Introduction

The Smith-Kettlewell RERC focuses on the development of new rehabilitation technology and methods for blind and visually impaired individuals, a diverse group for whom a variety of different approaches can be required in order to solve particular problems. Our research addresses all age groups and the broad range of life functions (e.g., vocational, daily living, communication, and orientation and mobility problems).

The present report summarizes progress over the past 5 year period (1995-2000) on the several priority areas prescribed by NIDRR for this RERC. More details of the projects herein, can be found in the Annual Reports published by the RERC, in the various journal articles and other publications produced over this period, or by contacting the RERC directly.

A. TECHNOLOGY AND METHODS FOR INFANT VISUAL IMPAIRMENT ASSESSMENT

1. New Electrode Technology

The early detection and assessment of visual impairments is vital in order to provide appropriate rehabilitation and prevent further deterioration in vision as infants develop into children. Innovative methods of infant vision assessment pioneered in our laboratories by Dr. Anthony Norcia have proven to be highly successful in this regard.

During the present project period we addressed the development and evaluation of a new class of active electrodes for infant vision screening. This development is especially important for vision screening in the multihandicapped population. The goal was to achieve electrodes which are easy to apply and remove, without the need for electrode paste, to provide a quantum leap in practicality and convenience in addressing the vision problems of this and other target groups. (Currently, the skin preparation, applications of gels, and the attachment of electrodes can be a major difficulty in the case of multihandicapped infants and children for whom minimum preparation and testing time is a necessity.)

In collaboration with Wen H. Ko at Case Western Reserve University, we engaged in development and evaluation of special active electrodes specifically optimized for Visual Evoked Potentials. (Earlier, we had tested multi-gigohm active electrodes previously fabricated by Wen H. Ko & Associates for an EKG application.) The new electrodes are of extremely high input impedance, with a unity gain operational amplifier incorporated into the electrode, and an active shield to optimize signal-to-noise performance. The electrodes will operate with no additional skin interface preparation at all; however, performance is superior when a simple conductive foam is interposed. This, however, can merely be added to the surface of the electrode and does not increase the time or discomfort of electrode attachment.

Through a collaborative SBIR grant (based on early test data we had supplied from the EKG electrodes) we tested prototypes of the new specialized active VEP electrodes on adults. We developed, and had approved by our Institutional Review Board, a protocol for subsequent infant testing.

Our evaluation showed that, as expected, the electrodes will operate with no additional skin interface preparation at all; however, performance is improved when a simple conductive foam is used. However, this does not reduce their usefulness, effectiveness or speed of application since the foam can be simply applied to the surface of the electrode in advance, and no skin application or preparation is needed.

Based on these results, the design was handed over to Wen Ko and associates for packaging refinements such as the incorporation of both pre- and post amplifiers allowing a longer, lighter cable, with the aim of producing a model suited to widespread use.

B. ORIENTATION AND MOBILITY

1. Talking Signs

a. Introduction

The Talking Signs system is designed to help address the problem of orientation and navigation for blind and visually impaired individuals. The system consists of infrared transmitters placed at the locations of signs, and a pocket-sized receiver carried by the blind user. The receiver decodes the message from the transmitter and verbally informs the user of its message. During the present RERC reporting period, we were fortunate to obtain Field Initiated Research funding to explore important additional applications of the Talking Signs technology. Highlights are summarized below.

b. Surface Transit (Bus) Study

We studied the use of Talking Signs in surface transit applications. Eighteen blind participants located bus stops using either the Talking Signs system plus their usual mobility skills or their conventional strategies, but without asking for assistance. They then located and identified a particular bus out of three lined up along a curb using either their normal procedure of asking the driver or using the Talking Signs system. Participants were generally more successful in locating and identifying bus stops using the Talking Signs system than using tactile signs, particularly when the bus stop was identified only by a pole-mounted sign rather than a bus shelter. Participants using dog guides had particular difficulty in the absence of the Talking Signs system. Our temporary installations on bus stops and bus shelters also provided information on beam coverage, power and mounting requirements for these applications.

We also designed and tested innovative devices for announcing the stops to bus passengers. We bench-tested a system designed to pick up infrared transmissions from Talking Signs placed on the bus stops. Our tests were designed to establish tolerances to timing and interruption of transmissions as the bus approached the bus shelter. The concept would allow the user aboard the bus to receive the message at will by activating his receiver.

This work led directly to transfer of the technology to a major manufacturer of onboard bus signs, which intends to incorporate Talking Signs in all its future products. Since this manufacturer currently services the majority of the market, it is hoped that this step will ensure the widespread adoption of our technology on buses. The owner of San Francisco’s bus shelter system, Gannett Outdoor, is intending to install Talking Signs on all its shelters.

c. Talking Signs at Intersections

The Problem:

Intersection crossing points are the places in any journey where the traveler is most vulnerable to danger. At controlled intersections in busy urban areas, the usual cue for determining when it is safe to cross is the sound of traffic beginning to move in the pedestrian's direction of travel. However, anywhere that green arrows allow turns in front of the pedestrian, safety is not assured by this cue. Many other important pieces of safety information cannot be determined by sound, such as the exact width of the street or whether a turning island exists.

Specialized Talking Sign Development:

To supply the blind traveler with added information about the intersection to augment what he obtains using regular orientation and mobility techniques, we developed and deployed specially modified prototype Talking SignsR units at certain signalized intersections in downtown San Francisco. The signs provide two types of information to a blind pedestrian; where he or she is located (eg “Traveling East on the 800 block of Grove Street toward Larkin Street."), and the condition of the traffic signal. The characteristics of an intersection such as turn lane, mid-block crossing, cut-through, island, pedestrian activated 4-way walk signal, free right turning lane, short walk cycle, use of pedestrian activated signal at island for walk signal to cross second half of street, etc could be added to the message or provided through a different receiver channel. The signs also let the traveler know whether he is straying from the designated path while crossing.

Evaluation:

Twenty persons having very little or no vision were asked to cross four complex signalized intersections in the Civic Center area of San Francisco under two conditions: with information provided by Talking Signs, and without information provided by Talking Signs. Participants included cane and guide dog users with a range of different travel skill levels. Some participants also possessed hearing losses. All crossings were made under normal daytime traffic conditions. Detailed data were gathered on ten specific measures in four categories: safety, precision, need for assistance, and knowledge gained about the intersection.

Results:

Mean street crossing performance over all trials with and without Talking Signs was as follows:

Statistical analysis indicated that participants were significantly more successful on eight of the nine measures when using Talking Signs than when not using them.

Conclusions:

Added information from Talking Signs at intersections significantly improved safety, precision, and independence in street crossing, as well as knowledge of intersections. This applied to good, frequent, independent blind travelers, using a long cane or dog guide, including those with hearing loss and persons who considered themselves relatively poor travelers who did not normally travel in unfamiliar areas.

Providing additional information at signalized intersection is one small, but very important piece of the challenge to provide a unified signage technology which allows all travelers to move efficiently, safely and independently. The Talking Signs can provide a standard mechanism by which a visually impaired traveler can go through an intersection to ATM or fare machine; fare machine to bus stop; bus stop to bus; bus to building; building to elevator; elevator to office; office to restroom, etc.

Full details of the evaluation study have now been published in Crandall et al (1999). Discussions are underway with manufacturers of intersection controls to have Talking Signs incorporated into their designs.

d. Symposium on Talking Signs at Intersections

In June 1997 we held a major Consumer Symposium on the applications of Talking Signs at Intersections. Included in the attendees were representatives from consumer organizations, public agencies, San Francisco Public Works Department, the Architectural & Transportation Barriers Compliance Board (ABTCB), researchers, and industry. Industry representatives included high-ranking officials from Mitsubishi Inc, who are arranging for Talking Signs technology to be implemented in Japan.

The purpose of the symposium was to present results of our research studies and to obtain user feedback. It also provided a forum for various public agencies to share their experiences with Talking Signs. In addition, since the meeting was held at the new San Francisco Public Library, it provided an opportunity for visitors to have “hands on” experience with the system both inside the building and in the surrounding area -- including intersections and the BART system.

On the occasion of the Symposium a list was compiled of all existing Talking Signs installations in San Francisco; approximately 500 signs were now been installed in the city at that time – a number which has more than doubled since then.

e. Vibratory Augmentation to the Talking Signs System

To augment selected types of information (particularly in traffic signal and street-crossing studies), an experimental vibratory output was developed. When a subaudible (50Hz) tone is impressed on the audio modulation of the transmitted recording, a low-pass filter/detector in the modified receiver turns on a vibrator (such as that used in "silent pagers"). The user, noting a pronounced vibration of the receiver, knows to pay special attention to that message which is so marked. Pilot testing of this concept indicated its possible usefulness as a future adjunct to Talking Signs.

f. Evaluation of Talking Signs with Developmentally Disabled Users

In collaboration with The Arc (Association of Retarded Citizens) of San Francisco, we performed an evaluation of Talking Signs in a local subway station. This human factors evaluation involved 19 developmentally disabled clients in determining how effective Talking Signs are for this group. We also explored the special adaptations to the technology (location and messaging) that might be needed to better match the system to these persons' needs.

a. Development of Solar-Powered Transmitters

We developed and tested a solar-based Talking Signs system for use in areas where mains power is not available. Sample applications are as varied as identifying bus stops and labeling hiking trails. We modified the transmitter to be much more energy efficient, thereby reducing the size of both photovoltiac panel and storage battery. We also tested the system in an outdoor environment over several of our "worse case" months (October and November were especially dark and stormy). The "charge/discharge" properties of the unit were constantly monitored and the data summarized. Our initial design proved that even on the darkest days, there was sufficient power in the system to provide message output (for up to 20 feet) during the daylight hours.

a. ATM Access

We built a prototype ATM mockup employing Talking Signs and presented it for evaluation by attendees at several conferences (national and international) involving accessibility issues. The mock-up demonstrated enhanced functionality not only for banking applications, but for all types of public transaction terminals. The application of Talking Signs, particularly for finding the terminal, complements the software approach developed by the Trace Center RERC.

a. Technology Transfer

Our active human factors research program, coupled with the ability to enhance the technology based upon the findings of these studies, has led to significant success in technology transfer. During the present 5 year reporting period, Mitsubishi formed a partnership with Smith-Kettlewell's technology transfer agent, Talking Signs, Inc. to produce and market infrared audible pedestrian signals in Japan. The first of these newly designed devices are currently installed in Yokahama. The partnership has since resulted in a new generation of Talking Sign devices and applications. One example is a new, state of the art, highly compact personal receiver. Another is commercialization of a personal transmitter/receiver combination product first developed by William Gerrey of Smith-Kettlewell, intended for use by a blind person staying in a hotel or other unfamiliar environment. In addition, an advanced version of Talking Signs with several output channels has been developed to tap the wider public market for such applications as museums and exhibits; different messages can be tailored to different audiences such as children, adults, blind persons, etc.

Collaboration with other industry partners was also explored. As a result of the research work conducted at the RERC in the spring of 1996, Luminator, one of the country's largest vehicle sign manufacturers, developed a prototype infrared "Vehicle Destination Sign" that allows a blind person to know what bus is coming when it is 100 feet away. We also engaged in a development project with a local company "NextBus" to provide an infrared audible output for the display of bus schedule information at bus stops throughout San Francisco. We continue to work with the City of San Francisco in helping solve other accessibility problems for information kiosks and building directories.

During the past five year period, the Talking Signs concept was also implemented for a time by TeleSensory, Inc, in the form of their “Marco” navigation system. Through our publication of Talking Sign transmission standards, TeleSensory was able to ensure that Marco was compatible with Talking Signs transmission standards, so that during its product life cycle, blind persons carrying a single receiver could use either system.

Talking Signs installations have been spreading rapidly. In San Francisco alone, equipped buildings and facilities include the Dept. of Public Works office, the new Main Library, Bay Area Rapid Transit/MUNI Powell Street Station, 30 new public toilets, McKesson Plaza, the new Courthouse, Yerba Buena Center, San Francisco City Hall, Outdoor Systems bus shelters, Muni key stops project, Rose Resnick Lighthouse, AFB Office, Brooks Hall Convention Center, CalTrain Station, and others, with many other installations in the pipeline.

Elsewhere, examples of installations include National Blind Union (Rome), Texas School for the Blind (Austin), Venice, Italy (rail station), Glasgow, Scotland (sight center), Center for the Visually Impaired (Atlanta), Carroll Center (for the visually impaired, Boston), Lighthouse International, New York, Iowa Braille Center, Mashantucket Pequot Museum.

Examples of currently planned installations in the US include Ronald Reagan National Airport, San Francisco International Airport, San Francisco Municipal Railway, Valley Transit Authority (Santa Clara, Ca), Embarcadero Ferry Landing, San Francisco, San Francisco Child Care Center, Pier 1, San Francisco, Hyde Street Harbor Project, San Francisco, Moscone Convention Center, San Francisco, University of California, Santa Barbara, UC Berkeley, Wright State University, Dayton, OH, Boston Science Museum, Cleveland Ohio Crawford Museum, Seattle Sound Transit, and many others.

2. Remote Sighted Guide

The Remote Sighted Guide is a concept designed to allow a blind user to instantly transmit, to a remote sighted operator, a picture of the scene in front of him. The operator would then answer the blind user's questions about the scene (where am I; which box is the corn flakes; etc.). This revolutionary approach combines the on-call advantages of technology with the intelligence of a sighted guide.

We developed a portable prototype system for studying the feasibility and usefulness of the remote sighted guide concept. The system consists of a miniature black and white CCD camera, an RF transmitter, a hands-free two-way communications system, and a portable 4-1/2" television set. The blind subject carries the camera and points it with verbal assistance as needed from the experimenter who is viewing the portable TV set display. The experimenter then describes to the subject the "information" displayed on the TV set.

The camera module, worn around the subject’s neck, houses a commercially available miniature CCD camera with NTSC output. In its resting position, the camera points straight ahead; it can be easily lifted from the resting position and aimed in any desired direction or held near objects of interest. The lens gives excellent depth of field so that objects between about a foot away and infinity can be in acceptable focus. Several different video transmission technologies were tested, and the final system has a range of approximately 100 feet -- a distance considered desirable and acceptable for concept testing. The viewing monitor can be carried in a special harness around the experimenter’s neck, and incorporates a hood to improve daytime outdoor viewing.

In field testing, the following types of information obtainable by the system were found to be of user interest:

Bus line numbers of bus shelters
Addresses on homes and businesses
Names of shops
Articles in stores
Restaurant menus
Waiting lines in businesses

Street signs proved difficult to read when there was excessive background glare. Street names on curbs, however, were easily read.

We subsequently explored alternative implementation technologies that could be economically extended to large numbers of users over a wide area. A test system was demonstrated using the transmission of images from a remote camera to a website. A wireless modem (Ricochet) system for cellular phones can be used on a miniature portable computer carried by the blind user, and transmit images to a website accessible by any of a number of remote sighted operators. Martin Brenner at CSU assisted us in the implementation of this demonstration system. Other internet-based approaches were investigated by our collaborators on the project, Blindsight Inc (see below). Either of these systems could be practically implemented on a large scale.

After the feasibility and usefulness of the concept was demonstrated, Blindsight Inc, a company specializing in computer vision applications, expressed strong interest in commercializing it for both visually impaired and the broader population of sighted pedestrians. For example, being able to transmit a picture of an assailant to the police, while summoning help, is just one application for the general public that could be commercially attractive. We developed a more comprehensive concept that would combine the Remote Sighted Guide with computer vision techniques (see below). Since the blind user would be carrying a camera, it would be relatively simple to feed its output into object recognition and display reading systems already being developed in our laboratories. In cases where the machine vision cannot interpret the display or scene in front of the camera, the blind user could transmit the image to the remote sighted operator for a more intelligent human interpretation. We therefore folded the Remote Sighted Guide project into our collaborative efforts with Blindsight, and began a search for funding for further development.

3. Maps and Map Reading

a. Talking Objects and Shapes: Preparing Children for Map Reading and Tactile Skills

Introduction:

This project is intended to provide technology to enhance the learning process for children in order to better prepare them for such complex tactile and spatial skills as map reading and tactile graphics interpretation.

It is easy for sighted children to visually identify two- and three-dimensional objects, discover their properties, and observe why, when, and how they are used. Blind children need to learn about things in a more didactic way. For objects, this often starts with tactile exploration, but even hands-on often gives little information compared to visual input. Parent and teachers supplement the tactile examination of objects with verbal information to help make up this deficit. These interactions are most beneficial, but it would be useful to a child to also have access to some verbal information about “things” in a more independent way as well.

The Talking Objects project aims to incorporate speech chips into existing objects to make certain toys and geometric shapes more functional and educational for blind children. Mr. Robert Langstroth, president of Companion Products International, is a consultant on this project, which is coordinated by Dr. Gilden.

Educational toys and games with electronic speech have become popular products for the general consumer market. However, most talking toys use prerecorded speech messages which cannot be altered. Recordable devices would have much greater flexibility and thereby greater educational value. Accordingly we have embarked on a project in conjunction with Companion Products International. of Milford, PA, to develop recordable educational devices for blind children.

Talking Animal Models:

A plastic toy sheep with internal recording and playback circuitry allowing for up to 20 seconds of speech messages was developed. Speech activation is accomplished via a unique, self-compensating (automatically adjusts to ambient light level) photosensitive (“shadow”) switch on the sheep’s surface. By opening the plastic body and pressing the internal record button, the parent, teacher, or therapist can program the toy with a message of up to 20 seconds in length. The child can then activate the speech or other sound message by simply moving a hand over the photocell.

The talking animal figures can have switches at several different locations. This enables a child to touch different parts and then hear the most recent information recorded. Although this might be brief such as “This is the toy sheep’s right front leg”, there is enough room even for a little poem such as “Who am I? I’m not blue. I don’t moo. I don’t peep. I’m a sheep”.

Message content can be updated as the child learns. For example, the messages could progress from preschool-level anatomy to information about domestic animals, cud chewers and ungulates.

Programmable Talking Hands-On Board:

A versatile hands-on talking board was prototyped. It uses an exceedingly flexible scheme to allow an instructor, therapist, or parent to “place” different speech messages on various locations of the board. This device can serve a multitude of functions depending on how it is configured. For example, when supplied with a visual overlay of text or graphics it can serve as an augmentative communication device or a game for those who can see but are non-vocal. When supplied with either a tactile overlay or even no overlay, it can be used by blind children and adults to acquire different types of spatial information. It could demonstrate how cardinal directions are used in maps, provide an audio-tactile display of the basic layout of a room or simple route, etc.

To keep costs down by using off-the-shelf components, Companion Products used a 14” by 9” picture frame for the board=s frame. The frame simultaneously houses the speaker, microphone, power supply, and all audio play and record circuitry, as well as between four and twenty magnetic-sensing switches.

The underside of the board contains a matrix of reed switches which are mounted with Velcro. They can quickly and easily be repositioned to redefine the various areas on the board. For ease of use there is no power (“on/off” switch). The reed switches are activated when a toy (“the traveler”) moves over them. This toy contains a magnet which closes whatever reed switch the traveler is positioned over. Momentary closure of the reed switch activates the speech circuitry.

Recording is accomplished by activating a hidden record switch in the frame. The volume is set automatically. Because this device is powered by 4 penlight cells, rechargeable batteries are unnecessary. In addition, the device powers down automatically after finishing a speech message.

Technology Transfer:

The prototypes were exhibited at several conferences, including the Fifth Annual Assistive Technology Expo at the Westchester County Center in White Plains, New York. They generated much interest on part of both distributors and rehabilitation professionals. Professionals from a residential program for the blind are considering using the sound recording system from the talking animal figures to label things which change frequently such as the food in their cafeteria.

Some attendees realized that the talking programmable board has applications beyond the needs of blind consumers. They see how it can also be used to serve as a basic communication board for those who cannot speak. They also see applications for populations needed more multisensory stimulation to grab their attention and hold their interest. This includes individuals with ADD and cognitive impairments. As a result Brynmar Associates planned to manufacture and distribute this device.

b. Virtual Maps for Blind Persons: "KnowWare"

Introduction:

Raised-line and textured maps have so far been the traditional form of map information available to blind people. These are bulky, expensive, not widely available, and extremely limited in resolution compared to print maps. These, in turn, are bulky, expensive, and limited in resolution compared to software maps. The latter also allow the user to pan and zoom.

In an attempt to provide blind map readers with some of the advantages offered by software maps, Dr. Gilden consulted with Dr. Myron Krueger, owner of Artificial Reality, on the development of a new concept in non-tactile audio-spatial maps for blind consumers. In this system, a user can explore a large surface area such as a table top with his hands, receiving auditory feedback about a “virtual map” that is effectively projected on the surface by a computer. Dr. Gilden named this system “KnowWare”.

Initial Development and Testing:

Supplementary funding for this collaborative effort was obtained through a grant from the National Institutes of Health SBIR program. Dr. Krueger’s laboratory in Connecticut developed the system, and Dr. Gilden’s laboratory undertook its evaluation.

KnowWare’s hardware components are a Silicon Graphics INDY computer with sound capability, a ceiling-mounted camera, and an illuminated surface. The software 1) presents map layout and features, 2) calculates the location of the user’s hand relative to the Avirtual mapA (in effect, an invisible map projected onto any plane over the light table), and 3) plays sounds (speech, music or other) that represent features on the map. As a user moves a hand over the light table, the image of an outstretched finger is transmitted by the camera so the computer can analyze its “geographic” location. The computer presents audio information which informs the user which features are being touched at any moment.

Five blind subjects participated in the initial phase of this project. They were asked to examine virtual maps generated by KnowWare of the United States, of individual states (enlarged when presented singly), and in some cases, of Africa. Africa was presented to those subjects who seemed capable of the challenge of learning about unfamiliar territory.

For tasks requiring locating and tracing boundaries, the five subjects exhibited a wide range of ability to function in KnowWare’s audio-spatial environment. Although all subjects seemed to require more time and take more concentration to examine and identify boundaries on KnowWareJ than on corresponding puzzle shapes of the states, they also received more information in the virtual map. Unlike tangible maps (puzzle maps or raised-line maps), the size, spatial detail, and verbal information of computer-projected maps can be infinitely varied C usually within seconds. During the exploratory research, this enabled a few of the subjects who examined enlarged projections of some individual states to make comments such as “I didn’t know that state had such an irregular coastline”.

Once they examined the map of the United States, all five of the subjects found it easy to quickly locate particular states.

Second Phase of Development and Evaluation:

In the spring of 1998, the RERC sponsored an effort to extend the current KnowWare system and to more rigorously test its performance. The first task was to rehost the original KnowWare software (implemented on a Silicon Graphics INDY workstation) to a personal computer. Secondly, Articifial Reality Corporation added a new feature which permits the operator to use the index fingers of both hands simultaneously to explore the shape of a geographic domain. Each finger receives a distinguishable tone when it is "touching" a map boundary and another when it is inside a map domain. This mode allows the user to quickly measure the width or height of a state or country, to get a sense of its general shape, and to identify narrow features such as peninsulas.
Finally, the investigators performed a rigorous comparison of blind people's ability to use KnowWare with their ability to access the same information from traditional raised-line maps. Fourteen adult braille readers participated. Half were exposed to and tested on KnowWareJ first, and afterward tested on raised-line maps. The other half were first tested on the raised line maps, and afterward exposed to and tested on KnowWareJ. Even though all of the subjects had used tactile materials extensively when they were in school, they were given only 10 minutes of familiarization with KnowWareJ prior to testing. The countries of Africa served as test materials because they are unfamiliar to most Americans. The subjects were given similar tasks on the two technologies. Subjects' overall performance using KnowWare was better than with the raised-line materials at the .03% level of significance.
This result was extremely encouraging. As a result of this work, along with the advent of inexpensive cameras attached to computers, and other emerging developments in hardware, software and Internet technology, it is becoming more likely that this virtual map concept, or some variation of it, will become commercially feasible. For example, it could prove an invaluable tool for accessing Windows, scientific graphics, diagrams, and the WWW. At the end of the current RERC reporting period, additional funding through an NIH SBIR grant was obtained to help pursue these goals.

4. Other Orientation and Mobility Projects

a. Emergency Exit Information

The RERC provided technical resources for the California State Fire Marshal's "Emergency Evacuation Information Task Force for People Who are Blind or Visually Impaired" monthly meetings in Sacramento. We organized an area-wide meeting of blindness specialists (O&M Specialists) which made three recommendations for the state fire code changes. The RERC sponsored a nationwide survey of blindness professionals to review these recommendations.

b. Standards Setting

The RERC has provided information for officials working on important regulatory actions relating to orientation and mobility for blind people. An example was supplying technical information for people involved in the drafting of the Transportation Equity Act of 1998 ("T 21") which reads, in part, "A safety consideration shall include installation, where appropriate, and maintenance of audible signals and audible signs at street crossings." The resulting language does not advocate any particular form of audible signage such as Talking Signs, but should serve to encourage improved access by blind pedestrians to intersections around the country.

Similarly, we have provided significant input to standards setting through connections with ANSII, the ATBCB, and the Transportation Research Board.

C. DESCRIPTIVE VIDEO STUDIES

1. Problem Definition

"Video description" (a narrative accompaniment to video programming, made accessible to blind and visually impaired viewers) presents a unique challenge, different from distribution of captioning for the hearing impaired. Captions (lettering on the screen representing audio dialogue) require so little bandwidth that they ride on the video signal without special handling by broadcasters. In contrast, an audio description track, via technology currently in use, requires special handling by originators of master tapes, centralized network facilities, network affiliates, and a channel on the satellite distribution grid must be devoted to it.

2. Standards for Transmission Format

What is required is a system by which an accompanying description channel be made "transparent" to all network-broadcast components involved. William Gerrey of the Smith Kettlewell RERC had earlier completed a project for OSERS on the technical feasibility of video description, concluding that the SAP (Secondary Audio Programming) channel provided a feasible method for widespread video description to be achieved. During the present project period he played an active role as a member of the Panel of Technical Experts (which included demographic and marketing experts) for an American Foundation for the Blind "Project to Conduct Research on Described Video's Audience, and Methods of Distribution" funded by OSERS. A wrap-up session of contributors on various aspects of video description was held in New York City on September 16-17, 1996. In a teleconference (adding three broadcast professionals, plus a member of Ameritek "New Media" project), a collective consensus was reached.

The Federal Communications Commission (FCC) issued a "Notice of Inquiry" in July 1996, to which individuals in our panel had submitted comments and suggestions. Among these were: "set-aside" bandwidth of upcoming "advanced television" (ATV) standards be dedicated to video description, and audio in the "vertical blanking interval" currently be made available for distribution of video description.

At the time, the FCC took a "wait-and-see" attitude to the submitted recommendations. The Telecommunications Act only required the FCC to report on video description; subsequent action of the Commission at the time, based on a "Notice of Proposed Rule Making" was to enforce guidelines only on "closed captioning" for the hearing impaired. Under these conditions, enforcement of "pass-through" regulations of cable providers, where "secondary audio programming" (SAP) is concerned, seemed unlikely for much of the cable translation equipment, since each of the dozens of channels may require its own SAP generator to deliver SAP to the consumer.

During this period of waiting, both the WGBH/TV Boston engineering staff (with funding from the U.S. Department of Education), and a private multimedia firm, Ameritek, undertook development of digital transmission standards which would accommodate extra audio bandwidth for video description.

No one organization can solve the problems of audio description for blind viewers (and much of the work involves specialized broadcasting expertise). However, we believed our RERC could best continue to play a role in the process of setting television transmission standards by actively making our expertise available as consultants -- including assistance with evaluating quality of proposed description audio channels with input from consumers.

Postscript: On Friday, July 21, 2000 (seven weeks after the present reporting period ended) the FCC announced new rules for video description on television. The new regulations, taking effect in April 2002, require that stations affiliated with the top four networks in the top 25 markets provide a minimum of 50 hours per calendar quarter of video description (about 4 hours per week). Cable and satellite providers must also provide this amount of description for the five largest non-broadcast networks they carry. All emergency news broadcasts must also contain an aural representation of key visual information (as well as captioning for deaf viewers). A crawl at the bottom of the screen containing emergency information must be accompanied by some indication, such as a tone, letting blind people know that there is important information being presented. The Smith-Kettlewell RERC warmly welcomes this new initiative.

3. Associating Description with Film

To solve the problem of incorporating video description in new movies, we proposed, with strong interest and offers of assistance from engineering staff at Dolby Labs, a trial system whereby a low-bandwidth speech channel carrying description could be incorporated onto movie film. Subsequently, an alternative approach has emerged through efforts by other video describers (WGBH and a Los Angeles private describer, Theatre Vision), to synchronize a compact disc (CD) of the description with movies. (Modern movie productions, especially those with stereo and "surround sound" already use a synchronous compact disc to accommodate the desired effects; coupling an added CD drive with signals from the camera is relatively easy.)

As long as the industry pursues this coupling of technologies, we recommend that those working to create accessible movies choose this "add-on" approach. The Smith-Kettlewell RERC continues to monitor actively the developments in the film industry. If an integrated single medium (replacing the synchronous CD approach) emerged, we would attempt to work toward reserving space on it for a description track.

4. Access for Deaf-Blind Persons

In collaboration with WGBH, we initiated a project to extend the concept of video description to deaf-blind viewers. The Descriptive Video Services division of WGBH/TV made available to us diskettes of two shows, "An American Experience" (documentary), and "Mystery" (drama). One diskette, in IBM format, contains all captions associated with the dialogue; the other, in Apple Macintosh format, contains the description of action. Unfortunately, neither type of file is "clean"; they are cluttered with special annotations and were not organized in formats desirable for our experiment.

The "American Experience" documentary was "cleaned up," and the two texts (captions and descriptions) were combined in proper order. Finally, a properly formatted braille version of that show was produced for “viewing” by deaf-blind consumers. The process indicated that unless "flags" mark points in the description text, repeats of dialogue (duplication captions), and other notes to the description narrator make the processing (cleaning) of the files an extremely difficult editing job. Thus, the project demonstrated that the necessary conversion could be done, but due to the editing complexity it was decided not to pursue it on a larger scale at this time.

D. ACCESS TO VISUAL DISPLAYS AND MODERN PRODUCTS

Work in the priority areas of visual displays and product access focuses on developing long-term, intermediate-term, and short-term strategies for access to by blind and visually impaired consumers. The two areas of products and displays are so interrelated that they are treated together in the present report.

1. Long-Term Strategies

We participate in concert with other research, manufacturing and service organizations in national efforts to develop agreed standards for visual display and product accessibility which manufacturers of consumer and other products could incorporate in their manufacturing processes at minimal cost.

One option we explored was the development of a standard for infrared links to and from products with visual displays. We have proposed this solution in the past, based on our experience with Talking Signs technology, and others have advocated its use for persons with both visual and physical impairments. The concept would require the information normally sent to a visual display to be routed through an infrared LED whose transmissions could be monitored by a suitable transceiver with speech and other output. This approach appears to offer the hope of a practical, low-cost solution if the needed standards and protocols can be devised and adopted.

We were founding members of an initiative by the Trace RERC known as the Universal Disability Infrared Link Protocol, bringing together persons interested in all disabilities to develop suitable standards. The group first met at the CSUN Conference on Technology and Disability in March 1995. To gain access to existing proprietary standards, and help steer future standards towards accessibility, Trace became a member of the IRDI, an industry group which includes major electronics and computer manufacturers involved in setting compatibility standards for infrared data transmission for computing devices. We contributed information on existing frequency bands and uses of infrared communication, and on Talking Signs compatibility.

In November 1996 we helped initiate a second initiative to engage in a dialogue on this subject with the Consumer Electronics Division of the Electronic Industries Foundation (EIF). We participated in drafting a letter to the EIF to be signed by a wide array of organizations related to technology for the blind assembled at the Third US/Canada Technology Conference in Baltimore.

As another example, Smith-Kettlewell research has been directly instrumental in providing guidance in formulating recent regulatory actions affecting ATM accessibility. For instance, the "Tech Act" for 1998, section on Communications Equipment, Section 1193.35 Input Controls and Mechanical Functions, Paragraph A, requires product input, control and mechanical function to be locatable, identifiable and operable through at least one mode without requiring the user to see. We also provided input into the ADAAG guidelines controlling ATMs and fare machines (see above).

Through these and other similar activities, the RERC has attempted to contribute to the movement towards universal design and a future in which accessibility features are built into all the nation’s infrastructure. Although no single organization can ensure long term accessibility of products and displays, we believe our supportive role with other organizations in this regard is most important.

2. Intermediate-Term Solutions

a. Computer Vision

We explored the use of new solid state camera, optical character recognition, and image recognition technology for the interpretation of the ubiquitous liquid crystal displays which make access to an increasing number of products difficult for blind users. The recent recruitment to Smith-Kettlewell of Alan Yuille, Ph.D., from the Harvard University machine vision laboratory gave us new capabilities in this area. In collaboration with Blindsight Inc (Mark Notzberg PhD, President), we obtained SBIR funding from the National Institutes of Health to help pursue this project.

In 1998 we completed development and user testing of a prototype system using a miniature notebook computer with a camera attached. The prototype was tested on half a dozen different displays using five outside blind subjects. Results showed that the displays could be read with over 95% accuracy. The prototype was demonstrated at the CSUN and RESNA conferences, and was most favorably received by blind users.

In this initial version, obtaining correct camera alignment necessitated the use of guides attached to the displays. We subsequently developed algorithms for a more advanced version which could be hand held at some distance from the display with a range of permissible misalignment. This more advanced version includes the capability of finding the visual display amid the background clutter.

We envisage two configurations for the commercial version of the device; a low cost version using the clients existing desktop or notebook computer, and a self-contained hand held unit, shaped along the lines of a TV remote or flashlight, containing its own computing power.

b. Remote Sighted Guide

The remote sighted guide concept (in which a blind user transmits a picture of the scene in front of him to a remotely located sighted person who interprets the image for him) can be applied to product and display access as well as problems of orientation and daily living. This project is described in Section B (Orientation and Mobility) above.

c. ATM Access for Blind Persons

Location and effective use of ATMs and other public terminals (vending machines with displays, kiosks, hi-tech public telephones, and ticket and fare machines) is a significant and rapidly increasing problem for those with visual impairments. Access to these terminals includes finding them, and (once they are found) reading the printed information on the screens as well as rapidly identifying the location and functions of the various controls.

During the reporting period we we have developed an improved prototype demonstration ATM access mock-up based upon the use of Talking Signs infrared technology. It enables a blind individual to find the ATM and to identify the location and functions of the various controls.
This approach to ATM accessibility has been demonstrated to many banking officials, blind consumers and rehabilitation professionals, and is designed to complement the infrared interface initiative mentioned in section D1 above and the demonstration systems for accessible information kiosks developed by the Trace RERC.

As a result of our research and dissemination efforts, supported by our consumer colleagues and others, new initiatives in ATM accessibility have recently been adopted, including new rules specifying accessibility enhancements to interactive transaction machines. These rules contain language supportive of "wireless" technologies as an appropriate alternative audible interface.

3. Short-Term Solutions

While pursuing the long and intermediate term solutions outlined above, we recognize the needs of blind employees and consumers for immediate access to displays and instruments. Accordingly, we undertook a number of projects to make specific products and displays accessible in the short term. These activities include the development of specialized adaptations to make displays on existing devices readable; the development of alternative devices with accessible display modalities; and the evaluation of commercial products to assess their accessibility for blind users.

a. Talking Frequency Counter

In both professional electronics and amateur radio, measurement of frequency is a commonly required task, both as a specific problem when the frequency of a transmitter or oscillator needs to be checked, and as a display access problem where a radio has an inaccessible display. Several "talking" frequency counters have been briefly available over the past ten years or so, but these have usually been both expensive and on the market for relatively brief periods. A major reason for this limited availability is that most such units have been built around adapted commercial counters whose displays have been made to "talk." With the increasingly short product life of any particular model of electronic device, and with visual display technology changing rapidly, the techniques used to make such devices accessible become outdated as soon as a particular commercial unit becomes obsolete.

In an effort to solve this problem we undertook the design of a talking frequency counter based on "multi-sourceable" components. The design used the "embedded microcontroller" and "digitized speech" techniques previously developed at Smith-Kettlewell, and a frequency counter based on a small production commercial unit which was being modified for our use (see below.)

The design included frequency measurement and a number of unique features including a means of "finding" a frequency entered from a keypad, and a means of limiting the number of digits actually spoken during each "reading" of frequency. These features would help compensate for the inherent slowness of spoken outputs.

The core frequency counter circuit used in the design was marketed by Radio Adventures Corporation of Franklin P.A. The original design drives an "LCD" display, but the company agreed to re-program its embedded microcontroller to deliver a variety of serial interfaces making it very useful with Smith-Kettlewell's digitized speech system. The "dual processor" system to would make several of the features mentioned faster and more accurate while still allowing for a device constructed of commonly available components.

During the process of prototype design and fabrication, it became clear that more work than first anticipated would be needed for both the adaptation of the commercial device, and development of hardware and software for the external microcontroller. The project was put on hold pending a solution to these problems. Although the "Radio Adventures" component did not become available, the subsequent increased availability of low cost development systems for the PIC processor family with built-in pre-scaling capabilities and true single-chip functionality renewed interest in the feasibility of the project towards the end of the reporting period. We believe that this new technology will be more promising for future development of frequency counters (and other metering equipment) for blind users, and we plan a new project to investigate such a solution.

b. Radio Frequency Power and "VSWR" Meter Evaluation:

We evaluated and tested a commercial radio frequency power and Voltage Standing Wave Ratio (VSWR) meter, the R.F. Applications model P2000CW. This unit uses Morse code as an output medium for measuring the output power from a radio transmitter, as well as the "Standing Wave Ratio" or efficiency of the antenna system. Morse code is used to take static measurements, and a "dynamic" output system allows for making adjustments to antenna tuners or transmitters while "hearing" the changing values. This unit is fairly competitively priced with similar units having visual displays. Assuming that the user is relatively familiar with the basic use of Morse code, it performs well and is quite easy to install and use. The unit is manufactured by a small builder of similar meters with visual displays and of other products for the amateur radio market. It is suitable for both blind amateur radio operators, and professional electronics technicians.

Morse code was selected by the manufacturer since the intended market is amateur radio operators (most of whom know the code), and because this system keeps the component count and cost of the unit lower than would be the case where speech output is to be employed.

The complete technical evaluation was published in the Smith-Kettlewell Technical File.

c. Evaluation of the “Aaria” Palmtop Computer

In 1995 the newly introduced Robotron "Aria" was obtained for evaluation as both a portable accessible computer and as a portable programmable aid to making accessible devices and displays which already have serial ports accessible. This small unit used a braille keyboard for compactness, and contained a custom-written speech synthesizer as well as word processor, calculator, telephone directory, diary, and clock/calendar software. Unlike similar equipment which was already quite popular, Aria used the commonly available DOS operating system and , and of "PCMCIA" memory card data storage. This flexibility and open operating system standard offered the potential for many applications where a portable, accessible computer is required as a normal part of a work environment, and where other equipment already contains serial RS232 ports but requires special software to interpret encoding systems specific to manufacturer and equipment. Examples could include a digital oscilloscope as well as other electronic measuring equipment and such items as scales, chemical pH meters, telephone "Calling Line Identifier" equipment and a similar range of devices.

Mr. Tom Fowle of our RERC staff assisted the manufacturer in evaluating and upgrading the unit. The manufacturer also offered to provide a technical manual to allow us to program Aria for specialized access tasks. Although release of the manual was delayed, preventing us from trying some of the intended applications, it was clear that the system “as is” could be used for basic access to worksite instruments providing RS232 ports.

d. Hand-Held Camera/Display System for Low Vision

We investigated the feasibility of hand-held camera and display systems for general-purpose use by persons with low vision. Our concept was for a hand-held “flashlight”-shaped camera system that could be conveniently held in a variety of positions and orientations to read displays and controls on consumer and workplace equipment as well as general-purpose viewing of difficult-to-see labels and other images in tasks such as shopping. Current “video magnifier” or “CCTV” systems are limited in scope in that they are specifically designed for reading printed material on a page. Even the new devices with head-mounted displays also use head-mounted cameras, leading to substantial bulk and lack of cosmetic acceptability of the combined device, as well as limited mobility of the camera itself.

We constructed a simple prototype system using a very compact, low-cost, commercially available CCTV camera (originally designed for concealed surveillance) which can be conveniently hand held, and a small CRT display. The concept was discussed with a manufacturer of low cost CCTV systems, who developed a prototype device performing similar functions. This version has a pistol type grip for the camera, with autofocus and a trigger controlling the zoom ratio. Future advances in camera and (particularly) display technology will inevitably make commercialization more practical.

e. Home Entertainment Equipment Access

Video-Cassette Recorders (VCRs) and players, hi-fi stereo equipment, and other complex consumer products and appliances have traditionally been difficult to operate even for people with optimal eyesight. Aside from the complexity of the controls, a major problem is the low contrast and low luminance used on the control panels (often using black switches on a black background), due to an emphasis on elegance rather than visibility. Since devices such as VCRs and stereos are often used in low lighting levels, even an individual with 20/20 vision under normal, brightly lit conditions becomes effectively impaired in dealing with the controls.

To complicate things further, some of the later model appliances (particularly microwave ovens) have no physical switches, the controls being composed of a membrane switch panel. In general, there is no standardization of location of switches, as is the custom in the automobile radio where the volume control and on-off switch is always on the left side.

Pending the advent of more universal solutions (which we are also pursuing C see Section D.1above), we have begun work on a project to provide low-cost, high-contrast markings for the switches and controls on these consumer products. We explored the use of die-cut stick-on labels and overlays, which might also enhance the tactile feel of the switches. Initial tests of this approach indicate that it can greatly enhance ease of use for all users C not just those normally considered visually impaired.

We conducted an informal investigation on the visibility of different colored labels designed to enhance contrast. After testing numerous colors and materials using eight subjects, we found that the color of “day glo yellow green” or fluorescent yellow green was preferred by most. The secondary light radiation of the fluorescent material gives the added advantage, and boosts the apparent contrast beyond that available from labels of conventionally printed colors. This color appears to be appropriate for both black panels and the flat metallic silver panels which together comprise the majority of VCR and stereo front panels.

We were able to obtain this color of material in an adhesive vinyl sheet form. To adapt this material to the situation we cut out tactile shapes for the given controls, adopting standard symbol shapes for “play," “stop," “fast forward," “rewind," etc. common to many VCR manufacturers. We conducted experimentation with different shapes and sizes of labels in order to settle on a set which will suit the maximum number of applications. We determined that by using two different sizes of each chosen shape, the great majority of controls on the various appliances we have used for testing can be accommodated.

We determined that any of a number of adhesive label manufacturers could produce sheets of these stick-on labels, given a design submitted by us. The finished sheets could be sold through low vision aid catalogs as kits that visually impaired consumers or family members could apply themselves.

f. An Accessible Appliance Timer

Until such time as appliance manufacturers make products "talk" their display information, or it becomes convenient to attach talking or braille readout devices to them, we have sought interim solutions to the problem of appliance control. For example, blind tape-recording enthusiasts and users of other home appliances have long awaited an accurate device for turning equipment on and off at predetermined times. Since talking clocks have become so readily available, and are accessible to blind users in terms of alarm-setting controls and feedback, and many blind consumers have more than one of them, we decided to explore the feasibility of harnessing these low-cost devices to program other appliances. We designed a sound-triggered relay circuit actuated by commercially available talking clocks. If the latter are not already available to the user, two clocks and the sound-triggered relay circuit could cost under $100.

The finished device consists of a relay module to which any two talking clocks can be held in close proximity by any convenient means (e.g., rubber bands). The alarm sound from one clock turns the power on, and the alarm of the second clock turns the power off. No modification to the clocks is necessary; audio to operate the system is captured by an inductive pickup in the relay module. This does mean, however, that the clocks used must have magnetic loudspeakers, as most do.

We are aware that this type of system is not an optimal long-term solution; however, it offers a practical and economical solution in the interim. Since many of our blind constituents are builders of electronic equipment (through the Smith-Kettlewell Technical File), it is likely that a significant number of blind people might benefit from this design. We will also make the device available upon request through our privately funded Rehabilitation Engineering Service.

g. Hearing-Aid Battery Tester for Deaf Blind Persons

Commercially available battery testers are accessible neither to blind nor deaf-blind users. Smith-Kettlewell had earlier designed a battery tester for a high school student at the Arizona State School for the Deaf and Blind. In this design, a commercial "button-battery tester" was modified with an attachment which gave the user vibratory indications as to the condition of zinc-air cells used in modern hearing aids. Unfortunately, the Mallory tester upon which that early device was based is no longer available, being supplanted by a much more complicated tester of general application (also meaning that it has unwieldy probes which require holding three items at once, thus interfering with attending to the vibrator).

A simpler implementation was therefore designed and built from scratch. On this unit, a brass plate, on which the cell under test is placed, serves as one contact; a flat-head screw attached to a wire, serving as the other contact placed atop the cell, is the other terminal. (At this time, hearing-aid cells exist in a wide variety of diameters and thicknesses, so a flexible system to accommodate any shape or size is needed.) One disadvantage in this new arrangement is that no visual meter or display is available, something that might be desirable in training use of the device to clients; however overall operation is greatly simplified.

A search for suitable vibratory transducers revealed an electromechanical buzzer ( a magnet on a cantilever which produced vibrations sufficient to feel) is no longer available. "Sounders" of modern availability are Piezo-electric units that generate a high pitch, and thus no vibration can be felt. Therefore, a new vibratory output must be investigated. (The device referred to here contains an old buzzer that was salvaged from an unused instrument.)

Although slightly more expensive, it is anticipated that a modified electromagnetic loudspeaker (with an added plastic dome, such as a section of ping-pong ball) will provide an adequate tactile output. Even better would be an electromechanical vibratory device which imparts enough energy to the case of the instrument that the user would not have to compromise holding the battery assembly while reaching for a vibrating surface.

The output of the unit is ingenious; the cell voltage is compared with a triangle wave whose level is biased so that a "dead" cell is just at the lower peaks of the wave. A cell whose voltage compares favorably with the absolute peaks of the wave will cause the tester to make very short indications, or none at all. A cell which has been used somewhat will cause the tester to emit longer "buzzes" -- perhaps of a 50% duty cycle for a cell with some life left. A cell that should be replaced causes the tester to emit long buzzes, or even a steady signal.

Pending identification of suitable vibratory transducers for future versions, the prototype was sent to the Arizona student for evaluation. He reports frequent and successful use of it. A second prototype was fabricated for evaluation by another consumer, T.T. of San Diego, who gave us the following review:

"The SK vibratory battery tester is working well and has seen plenty of use. On a scale of usefulness, 1 being useless and 10 being an absolute necessity, I would rate it an 8. Up to now, I have tested batteries by whether the hearing aids are working at all. This tester gives me an idea as to how the batteries are doing; this allows you to prepare. I do appreciate it; it has come in very handy."

h. Cassette Players for Talking Books: A Rejuvenator/Evaluator for Talking-Book Batteries

(1) Problem and Solution Approach

The most commonly failed component of cassette players issued by the Library of Congress for the "Talking Book Program" is the Nicad battery pack. Despite disadvantages of such a special battery system, the latest equipment (even a record-playback machine made by the American Printing House for the Blind) incorporates it, so for the time being it is here to stay. From time to time, our RERC has been asked by state services to develop a device with which batteries of returned machines may be quickly checked. For example, can a battery be brought up to its full terminal voltage, and if so, will it hold a charge? If there is a shorted cell in the pack, can the short be cleared?

Consequently, we undertook a project to enable these functions to be performed in a manner accessible to a blind user C either at home or as a staff member of an agency which services returned batteries and talking book players. The resulting device, the latest of our designs of a battery-assessment instrument, allows quick charge and discharge characteristics to be "observed," and an attempt to "clear" shorts in the string of cells to be made by a blind operator.

(2) System Requirements

The battery pack, a sealed unit containing six "sub-C" cells, is rated as having a 1.2 amp-hour capacity. Its terminal voltage is typically 7.2 to 7.5 volts. To make optimum use of it, it should be charged at 120mA for 14 to 16 hours, then left to run its device until its voltage drops below 6 volts or so, then charged again.

Unfortunately, the cassette machines using these batteries do not "override" the battery when they are plugged in; in fact, they need the battery as part of the power supply filter system. They are often left plugged in, thus charging their batteries until chemical reactions within the cells damage the capacity. Furthermore, infrequent use of a machine on battery power causes a type of "memory effect" which reduces battery capacity. Worst of all is when stronger cells in a string last long enough to "reverse charge" weaker ones; shorted cells in the series will be the result. Finally, batteries left on the shelf often develop shorted cells; thus, even replacement ones can be at fault.

In testing and salvaging these batteries, it makes sense to perform the following operations:

When charging, does the battery reach a level comparable to its terminal voltage (usually about 20% higher than that on cells which are not leaky)?

When discharged at the same rate as the charging current, is the time taken to drop below an unacceptable voltage somewhere between one-half to two-thirds of the charging time?

If the rated voltage cannot be reached, does a mild high-current shock to the battery clear shorts in cells so as to restore them back to life?

These are features we designed into the test set described below.

(3) System Design

A "quick-installation" battery holder for these special units was constructed on the top panel. Also on the top panel are the following controls:

1. A 3-position rotary switch selects: (1) a normal discharge load of 10% amp-hour capacity, which also allows a test of battery voltage under a light load; (2) a normal charge rate of 120mA (for overnight charging); and (3) a quick-charge/discharge assessment (at approximately 1 amp).

2. Associated with position 3 of the rotary switch listed above, a momentary switch (pushbutton or "momentary toggle") connects the high-current charging circuit; releasing this switch puts the battery under a heavy load of similar current drain.

3. A high-current toggle switch connects a large electrolytic capacitor across the battery with the intention of "burning out" minor shorts within cells of the pack.

4. A small toggle switch allows the audible indicator to be turned off -- especially desirable when slowly charging a battery.

The heart of the metering circuit is "The Smith-Kettlewell Auditory Battery Tester," published in Vol. 11, No. 1 of the Smith-Kettlewell Technical File (Winter 1990). In the range of interest, the instrument emits "tone bursts" whose length increases until the battery is adjudged to be dead, whereupon the bursts blend together into a steady tone. Put another way, the shorter the "pips" (bursts of the tone), the better the battery. An added feature to the output of this instrument is that the pitch of the tone drops as the battery voltage declines; thus, a good qualitative indication of the discharge rate can be had.

(4) System Operation

In operation, the user might start with the instrument in the simple battery-test position -- with the rotary switch in position 1 and with the audible indicator turned on. (The audible indicator emits a low buzz, in the test position, for either no battery or a dead one, just to let you know it is on.) Placing the battery in the holder, the response of the auditory output is then noted; if the pitch rises and the tone is interrupted to produce "pips," the battery is at least partially charged and all cells are working.

If in the above test the buzz does not change in pitch, or if it does but is steady and not interrupted, an attempt must be made to put a charge on the battery. Switching to position 2, the normal charge position, a favorable indication would be to hear the pitch rise and interruptions in the tone begin; this would mean that all the cells are probably unshorted, and that all the battery needs is to be charged.

A well-used battery may seem like it is charging, but it may not hold a charge when put into service. For this test, the rotary switch should be advanced to position 3. With the auditory indicator still left on, hold the momentary switch in the charging position. (The indicator may fall silent while charging in this high-current mode, since the battery voltage may exceed its range of indication.) After some time, perhaps 30 seconds, release the momentary switch and listen to the indicator. In this mode of heavy discharge, the bursts of the indicator tone will grow longer until they blend together, and the pitch will drop. The battery is still usable if it takes longer than 15 seconds for the "pips" to blend together and the pitch descends rapidly. It is best to keep some charge in the battery if it is to be set aside for later use; you may use the quick-charge feature to bring it up to a "good" indication, then quickly remove it from the instrument so as not to discharge it through the load.

If a 30-second quick charge does not bring the battery up to the point where the tone is interrupted, imparting a shock to it is the last opportunity to salvage it. With the rotary switch in position 2 (normal charge), flipping the high-current toggle switch so as to connect the capacitor to the battery may clear shorts in the package and render the battery useful again.

(5) Evaluation and Dissemination

An early precursor of this device, built for the Nevada rehabilitation system some years ago, proved useful. The latest device has proven capable of accomplishing all the above functions in a manner fully accessible to a blind operator. Its design will be documented for release to the National Library Service as well as being published in our Smith-Kettlewell Technical File.

i. Accessibility of Electronic Music and Studio Equipment

Mr. Jay Williams, formerly of our staff and subsequently a consultant to the RERC, completed a comprehensive primer of accessibility to electronic music. The primer was based on work he accomplished while at Smith-Kettlewell, updated during subsequent professional experience.

Blind musicians and composers stand to lose ground in accessing computerized music, both in performance and in creating (composing) music. With RERC-funded support, a project was conducted to try and prevent and overcome these barriers. "Access to Computerized Music Production by Blind and Vision-Impaired Musicians" was compiled by Jay Williams, wherein hardware, "sequencer programs," and other applications software are evaluated for their accessibility, and advice is given on how to access them effectively. This document has now been published on the Smith-Kettlewell website, and is also available on diskette from the RERC.

Mr. Williams’ expertise in computer-controlled music and editing equipment has also been most helpful in related areas involving access to jobsite equipment. For example, a radio personality wishing to produce documentaries came to us asking if we could adapt a stand-alone "editing station" so that precision time within a program -- shown on a visual display -- could be made to talk. Mr. Williams recommended that, for a competitive price, a computer outfitted with an audio-editing program could serve that desired purpose with off-the-shelf equipment, thus saving thousands of dollars of engineering time.

j. Vibratory Volume-Level Indicator for Live Programming

Problem Statement

Available devices for monitoring program level for broadcasting or quality recording provide the blind audio engineer with auditory signals which alert him/her to the instances where sound levels exceed 0dB (or -4dB, often specified for master recordings intended for transfer to CD's). Where the engineer is not isolated from the microphones, these audible indications -- when made prominent enough to be noticeable -- may be picked up by the microphones, even if the engineer is listening for them over headphones. Furthermore, the audible level indications interfere with adjudicating other aspects of programming -- the "mix" (mixture of sources), acoustic quality, etc. Current visual instruments typically present sound-level information via flashing lamps, and present the sighted audio engineer with no such concerns or conflicts.

Solution

An obvious alternative to lamps or audible signals is a tactile volume level indicator. In 1978, we designed such an instrument using the Holmlund/Alden Vibrotactor, a device with a vibrating plunger which we used in our TVSS and other vibrotactile devices. This 1978 design has two disadvantages. First, the vibrotactile stimulators are no longer available. Second, those tactors were designed to stimulate the skin directly; ideally, the user's finger would be positioned over a hole through which the plunger could tap or vibrate against it. This is not the most convenient mode for an audio engineer whose hands may be occupied with other tasks.

As an alternative, we are exploring the feasibility of developing a tactile volume level display using pager vibrators from two manufacturers. While their response is not as immediate as the Holmlund/Alden tactor (they use motors spinning eccentric weights), the stimulus presented by them is sufficient to be felt with the unit tucked into a waistband or strapped to the wrist. Also, they are readily available.

Evaluation

Before finished prototypes were distributed to three blind recordists involved in doing live recordings, preliminary in-house demonstrations comparing this system to visual instruments were made with volunteer experts in sound recording -- both sighted and blind.

k. Vibratory Vacuum-Cleaner Attachment for Deaf-Blind Users

Problem Statement

When using a vacuum cleaner, it is not uncommon to pick up something large enough to obstruct the air passage, thus rendering the vacuum cleaner impotent until the obstruction is removed. Many people use the pitch of the motor as an indication that this has occurred. Obstruction of air flow removes load from the motor, causing it to run faster and generate a high-pitched sound. Deaf and deaf-blind individuals and others who cannot hear or are unaware of this motor speed variation have no indication that obstruction of the air passage has occurred.

Solution

We designed and tested a prototype attachment which senses the degree of vacuum (or pressure at the output in the case of "upright" cleaners). A significant change in air flow causes this attachment to vibrate sufficiently to be felt at the cleaner's handle. The availability of low-cost pressure/vacuum sensors and pager vibrators may lead to a viable product not only for deaf and deaf-blind persons, but the broader market as well.

l. Tactile-Output Light Probe

Problem:
Light probes are perhaps the most common tools used by blind employees to read a variety of displays, especially those such as telephone consoles incorporating multiple light indicators or LEDs. Auditory output in the form of tones of varying pitch and loudness (corresponding to received or reflected light intensity) has traditionally been used; however this can be obtrusive and is not helpful to those with hearing impairments. Accordingly, there has long been a need for a satisfactory light probe with a vibratory output.

The "tactor" (output transducer to be felt by the user) has heretofore been a specially made component, or an adapted earphone/loudspeaker unit. Any such tactor, whether adapted or fabricated from scratch, was expensive to produce, sometimes quite large in size (in order to deliver sufficient power to the skin). Additionally, it has long-since been known that the haptic sense cannot discern variations in frequency as well as the auditory modality can, and that the bandwidth of frequencies that can be felt is limited to a few hundred Hertz.

Some prototype light probes using those technologies required both hands to operate them; the large transducer ("tactor") was often in a separate box with the battery and electronics. Finally, adapted loudspeakers which delivered sufficient power to be felt were also loud auditorily, negating the privacy and unobtrusiveness advantages of tactile output.

These combined problems have so far precluded the availability of commercially viable vibratory light probes.

Potential for New Designs Using "Silent Pager" Vibrators:
With the advent of "silent pagers", vibrator motors -- small DC motors which swing an eccentric mass to creat strong vibrations -- have become readily available at low cost (under $12 in small quantities). These are comparatively efficient, drawing less than 150 milliwatts at full speed. They come in two basic case styles, cylindrical and "pancake"; both types are quite small. Although their available range of frequencies is limited to about 100 Hertz at top speed, they serve to augment dynamic information by producing the effect of combining output power with vibration rate; a 20% change in drive can easily be felt by the user (a subtlety not apparent with purely pulsating tactors whose output power is constant over a wider range of frequencies).

Another advantage of this output device is that it is relatively quiet. For this reason, hearing people using light probes in social situations may prefer a "silent" vibrating one.

We developed two prototypes for trial, one for each style of motor. Since the motors themselves are equivalent to current-controlled oscillators in audible light probes, a power MOS/FET and two resistors constitute the bulk of the circuit required in these prototypes. When operating, both probes seem equivalent in presenting the dynamic information of light level.

Remaining Problems and Solutions
It was noted in our evaluation trials that there is a failure of the output vibrator to start if a low level of light is encountered which then causes the armatures of the motors to rest in a "dead spot" of their rotation. A slight tap on the side of the probe usually starts the motor, but if the user is unaware that no light is present, the dormant probe is always viewed with suspicion.

This failure occurs less frequently if the plane in which the mass is swinging is horizontal; thus, strategic orientation of the vibrator in the probe can help. However, depending on the task, a light probe may be called upon to operate in any orientation -- horizontally in an elevator while checking as to when your floor button goes dark, straight up if you are checking a ceiling light, at an incline if you are checking to see if a light down the hall has been left on, etc.

We are exploring several potential solutions to this problem, including a more sophisticated drive system -- using a current-controlled chopper to deliver high-current pulses of varying pulse width, use of two vibrators, oriented on different axes, and a continuing search for improved vibrators which are not subject to this difficulty.

m. Adapting the Inventronics "AccuTuner" Display for Use by Blind Piano Tuners

Background:
Computer-assisted piano tuning is at a technological stage that it is being seen as equivalent to "aural tuning". With computer assistance, tuning is done faster, and it is generally accepted that the stress load on the piano tuner using this equipment is greatly reduced. Of particular importance is that instructors of tuning, and adjudicating of newcomers seeking certification from the Piano Technicians' Guild, both rely on a particular instrument, the "Sanderson 'AccuTuner'" by Inventronics, is the approved computerized instrument. Thus, any blind examiner/trainer must have access to the AccuTuner if he/she is to be employed by the Guild.

The RERC at Smith-Kettlewell was approached by the Piano Technicians Division of the National Federation of the Blind to assist in adapting the Sanderson AccuTuner. Dr. Albert Sanderson, President of Inventronics, kindly agreed to provide the technical information required to connect readout equipment to his device.

Description:
The AccuTuner has three readouts. Two are digital displays driven by serial-in parallel-out shift registers; Dr. Sanderson offered to provide us with the serial codes for the various messages, and no special difficulty is anticipated in making these messages "talk" appropriately. The remaining readout is a circular pattern of LED's which appears to revolve in accordance with the pitch of the piano string being tuned. As the string's pitch approaches that of a frequency standard in the AccuTuner, the motion of the pattern slows down, coming to a stop when the string is "in tune". Dr. Sanderson agreed to provide us with connections to two bipolar triangle-wave signals (90 degrees different in phase) which are used to derive the pattern.

We were advised that it is important that, even though fine pitch comparison such as that used in aural tuning, is not being used by the technician with this instrument, it is best that we avoid interfering with listening to the piano while tuning.

Three approaches to the dynamic display were proposed for consideration:

(1) A vibratory output to the dynamic display was tried, but the rapidity with which the display "rotates" for even moderate out-of-tune conditions exceeded the responsiveness of vibrating "tactors". For the time being, a tactile output scheme has been put on hold.

(2) We proposed an audible presentation of spatial character; in this concept, a hat brim containing small loudspeakers emitting pink noise or other sounds -- not to be confused with pitch -- would simulate the array of LED's on the AccuTuner. A survay of members of the National Association of Blind Piano Tuners, a division of the National Federation of the Blind, indicated that this would only be acceptable as a last resort.

(3) The approach which has received popular aclaim is to use eight sequentially filtered bursts of white noise. Eight band-pass filters acting on "white noise" produce distinctly different sounds which represent the eight lamps of the rotating display LED's. A Q of 10 results in sounds which are distinctly different in "pitch", yet these sounds are broad enough in spectrum so as not to be confused with tones from the piano's strings.

The resultant eight distinct filtered noise bursts ascend in "pitch" to indicate a clockwise rotation of the visual pattern; a descending series of "pitch" would indicate counterclockwise motion of the visual display. A ticking sound in accordance with flashing of the sharp indicator was added.

The ultimate criterion is to demonstrate whether or not a user can, with equivalent ease, determine when the dynamic display is frozen. Like with the visual dynamic display, cycling of the white-noise bursts can, at fast rates, become blured and indistinguishable as to the direction, but the ticking of the audible sharp indications seems as effective as the visual flashing lamp.

Trials, as yet, have been simulations using a switched-capacitor filter operated by a variable-rate sequencer circuit. This approach has received enthusiastic responses from both members of the National Association of Blind Piano Tuners, and the technical staff of Inventronics, who are piano tuners as well as makers of the Accutuner device.

Current production of the Sanderson Accutuner is undergoing revision to produce a model with added features. We have been assured cooperation, from the factory, that a suitably modified unit will be available for adaptation.

n. Accessibility Evaluation of Modern Ham Radio Equipment

Introduction

Amateur radio is a major recreational occupation for blind individuals, and in addition to its hobby value it is an important adjunct to the nation=s emergency response system. Modern radio equipment is becoming more and more difficult for blind persons to operate due to the presence of LCD information displays and of controls which toggle between several different functions (with the function usually indicated on the display). Also, some transceivers are designed to operate under computer control using the Windows operating system. During the present reporting period, we conducted an in depth evaluation of the access options for this type of state of the art equipment. The evaluation was performed by Thomas Fowle, himself a blind amateur radio operator.

The Equipment
The Kachina Communications 505DSP Amateur radio transceiver allows for operation on the "HF" amateur radio bands between 1.8MHz and 30MHz. This transceiver, unlike more usual radio equipment which is operated with knobs and buttons, is connected to a personal computer. It is controlled entirely through software running under the "Windows" operating system. Extensive use is made of the "Point and Click" interface paradigm involving such essentially graphical concepts as "Slide Bars" "Soft buttons" and "Virtual Meters." As an example of equipment designed to be operated entirely through the windows interface, this radio Transceiver was tested with two windows screen reading programs, "ASAW" from MicroTalk, and "OSW" Outspoken for Windows from Alva Access Group. This testing was performed using both speech and braille screen reading systems.

Conclusions
The evaluation showed that through the use of the access systems and special operating strategies, a blind person could access the equipment with some limitations. The results were of interest in terest not only for radio equipment but other types of modern equipment which rely on the graphical type of user interface. Our full evaluation report was shared with Kachina communications with a view to adding features which would help make the system and others like it more universally accessible.

Acknowledgements:
We wish to thank several individuals who made this project possible. Mr. Cameron L. Earnshaw, Vice President of Kachina Communications was most helpful in providing this equipment on a 45 day loan as well as documentation not already available on the Kachina web site. Mr. Roberto Gonzalez of Alva Access Group technical support was also most helpful in allowing me to use the beta test of OSW Ensemble V2.0 and in providing assistance with this program. We also wish to thank Susan Fowle NY6D for her invaluable assistance in both learning and operating the Kachina software and editing the evaluation report.

A. TECHNOLOGY FOR DEAF-BLIND PERSONS

1. Dexter, the Fingerspelling Hand

The Dexter project, on which Smith-Kettlewell has been collaborating with Upstart Robots, is aimed at developing a robotic hand which can perform the finger spelling code used by many deaf blind persons. During the present reporting period, we collaborated with a private company, Upstart Robots, with supplementary NIH SBIR funding, to further refine the system.

Under the Upstart Robots collaboration, a new mechanical hand design was developed which was capable of forming even the more complex letters of the code without compromise. Another notable accomplishment was the realization of telecommunications capability. Telecommunication programming was completed by Thomas Fowle of our staff, to the point where it could be demonstrated to telecommunication companies in conjunction with the prototype hand.

A new algorithm for generating the transitions between letters was invented, with the advantage of not needing any "exception" coding for the more difficult letter-to-letter pairs. The software has been rigorously commented and documented. Electronic problems were initially encountered with the I2C bus, but were successfully overcome.

During development, it was found to be desirable to provide feedback on the flexion angle of each individual finger joint, in order to optimize performance in forming the letters. No adequate bend sensor could be found, so microswitches were introduced to sense when the tendon lines become taut. This innovation proved to be a feasible solution for joint angle feedback. Problems of spring breakage due to fatigue in the knuckles were overcome through development of a silicone rubber casting. A test joint made with a silicone rubber tube lasted for one million joint operations with few signs of fatigue.

Another innovation successfully developed and tested was a tether to separate the hand from the drive box and provide users with a more mobile, comfortable-to-use hand. This was accomplished without degrading hand performance.


2. The TeleLite: A Next Generation TDD for Deaf-Blind Persons

Using the telephone is still difficult for many deaf-blind people. At the suggestion of Blazie Engineering, Smith-Kettlewell staff embarked on a new project for design of a successor to the earlier TeleBraille, also developed by Smith Kettlewell, thus assuring that deaf-blind persons will continue to have access to a commercial braille modem.

With the experience gained by Mr. Fowle and other RERC staff in the Dexter and TeleBraille projects, several engineering innovations were developed which should make a small, relatively inexpensive "TDD" modem practical for use with the BrailleLite, a palmtop braille output computer made by Blazie. The resulting device, known as the TeleLite, is being designed to be readily usable with other computers accessible to deaf-blind persons, and for others who may require "TDD" communications in a relatively inexpensive and portable form.

The Telelite will allow any computer, but particularly the BrailleLite, to communicate using "TDD" standards on the telephone line. It will use a modified set of the standard "Hayes" modem command set so that it will be usable with any computer having a serial port and a "Terminal mode." TeleLite will connect to the BrailleLite through its standard serial port. Ascii characters will be converted to and from the TDD "Baudot" code and appropriate tones will be generated to place this information on the telephone line in the "TDD" standard format. Connection to telephone lines will be via standard RJ11 connectors or through a commercial acoustic coupler for pay phones and other situations where jacks are not available.

The new system will also incorporate a “Face to Face” mode to allow a deaf-blind person using this system to communicate directly with a sighted person or work with a sighted interpreter in a conference or meeting. The interpreter will use an inexpensive "print" TDD connected to the TeleLite to communicate with the deaf blind person. It will be possible for TeleLites to be "chained" so that several deaf blind persons may take part in a "round-table" conversation without the use of telephone lines or other equipment. This will be invaluable for training and conference situations.

The system will use internal rechargeable batteries incorporating the same advanced battery control and monitoring system as used in the BrailleLite.

3. A Vibro-Tactile, Multi-Alarm, Wearable Medication Reminder for Deaf-Blind Individuals

a. The Problem

The Helen Keller National Center for Deaf-blind Youths and Adults confirmed our hypothesis that on the whole, the deaf-blind population tends to take more different medications than the general population. A study conducted earlier by Dr. Gilden revealed that the need to be reminded to take medications is sometimes the single block to independence. There are some family members who are "tethered" to their deaf-blind loved one just because they need to be available to remind them many times throughout the day that it is yet again medication time.

There are various types of reminders on the commercial market, and some even target the medication-taking issue. These devices, however, are not usable by deaf-blind persons for two reasons: (1) they require vision to set the alarm times, and (2) they require hearing to detect the alarms.

b. Solution Approach

The Smith-Kettlewell Eye Research Institute RERC had the good fortune to be the chosen sabbatical site of Kazushige Magatani, Ph.D, Associate Professor of Electrical Engineering at Tokai University in Japan. He has joined our RERC for six months (with funding from his university) specifically to work with Dr. Gilden toward developing a medication reminder for deaf-blind persons.

Since deaf-blind people are generally familiar with analog clocks with raised dot markings, we chose the device to be a variation of a "braille" wristwatch. These analog watches require only touch both to set and read the clock time. Ours would only require touch to set and read the alarm times as well. The wristwatch style will make the reminder easy to carry as well as allow it to remain in contact with the skin for vibration detection. It will also enable the user to know the current time and compare it to the alarm (medication reminder) times. The analog clock will be motor driven.

In addition to the analog clock (the interface for the deaf-blind user), the reminder will also includes a digital clock to work with a microprocessor to control the alarm system. The times of the two clocks will be synchronized.

c. Design Details

The digital clock uses an 8 bit single-chip microprocessor which includes two 8 bit parallel interfaces, two asynchronous serial interfaces, and 3 channel timer interfaces. This microprocessor also controls the alarm settings, the synchronization between the analog and digital clocks, an RS232C interface, and alarm vibration. To synchronize the analog and a digital clocks, the position of the hands are detected by using 2 small magnets affixed to the hands and two rings of 12 Hall sensors are set around the dial. The outer ring of Hall sensors will be positioned to line up with the magnet on the minutes hand, and the inner ring of Hall sensors is positioned to line up with the magnet on the hour hand. The microprocessor adjusts the digital clock in accordance with which Hall sensors detect the magnets on the clock hands. In this synchronized mechanism, the maximum error of synchronization is 5 minutes.

To set the alarm times, the deaf-blind user moves the clock hand to the desired alarm time, and then pushes the alarm set button. It will be possible to set alarm times as frequently as every 5 minutes. The medication reminder will be able to memorize up to 32 medication times per 24 hours, and differentiate AM from PM. The actual alarm signals will be vibration.

It is anticipated that the needs of some deaf-blind consumers would best be served by having a pharmacist or other healthcare professional set the clock and program the alarm times. We have therefore also incorporated a serial interface (RS-232C) for connecting the device to a computer for remote programming.

d. Conclusion

This new medication reminder is being developed specifically to meet the needs of deaf-blind persons. Its design will foster medical independence. It is anticipated that the transfer of responsibility from the caregiver to the deaf-blind consumer will be of great benefit to both parties. Work is continuing on further development of this prototype system.

G. ACCESS TO ELECTRONIC INFORMATION SYSTEMS (INTERNET/WWW)

The goal of this project is to help find ways of enhancing access by blind and visually impaired individuals to all forms of electronic information, including the Internet and World Wide Web.

1. World Wide Web Access

To help find ways of enhancing access by blind and visually impaired individuals to the Internet and World Wide Web, William Loughborough of our staff has contributed to many collaborative initiatives. We first compiled and published a web site at: http://www.ski.org/rehab/webaccess to provide an information resource on this subject for blind persons. It lists many informative links to sites by and for blind World Wide Web users and those who seek to further the goal of improved accessibility.

Mr. Loughborough has participated in panels at five World Wide Web conferences where the World Wide Web Consortium (W3C), the standard setting body for the Web, introduced the Web Access Initiative (WAI) with the goal of making the Web more accessible to people with disabilities. An outgrowth of the WAI resolution was the International Program Office (IPO) which organized several working groups to further the goals of the WAI. Mr. Loughborough participated in mseveral face-to-face working conferences at which the needs for accessibility were emphasized to the authors of new language standards for the web (HTML 4) and in particular Cascading Style Sheets, a method by which users can cause information on the Web to be presented in ways usable by people with disabilities. He also attended WWW7 in Australia and WWW8 in Amsterdam. In all cases he contributed materially to the outcomes by participating in panels and working groups and through personal contacts with the WWW Consortium members.

Mr. Loughborough has also participated in many teleconferences in these matters and served on three of the Working Groups in the WAI: Education & Outreach, Authoring Tool Guidelines, and Evaluation & Repair. These groups have been establishing the standards by which the rules governing Web Access will be defined in the future. We have particularly targeted the working group on Authoring Tool Guidelines, in the belief that the best way to ensure future access for blind persons to the Web is to ensure that the software that people use to generate web pages will produce only accessible output.

We have also participated in and monitored several E-mail lists on the Internet whose mission is to enhance the cyberworld's accessibility, including the interest groups of the WAI, the webwatch group who "police" web sites, the browser/screen-reader group, and other access related sources for braille support, etc.

We recognize that no one organization alone can ensure accessibility to the World Wide Web. We therefore believe that making contributions to collaborative and joint efforts with other members of the blind rehabilitation community is the best approach to continuing improvement in equality of access to Internet-based information.

2. Evaluating Web Access Software

In monitoring and evaluating developments in this field, we have attended access demonstrations by various software vendors and tested some of the simpler web access programs in our laboratories. The difficulties of installation, setup and operation of these programs, as well as the rapidly changing Net environment, have proven to be a major stumbling block to full access; at a demonstration at the CSUN conference no vendor was able to demonstrate reasonably full access to the Web. Since then, however, there have been some promising developments.

3. Telecommunications Access Initiative

Our RERC was a founding participant in a new non-profit telecommunications initiative to benefit underserved telephone customers, including persons with disabilities. This non-profit entity (which requested an endowment from Pacific Bell of $50 million to be dispersed over the next ten years) would have the responsibility of promoting "full and equal access to basic and advanced telecommunications services, and to meet undeserved communities' needs for information carried by those services." Smith-Kettlewell was invited to participate in the Research Group in part because of its prior R&D activities into enhancing accessibility to written material for blind persons, and to telecommunications for the deaf-blind.

This year, the project came to fruition when a definite commitment was obtained from the telephone companies involved to provide the funding needed for these efforts, and the initial symbolic check was presented to the non-profit entity at a ceremony attended by us last month.

Due to efforts such as ours and those of blind consumers, the need to provide accessibility features is slowly being acknowledged by the Internet community. The new release of Microsoft’s Internet Explorer supports limited keyboard operation, unlike previous mainstream browsers. Netscape, previously impossible to operate without a mouse, has now publicly stated its intention to provide access features in future releases.

William Loughborough of our staff has been acquiring expertise in Internet access and Java programming. He has also been investigating the use of special software agents to ease access by blind and visually impaired persons to the entire Internet, and is now taking the lead in our accessibility project. In exploring available solutions, Mr. Loughborough identified PW Webspeak, a non-visual browser, as the most effective tool available to blind users at present. He has been in contact with Productivity Works, which makes PW Webspeak, and is assisting in making blind persons aware of its existence. He also belongs to several blind Internet user groups, to monitor their concerns and assist with access problems.

4. New Braille/Graphics Printing Technology

Present Braille embossers commonly used in office situations are relatively expensive and slow compared to standard printers. They are also so noisy as to make occupancy of the office very unpleasant during operation, unless the printer is enclosed in a specially made soundproof box. Finally, they also contain many highly stressed moving parts to perform the embossing, reducing their reliability.

The Smith Kettlewell RERC previously developed a new solid state technology to address these problems. During the present reporting period, Quantum Technology, a well known manufacturer of Braille embossers and other technology for the US and international markets, conducted a worldwide search for new braille printing technology. After reviewing the potential of numerous proposed and prototype systems, Quantum settled on the system developed at Smith-Kettlewell as being the most promising.

This printing concept uses primarily electronic rather than mechanical technology, reducing the number of moving parts -- and therefore the manufacturing costs and the potential for breakdown -- to an absolute minimum. It also has the advantage of being almost completely silent, with a noise level no higher than that of a laser printer. Quantum has produced a full-scale working prototype printer with an eye toward making it into a product to facilitate the accessibility of all types of electronic and computerized information.

5. Exploring New Refreshable Braille Display Technologies

Another important barrier to computer and electronic information system access is the longstanding unavailability of low cost refreshable, dynamic braille/graphics arrays. To date, the only refreshable braille displays available have one, or at the most two lines of braille, and cost $4,000 to $14,000, placing them well beyond the reach of most blind consumers.

Accordingly, we initiated an exploratory project to examine the feasibility of radically new approaches to this problem. Our approach uses a combination of proven physical principles and fluids with state of the art electronic manufacturing techniques enabling arrays of braille cells and tactile graphics to be fabricated en masse rather than individually. Supplementary Field Initiated funding was obtained to pursue this project and explore its feasibility.

H. ENGINEERING DESIGN REVIEW METHODS

The goal of this project is to develop and publish a practical guide to design review and technology transfer for inventors and developers in the field of assistive technology.

We have often found that inventors of new ideas for devices to help people with disabilities underestimate the difficulties of carrying their inventions from a concept to a practical device, and from a prototype to a commercial product. They also tend to overestimate the market size and ability to pay, as well as the likely commercial returns they might receive. We have also found (from observing and participating in a number of engineering design review and technology transfer approaches used by a variety of organizations) that overly formalized approaches to these problems can easily lead to reduced innovation as well as missed “common-sense” considerations.

We have therefore adopted a simplified, down to earth approach to the project, in hopes of helping others learn the practical aspects of invention development and technology transfer, while benefiting from the experience of our own and others’ past shortfalls and successes.

Deborah Gilden, RERC Associate Director, synthesized these experiences and observations, drafting a paper designed to act as a brief, practical guide to the invention, development and technology transfer process for other developers. The paper incorporated our own experiences and many discussions with researchers, consumers and business collaborators -- pointing out many often-overlooked facts.

The final version of the paper, entitled “Moving from Naive to Knowledgeable on the Road to Technology Transfer”, was invited for publication by the journal Technology and Disability. It appeared in a special issue on technology transfer in July 1997, making it available as a resource for the entire disability research field.

I. LOW VISION RESEARCH

1. Visual Impairment Problems in the Elderly Population

a. Overview

Under funding from the National Eye Institute, the RERC is conducting perhaps the most comprehensive study ever undertaken of visual impairment in the oldest age groups. Conventionally, previous vision studies of the elderly have covered age ranges only up to approximately 75, and most have studied only basic measures of vision such as standard visual acuity. However, the rapid growth of the population in older age groups is causing an explosion in the low vision population, and there is a great need for the problems of these older age groups to be studied and solutions to be found.

In collaboration with the Buck Center for Research on Aging, we conducted a population-based study of elderly persons residing in Marin County, California. We tested 900 individuals over the age of 55; 400 subjects were in the 75 to 100 age group. Vision tests included high and low contrast letter acuity; low contrast, low luminance (SKILL) acuity; contrast sensitivity; disability glare; flicker sensitivity and CFF; color discrimination; stereo acuity; and visual fields with and without an attentional component. Questionnaire data on practical problems in daily life due to vision impairment, and measurement of abilities in walking, reading and other tasks are also part of the study. A second round of testing has been completed and a third is planned, in order to gain the benefit of longitudinal data in making individual predictions about risk factors.

b. Initial Results

Results have already greatly extended our knowledge of the degree, nature and prevalence of visual impairments in the older age groups. Publications are in preparation and in press detailing the results (see listing at end of this report). A joint presentation with the RERC on Aging took place at the American Society on Aging conference in March 1998. One finding was that even though median acuity under brightly lit, high contrast conditions (seldom encountered in real life) declines relatively slowly with age, there is great variation within the population. Most significantly, acuity under real world conditions of low contrast, low luminance and glare decline drastically beyond 75 years of age. The following table indicates the extent of this problem.