An Evaluation of "Talking Signs" for the Blind

LESLEY A. BRABYN and JOHN A. BRABYN
The Smith-Kettewell Institute Rehabilitation Engineering Center
San Francisco, California

Requests for reprints should be sent to:
Lesley A, Brabyn
Smith-Kettlewell
2232 Webster St.
San Francisco, CA. 94115.

ABSTRACT

A new orientation system for the blind was evaluated in terms of its efficiency in assisting goal-oriented travel performance. Blind subjects were asked to locate rooms independently on one floor of a [large building. In one condition the subjects used audio speech signals presented via an infrared transmitter and hand-held receiver in the other condition they used conventions raised-print labels. Performance in terms of time required to reach goal was compared for the two conditions. The results indicate that it took significantly less time to reach the goal locations when subjects used the audio system. The role of such a system in aiding the orientation and travel efficiency of blind pedestrians is discussed at are additional applications of the system effective techniques developed by the subjects for using the device and suggested improvements for future versions.

INTRODUCTION

The Orientation Problem:

For the blind, the problem of determining absolute orientation in an environment designed for the sighted is a major one. Even when he or she possesses superior mobility skills, the blind traveler often finds that most cues that the sighted population uses to navigate through unfamiliar territory are unavailable. Street signs, store names, building addresses, and room numbers are only a few of the many identifying markers that the sighted use daily and take for granted. The blind must use more indirect methods to discern the same information. For example, although able to successfully negotiate them-selves through a busy street intersection, they may still have difficulty in discovering which intersection they are crossing.

The Architectural Barriers Act of 1968 (Public Law 90-480, amended through 1978) and Rehabilitation Act of 1973 (as amended through 1978 Sections 502, 506, and 507), mandated certain environmental changes in order to ensure that physically handicapped persons would be allowed the same ready access to public areas as that afforded to the non handicapped. The legislation focused public attention on the needs of certain handicapped groups and primarily resulted in efforts to improve wheelchair access. So far, however, little has been done to improve accessibility for the blind. Functional blindness has been estimated to affect approximately 498,000 persons in the U.S., with 46,000 new cases on average occurring each year (National Society to Prevent Blindness, 1980). This estimate does not include an additional 10,917,000 persons classified as visually impaired (National Society to Prevent Blindness, 1980). With so many visually handicapped people in the U.S. population, it is surprising that so little attention has been paid to the specialized needs of this group.

Possible Solutions:

The techniques currently used by the blind for determining the rudimentary cues needed for successful navigation are not entirely satisfactory. In order to discover "which bus?" "what room number?" "where's the elevator?" etc., the blind traveler must either actively seek out a sighted person for information or assume a passive and helpless stance until someone offers assistance. The use of a sighted guide does not rectify the situation, for in the basically dependent role of the guided, the blind person's decisions concerning daily activities must be contingent on the guide's availability. There are also disadvantages to a Braille- based system of markers, for, contrary to common public opinion, Braille is not the universal language of the blind. Goldish (1967) estimated that only 25% of the blind population in the UPS. are able to read Braille, 18% use it, and 8% use it as a primary reading mode. A study by Gray and Todd (1965) showed a close relationship between the ability to read Braille and the age at onset of blindness. Of the congenitally blind, 72% were found to be Braille readers, whereas only 11% of the group who had become blind after the age of 50 read Braille. With the incidence of vision-related problems concentrated primarily among the elderly (Genensky, 1978), this means of communication does not seem to be the answer, due to the low literacy rate. A possible alternative system of navigational markers might use raised-print panels wherever navigational signs are located It is not clear what proportion of the blind population could read such signs, nor at what speed, but the effort needed to locate such signs would hardly be a dignified or efficient method for the blind traveler. An ideal system of navigation and orientation lot the blind should therefore be capable of independent operation by the user at a distance and should provide the desired information in a universally understood form. To meet these criteria, an electronic navigation system (ENS), called "Talking Signs," has been developed (Loughborough, 1979). Through this system, the message of a sign can be translated into modulated infrared light waves and transmitted continuously via a miniature infrared source that is invisible to the sighted population. The transmitting light source is placed at a location near the written sign, and the user of the system has only to point a small, hand-held receiver in the general direction of the transmitter for the sign to be decoded into speech output. More comprehensive descriptions of the technical design and specifications of the system can be found elsewhere (Rehabilitation Engineering Center, 1981).

Objective of the Study:

The ultimate goal of an electronic navigation system of the type described is to make absolute orientation markers easily accessible to the blind, the visually impaired, and the reading impaired. The basic physical parameters of such a system have been defined and described (Schenkman, 1981). The speech quality of the transmitters and receivers has been tested and the verbal messages have been found to be easily recognizable (Brabyn and Brabyn, 1982). Another necessary step in the evaluation of the system is to demonstrate whether it actually enables the blind traveler to find specific locations.

The objective, then, of this study was to compare the orientation performance of blind persons using an ENS with their performance using raised-print labels. The raised-print signs were chosen as a standard for comparison since they appear to represent the best ''conventional" possibility for a navigational orientation system. The performance criterion was the time taken by the blind travelers to locate specific rooms in an unfamiliar building, (a typical problem faced by the blind traveler) using the ENS and raised- print labels, respectively.

METHOD

Subjects:

Twenty-one totally blind adults (5 females and 16 males) ranging in age from 28 to 77 years (with a mean of 36.4 years) participated in the study. Thirteen of the subjects were congenitally blind the remaining eight subjects had been blind for a mean of 73.8% of their lives. All subjects used the long cane as their primary mobility aid and had received from 0 to 4.5 years of formal mobility training. Eighteen of the subjects rated their previous experience with raised lettering as ''considerable" the remaining three described it as "some to moderate." The subjects, who were paid for their participation in the study, were uninformed as to the numerical sequencing arrangement of the rooms on the experimental floor of the building and had little or no prior experience using any type of electronic navigation aid.

Apparatus:

A floor of the research building was equipped with ENS transmitters, one placed in the middle and above each of the 14 door-ways at a height of 218 cm from floor level. The transmitters were programmed to indicate such specific locations on the floor as room numbers, exits, and restrooms. Each door was also fitted with a corresponding 5 cm x 10 cm raised-print label using 2-cm-high characters, centered and positioned at a standard height of 132 cm on the door itself.

Before the experimental session began, each subject was taken to a separate room away from the experimental floor and was given a brief training period with the ENS and the raised-print labels. The training, which lasted between 5 and 10 min., consisted of the subject's familiarization with the speech output of the ENS, instruction on how best to hold and operate the receiver, and practice at locating the beams from three test transmitters. The subject also examined examples of the raised-print labels and practiced locating and reading them at the standard height on a test door.

When the subject reported feeling reasonably comfortable using both systems, he or she was taken to the test floor and was asked to locate and stand in front of one of the 14 different doorways, which the experimenter designated by its name or number. In half of the trials, the subject used the ENS to find and identify the target doorways in the other half, the raised-print labels were used. In all, 14 trials were given per subject, 7 trials in each condition. The order of locations and method of finding the target sites were randomly assigned with the restrictions that no subject was required to find the same location twice, and the total distance traveled was equated for each subject and each condition. During the experimental session, the subject was permitted to use a cane. In order to simulate real-life conditions more closely, no attempt was made to curtail the usual flow of traffic on the floor. However, to enable the independence of travel in the two conditions to be assessed, subjects were not allowed to ask directions of passers-by. Each trial was begun from the same starting place, and the speed of the subject's progress was timed by the experimenter with a stopwatch. The length of each trial was calculated from the time the subject took the first step until he or she reached the target location and verbally identified it as being the correct doorway. In both conditions, the subject was required to place one hand on the doorway being identified before the timer was stopped. In order to monitor the experimental situation and keep the subject from possible harm due to the unexpected appearance of obstacles, the experimenter stayed close (but not obtrusively so) to the subject. The experimenter wore rubber-soled shoes and remained silent throughout each trial.

RESULTS

For each subject, the mean time required to reach the target locations was determined for the two conditions. Scores ranged between 17.0 and 132.6 s, with a mean of 64.4 and standard deviation of 35.8 for the ENS condition. In the raised-print condition, scores ranged between 30.1 and 241.0 s, with a mean of 86.2 and standard deviation of 58.7. Seventy-one percent of the subjects took less time to reach the target locations using the ENS, and a one-way analysis of variance for a repeated measures design (Keppel, 1973) indicated that the difference between the two methods was significant, F(1, 20) = 5.00, p 0.05. A summary of the analysis of variance is presented in Table 1. In order to determine if differential learning rates could have contributed to this significance, learning curves were plotted for the two conditions. When best-fit linear functions were drawn through the curves, the difference between the two slopes was only 7.5 deg, whereas the intercepts differed by 40.1. The learning curve had flattened by the third trial for the ENS condition and by the fourth trial for the raised-print condition. This suggests that only marginal differences exist in the learning rates between the two conditions. A correlational analysis of the data (Horowitz, 1974) showed that there were no correlations in either condition between performance and either age, amount of formal mobility training, or percentage of life blind.

Table 1: Summary of the Analysis of Variance

The First column is "source" and includes three parameters: Method, Subjects, and Method x subjects

The second column is "df" and includes three parameters: 1, 20, and 20

The third column is "MS" and includes three parameters: 5014.03, 3725.76, and 1003.16

The forth column is "F" and includes one parameter on the first row: 5.00

The fifth column is "p" and includes one parameter on the first row: 0.05

The "Total" for the second column parameter (df) is 41.

The "Total" for the third column parameter (MS) is 2429.16.

DISCUSSION

The results of this study show that an electronic navigational system of the type described can significantly assist the blind traveler in finding specific locations. It is noteworthy that these positive results were obtained indoors, since in outdoor situations raised print or braille might be more difficult to locate, and the "remote reading" abilities of an ENS would be even more advantageous. Subjects had no prior experience with electronic navigation systems and received minimal training during the experiment with the device being tested however, 86% of the subjects expressed considerable previous familiarity with raised-print labeling. Thus, we feel that the performance test used in this experiment was conservative. Although engineering and technology can provide many travel aids and other devices for the handicapped, the success of such devices depends on their acceptance by users. The consumer must like using the aid and it must provide a solution to an actual problem.

Consequently, the most significant finding of this study may be the subjects' positive responses to the remote system. Subjects experienced little difficulty in learning to operate the system, and they thought that it enabled them to reach the desired locations independently and inconspicuously. They also remarked upon the convenient size of the receiver it fit comfortably in the hand when in use and could be carried in a shirt pocket or a small handbag when not needed. During the course of the experiment, the subjects developed several innovative techniques for accurately locating the transmitter beams. For example, two subjects first determined the correct vertical angle at which to hold the receiver for locating the transmitters, then walked briskly through the corridors while keeping the ''on" button depressed and the receiver properly tilted. By using this method, they were able to "read" each room number as they went by. This method produced very low room-detection mean times of 33.6 and 28.4 s. The participants suggested additional uses and applications of a remote navigational system based on their own needs. These included marking of "walk" or "don't walk" signs, schedule and boarding location information in public transport stations, floor numbers in elevators, and store names and addresses on buildings. They further suggested that a portable transmitter be developed that could be temporarily positioned on a location that the user wanted to find again, such as a hotel-room door. Since the version of the ENS tested in this study was still in its prototype stage, suggestions for improvement of subsequent models were elicited from the subjects. Complete details of these features appear elsewhere (Rehabilitation Engineering Center, 1982), and several of the suggested improvements have been incorporated in the most recent receiver model, which is soon to be available for evaluation.

In general, the statistical results of this study as well as the positive reaction by the subjects to the remote type of navigation system under evaluation were most encouraging. Clearly, there is a need for an orientation system such as this. The technology is available, the system does improve independent travel performance, and the user community appears to be most receptive to the idea.

REFERENCES

Brabyn J. A., and Brabyn L. A. Speech intelligibility of the Talking Signs. Journal of Visual Impairment and Blindness, 1982, 76, 77-78.

Genensky, S. Data concerning the partially sighted and functionally blind. Journal of Visual Impairment and Blindness, 1978, 72, 177-180.

Goldish, L. B. Braille in the United States: its production, distribution, and use. New York: American Foundation for the Blind, 1967.

Gray, P. G. and Todd, J.E. Government social survey: Mobility and reading habits of the blind. London: Ministry Of Health, Her Majesty's Stationery Office, 1965.

Horowitz, L. M. Elements of statistics for psychology and education. New York: McGraw-Hill, 1974.

Keppel, G. Design and analysis: A researcher's handbook. Englewood Cliffs NJ: Prentice-Hall, 1973.

Loughborough, W. Talking Lights. Journal of Visual Impairment and Blindness, 1979, 73, 243.

National Society to Prevent Blindness, Vision problems in the United States: Data analysis. New York: Author, 1980.

Rehabilitation Engineering Center, The Smith-Kettlewell Institute of Visual Sciences Annual Report of Progress: April 1980 to March 1981. San Francisco: Author, 1981

Rehabilitation Engineering Center, The Smith-Kettlewell Institute of Visual Sciences Annual Report of Progress: April 1981 to March 1982. San Francisco: Author, 1982.

Schenkman, Personal communication, 1981.