  The plastic is magnetically guided onto paper to form Braille
dots.  Until 1990, Howtek sold an embosser, the Pixelmaster, based on this
principle, but it could only raise dots eight thousandths of an inch.  Even twelve
thousandths of an inch is barely tolerable for Braille, and far below standard dot
height. So the Pixelmaster, which could also produce printed text, was considered
appropriate only for tactile graphics and visible print.  Technologically, the
Pixelmaster could have probably produced Braille of normal height, but it was
originally designed for advertising.  The plastic "ink" was originally intended for
producing more brilliant colors displays and not to produce Braille.  Similar
technology was tested in Japan and found to produce the proper height of dots,
although dot shape was a problem.  Plastic dots on paper are much more durable
than conventional embossed paper dots, but it is unclear whether their feel can
be made acceptable to the users.  Dots that are too smooth are harder to read.

      Smith-Kettlewell is in the very early stages of developing a thermal
embossing technology, said to offer the possibility of fast and silent paper
Braille.

      Multiple copies of Braille text can be made with heated plastic sheets that
conform to the shape of the original Braille page.  This process is called vacuum
forming.  In larger quantities, printing press techniques can be used to emboss
Braille, but copying Braille is still expensive.  Even in quantities of a few
hundred, a relatively fast Braille embosser still has advantages over Braille
copying technology, including the feel of paper vs plastic.

      Sheets containing encapsulated ammonia are being used to produce some
tactile graphics, but the special plastic sheets required are too expensive to be
used as a substitute for paper Braille.  They are used with thermal copiers, but
have the feel problems associated with plastic too.

10.0  COST CONSIDERATIONS OF ADVANCED BRAILLE TECHNOLOGY

      As explained in the preceding section, advanced technology Braille must
cost less than 20 to 25 dollars per dot to be cost effective when compared to
existing technology.  A substantial price reduction would be needed to make
multiple-line displays marketable, and a 40-cell by 25-line full-page display would
have to cost less than $2.50 per dot to be in the price range of existing market
forces.

      Piezoelectric displays, the dominant technology at this time, show little hope
of dramatically dropping in price in the near future unless a much less expensive
material for them is found or new manufacturing techniques formulated. 
Thermostat metals, though cheaper, are not energy-efficient, making them a poor
choice for displays.  Highly mechanical approaches tend to be expensive,
unreliable, or both.  Shape memory alloys tend to be somewhat expensive, though
moderately energy-efficient.

      Smith-Kettlewell's electromagnetic technology seems to be able to offer a
significant price breakthrough, but it is limited to one-line displays.

      Soft Braille offers an apparently inexpensive alternative to refreshable
Braille technology, but its costs may be misleading.  Refreshable Braille makes it
feasible to move to different points in a document without too much confusion,
but Soft Braille is best suited to cover-to-cover reading which is not the reading
pattern except for some pleasure reading.  Also a disposable belt model must be
judged on the basis of the cost and inconvenience of replacing the belt.

      On the horizon, magnetostriction seems to offer little hope for a price
breakthrough at this time.  On the other hand, ER fluids seem to have the
potential of producing a low cost Braille display.  However, there are significant
engineering hurdles to be overcome.  Polymer gels offer future hope of low-cost
displays, though they are still in the basic research stage.  New developments
in piezoelectric materials merit further study.  Superconducting solenoids and
silicon micromachines are probably beyond the ten-year scope of this study, and
it is not clear whether electrode-based Braille will ever have the feel demanded
by persons with vision impairments.

      Based on the outlook for affordable full-page Braille displays, two
compromise approaches should be considered.  First, smart materials are a way
of increasing the reliability of existing mechanical locking mechanisms: 
Eliminating the need for locking mechanisms might be better in the long run, but
that may not be technologically feasible without the expense of piezoelectric
displays.  Second, further investigation into the effectiveness of a sliding one-
line display is justified by the lack of compelling evidence that a full-page
Braille display is technologically feasible in the next three to five years. 
However, research and development efforts should be devised to push the
technologies described above beyond the laboratory.

11.0  COST BENEFITS TO PERSONS WITH SENSORY IMPAIRMENTS WITH EARLY
      INCLUSION OF BRAILLE

      Braille displays are presently costly, and therefore many persons with
vision impairments currently use synthesized speech instead.  More affordable
Braille displays would give persons with vision impairments more options between
speech and Braille, and increase literacy among persons with vision impairments. 
Lower-cost Braille displays would allow larger displays to be purchased per
dollar.  Perhaps the biggest short-run cost benefit would be the improved
earning potential that a better Braille display could give blind workers, especially
in the computer programming and office environments.

      In the long run, an affordable full-page Braille display would contribute
to Braille literacy, education of the blind, and access to computers, empowering
persons with visual impairments.

12.0  PRESENT GOVERNMENT INVOLVEMENT IN ADVANCED BRAILLE TECHNOLOGY

      At present, no Government programs are known to be supporting Braille
display development, although virtually all previous development have been
Government-sponsored.

13.0  ADVANCED BRAILLE TECHNOLOGY TIMELINE

      Paperless Braille technology has settled on piezoelectric actuators because
of their reliability and energy efficiency.  To compete with piezoelectrics, any
new technology must provide reliability and energy efficiency at a lower cost per
cell.  As computers become increasingly portable and dependent on battery
power, only the most energy-efficient Braille displays can be used in these
portable units.  As the most energy-efficient proven technology for Braille
displays, piezoelectrics should be reevaluated to determine if new manufacturing
technologies can lower the cost of piezoelectric displays.
      
      Smith-Kettlewell's new electromagnetic technology, already in the process
of becoming a commercial product, may provide significantly less expensive one
line displays.  Expected in 1993 is Blazie Engineering's pneumatic display which
may make full-page displays affordable.

      Recent Soft Braille displays offer a compromise between low cost and limited
performance.  Their price is revolutionary, but their capabilities are extremely
limited for most applications other than reading for pleasure.

      Magnetostriction is not likely to be a major factor in Braille displays, but
clever designs or new materials breakthroughs could overcome the cost barrier. 
ER fluids, with the potential for low cost and high energy efficiency, should be
ready to begin development in 1992.  Polymer gels should be available in
actuators between 1993 and 1994.  If their energy-efficiency proves to be high
enough to make them practical, they should be evaluated for Braille displays. 
However, they are still in the basic research phase in 1991.  Superconducting
solenoids and silicon micromachines are beyond the ten-year time scale being
considered by this study, and electrode-based Braille seems unlikely to be
practical at this time.

      A smart materials approach, with shape memory alloys and piezoelectric film
sensors, merits some consideration, but it offers only a limited chance of success
in competing with piezoelectric displays.  That approach is moderate-risk,
moderate-gain, and probably moderate-cost, since some of the development has
already been done.

      Experimenting with a one-line display that slides up and down a page, that
may be compressed vertically, is a higher-risk approach, but it offers potentially
very high gains if it makes a full-page display unnecessary.  Cost should be
relatively low in this approach, making it especially attractive.

14.0  PROPOSED ROAD MAP FOR INCLUSION OF BRAILLE CAPABILITIES

      The U.S National Aeronautics and Space Administration (NASA) and the
Department of Defense (DOD) fund actuators for specific systems.  Neither is in
the business of developing advanced actuators for Braille.  However, in a
cooperative funding teaming arrangement for research and development on small
lightweight actuators there probably would be a great deal of interest.  For
example, both the DOD and NASA fund research and development efforts for
devices for the handicapped through small business and university innovative
research grant programs.

      The Department of Education, as a first step in Braille cell development,
should explore the possibility of a cooperative effort with both NASA and the DOD
in the area of small actuators for use in full-page Braille cells.  This would lead
to a research and development program for small low power consumption
actuators for use in braille display devices, robotics, and space applications.

      A program would then be established to develop single-cell actuators that
could be used in braille displays.  NASA or the DOD could sponsor the initial
basic research under small business or university educational grants for cells of
one to six elements.  This could be followed up with a program by the
Department of Education under a grant for a Phase I program to develop a
single-line Braille display of 80 characters.  Finally, a grant would be awarded
for an 80 column by 25 line display.

      Key to each phase would be three to four research organizations competing
for the next phase of the program.  For example, NASA might start with four to
five organizations for a one year concept design study at approximately $125,000
per year, for small low power actuators based on advanced technology.  Designs
would be sought for both single actuator and multiple cell designs.  NASA would
then select the three organizations with the most feasible designs for a second
phase to integrate the individual actuators into a prototype Braille cell display
and fabrication effort at $150,000 each for one year.

      The basic research effort would then be followed up with a Department of
Education effort to integrate one or two of the organizations' actuator designs
into low-power, full-page Braille displays over a two to three year program.  The
Department of Education and NASA would jointly fund the efforts and cooperate
in the exploitation of the technology.

      Finally, the Department of Education would fund a production study and
transition the devices to full scale development.

15.0  POTENTIAL PROGRAM SCHEDULE

      Figure 2 presents a potential program schedule for the development of a
braille display unit using advanced technologies.   The Department of Education
could act as the program administrator, with NASA and the DOD providing basic
research and development expertise at critical times throughout the development
effort.  Within five to six years a full-page Braille display could be on the way
to full scale production.
