Accessible Pedestrian Signals: A Guide to Best Practices (Workshop Edition 2010) (2011)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Accessible Pedestrian Signals: A Guide to Best Practice C- 261 Appendix C: Research on APS Introduction Th roughout the preceding chapters of this Guide, references were made to various studies to support the recommendations given in those chapters. Interested readers may wish to see more details on the studies underlying these recommendations. Th is appendix was designed for such readers. Th is appendix reviews the research literature on APS including: • Research on problems of blind pedestrians • Eff ects of APS on particular street crossing tasks • Eff ects of APS on independence and confi dence • Eff ects of WALK signal characteristics • Accuracy and speed in identifying WALK indication for a given crossing • Research on source of the WALK signal • Research on other APS features • Other concerns and needs • Need for additional research

C - 262 Appendix C: Research on APS Research on problems of blind pedestrians INTRODUCTION Although APS have been widely used in Japan and Sweden since the 1960s, the early development of APS in those countries was not, as far as these authors have been able to ascertain, based on research. Nor was there any research basis for the “cuckoo/ cheep” APS that has been most commonly used in the U.S. Th e fi rst substantial research on APS appears to have been done in 1976 by Frank Hulscher, an electrical engineer with the Department of Motor Transport, New South Wales, Australia. Hulscher’s research was the basis for the well developed, fully standardized, and highly successful APS system in use in Australia today. Substantial research on APS in the U.S. began with a project undertaken by the San Diego Association of Governments in 1988. Th e results of this project were the basis for a policy of implementing standard signals at intersections in San Diego where a city access committee recommended them, following a systematic evaluation. Since 1996, there has been a concerted research eff ort related to APS. Several research studies and surveys have documented problems of pedestrians who are blind or who have low vision at signalized intersections without APS. (Barlow, Bentzen, and Bond, 2005; Bentzen, Barlow and Bond, 2004; Bentzen, B.L., Barlow, J.M. & Franck, L. 2000; Carroll, J. & Bentzen, B.L. 1999; Murakami, Ishikawa, Ohkura, Sawai, Takato and Tauchi,1998, San Diego Association of Governments, 1988). Several studies in the U.S. have compared travel by blind pedestrians with and without APS. (Barlow, Bentzen, Bond and Gubbe, 2006; Crandall, Bentzen and Myers, 1998; Crandall, Brabyn, Bentzen and Myers, 1999; Crandall, Bentzen, Myers and Brabyn, 2001; Marston and Golledge, 2000; Uslan, Peck and Waddell, 1988; Williams, Van Houten, Ferraro, and Blasch, 2005). In these studies, some of the APS have been signals mounted on the pedestrian signal head, while others have been receiver-based systems, and others been pushbutton- integrated devices. Information in this section is drawn from these studies. A brief summary of each follows. Surveys of blind pedestrians and orientation and mobility specialists San Diego Association of Governments (1988; and Szato, Valerio, and Novak, 1991a, b, c, hereafter referred to collectively as San Diego research) surveyed 71 national, regional and local organizations representing and/or serving elderly persons and persons who were visually impaired to determine their involvement in installation of audible signals and the level of support for audible signals; 36 responses were received. A separate survey was also sent to members of California Association of Orientation and Mobility Specialists to gather information about their experience with audible signals. Surveys were mailed to 67 members and 27 responses were received.

Accessible Pedestrian Signals: A Guide to Best Practice C- 263 In 1998, the American Council of the Blind (ACB) and the Association for Education and Rehabilitation of the Blind and Visually Impaired (AER) conducted similar surveys to determine problems experienced by blind pedestrians during street crossings. Problems with audible signals currently installed were also identifi ed by the surveys. • ACB survey (Carroll, J. & Bentzen, B.L. 1999, hereafter referred to as ACB survey) — surveys administered orally, in groups, to 158 pedestrians who are visually impaired • AER survey (Bentzen, B.L., Barlow, J.M. & Franck, L. 2000, hereafter referred to as AER survey) — mailed to 1000 orientation and mobility specialists. 349 surveys returned. Murakami, Ishikawa, Ohkura, Sawai, Takato and Tauchi (1998) conducted a survey of 50 blind pedestrians in Japan. More details about the survey results, and references to studies which support the results, are provided in the section on crossing problems. Crossing data Uslan, Peck and Waddell (1988) compared crossings by 27 legally blind pedestrians at three intersections in Huntington Beach, California, having “bird call” signals and one control intersection without APS. In research by Th e Smith-Kettlewell Eye Research Institute (Crandall, Bentzen and Myers, 1998, Crandall, Bentzen, Myers and Brabyn, 2001, Crandall, Brabyn, Bentzen and Myers, 1999, hereafter referred to collectively as SKERI research), 20 blind participants made a total of 80 crossings at 4 fi xed-time signalized intersections in downtown San Francisco, both with and without Talking Signs®. Talking Signs is a receiver-based APS. Intersection signal phases were pretimed and pushbutton use was not required. Th ere were nine measures, including measures of crossing timing (safety), orientation (precision), and independence. Marston and Golledge (2000) compared crossings by blind participants with and without APS, using the receiver-based APS system, Talking Signs, in a study investigating the use of Talking Signs remote infrared audible signs for a number of transit tasks. Th e eff ects of a pushbutton-integrated APS, a receiver-based APS manufactured by Relume, and typical visual pedestrian signals without APS on the street crossing behavior of 24 totally blind participants were compared by Williams, Van Houten, Ferraro, and Blasch (2005). As part of a project on blind pedestrians at complex intersections, funded by the National Eye Institute, National Institutes of Health, a series of studies on crossings by blind pedestrians with and without APS is in progress in four cities, Portland, Oregon, Charlotte, Tucson, and Cambridge, Massachusetts. Objective data on measures of street crossing performance by sixteen participants who were blind was

C - 264 Appendix C: Research on APS obtained at two complex, signalized intersections in each city. Measures were similar to those used in the SKERI research, including nine broad measures of crossing timing, orientation, and independence. Results from all four cities are not yet analyzed, but slightly diff erent analyses have been reported in several articles. Results from pre-installation testing in two cities, Portland and Charlotte, are reported in Bentzen, Barlow and Bond (2004. hereafter referred to as NEI-2 cities) and Barlow, Bentzen, and Bond (2005, hereafter referred to as NEI-3 cities) reports on three cities, Portland, Charlotte and Cambridge. Barlow, Bentzen, Bond and Gubbe (2006, hereafter referred to as NEI Portland pre-post) compare results of testing with and without APS in one of these cities, Portland, Oregon. PROBLEMS WITH SPECIFIC TASKS AT SIGNALIZED INTERSECTIONS BY PEDESTRIANS WHO ARE BLIND OR WHO HAVE LOW VISION Locating the crosswalk Pedestrians who are visually impaired use a variety of non-visual cues and strategies to locate crosswalks, none of which is absolutely reliable. Cues include proximity to the corner or to traffi c on the street parallel to the pedestrian’s direction of travel, location of curb ramps when present, location of idling cars on the street to be crossed, and presence of other pedestrians. When APS are present, they may also be used as an indication of the location of crosswalks. ACB and AER surveys did not include specifi c questions about locating the crosswalk. 78% of individuals in the Japanese survey indicated that locating the crosswalk was diffi cult at intersections without APS. In SKERI research, without APS, participants requested assistance in locating the crosswalk on 19% of street crossings. Participants were permitted to begin a crossing from any location that satisfi ed them, whether or not it was actually within the crosswalk lines. On 30% of trials, subjects who located the crosswalk independently began crossing from outside the crosswalk. (Participants were permitted to start crossing from any location so long as they were in no immediate danger.) Th e NEI-3 cities results indicated that participants started from outside the crosswalk area on 28.3% of crossings and requested or required assistance locating the crosswalk on 15% of crossings. Orientation (Aligning to cross and maintaining alignment while crossing) Pedestrians who are visually impaired typically use a number of clues to help them align to face the direction of travel on the crosswalk before beginning to cross. Th ese include vehicular sounds such as the direction of parallel traffi c fl ow and the location of idling traffi c on the street to be crossed. Another good cue, or strong predictor,

Accessible Pedestrian Signals: A Guide to Best Practice C- 265 of the direction of the destination corner is the direction pedestrians are traveling before they arrive at the starting corner. Additional cues that are useful in familiar locations are walls, fences, hedges, grasslines, and objects near the curb that have a straight surface that is either parallel or perpendicular to the direction of travel across the crosswalk. Th e direction of slope of the curb ramp, and the alignment of a truncated dome detectable warning near the bottom of the curb ramp are not reliable indicators of the direction of travel on the crosswalk, and when used either intentionally or inadvertently, may lead pedestrians who are visually impaired into the center of an intersection. Although there are numerous clues to the direction of travel on the crosswalk, none of them alone or in combination provide a defi nitive direction to pedestrians who are crossing at unfamiliar crosswalks. Two clues used to help pedestrians who are blind maintain their alignment while crossing are the distance and direction of parallel traffi c fl ow, and the location of idling vehicles on the perpendicular street. Even at familiar crossings these clues may not be suffi cient to provide positive guidance to pedestrians who are blind as they are crossing a street. In the AER survey, 97% of O&M specialists who responded indicated that their students had diffi culty aligning to cross the street while 66% indicated that their students sometimes had diffi culty knowing where the destination corner was. Th e most important reasons stated were that traffi c was intermittent or the intersection was off set. In the ACB survey, 79% of respondents indicated that they sometimes have diffi culty fi guring out where the destination corner is. Respondents to the Japanese survey indicated that “direction taking at the starting position” and “keeping direction while walking in the crosswalk” were a problem, even with APS. Th e broadcast sound from speakers mounted on the pedestrian signal head has not seemed to provide usable directional information. ACB and AER surveys indicated that pedestrians who are blind are sometimes not able to localize the sound of an APS in order to use it for guidance across the street. ACB – 6%; AER – 39%. 85% of ACB respondents indicated that they were sometimes confused by unexpected features such as medians or islands while 64% of O&M respondents indicated that their students had diffi culty with medians or islands. In SKERI research, participants started crossing from an aligned position on 48% of crossings without use of APS, and completed their crossings within the crosswalk on 58% of crossings. In the NEI-3 cities research (without APS), participants started from an aligned position on 73.4% of crossings. Participants requested assistance or required intervention for safety while aligning to cross on 10% of crossings. Of participants who aligned independently, 58.4% completed their crossings within the crosswalk. Using pushbuttons At 90 – 95% of signalized intersections today, signal timing is constantly adjusting to accomodate current vehicular and pedestrian traffi c, and pedestrians are required to push a button to actuate a pedestrian timing to cross the street. Unless a pedestrian

C - 266 Appendix C: Research on APS pushes the button, the WALK signal will not come on, and, when there is a green signal for parallel traffi c, it is timed to accommodate vehicular traffi c, not pedestrians. A pedestrian who does not push the button, and who crosses with the green signal for parallel traffi c, may not have enough time to cross the street. Th e vehicular signal timing may be as much as 20 seocnds less than the pedestrian signal timing (Barlow and Franck, 2005). Th erefore it is very important that pedestrians who are visually impaired use pushbuttons. Blind pedestrians experience a number of problems with pushbuttons. In the ACB and AER surveys, many respondents indicated that they or their students had diffi culty with pushbuttons: ACB – 90%; AER – 94%. Th e following reasons were given. • Users couldn’t tell whether they needed to push a button. • Users had diffi culty locating the button. • Users couldn’t tell which crosswalk was actuated by the button. • Th e pushbutton was so far from the crosswalk that users couldn’t push the button and then return to the crosswalk and align to cross before the WALK interval began. Uslan et al, (1998) also found that the major problems 27 legally blind participants had with “bird call” type APS, were in locating the pole and the pushbutton, and determining which pushbutton was for which crosswalk. Participants traveling with dog guides experienced the most diffi culty locating the pole. Sometimes participants fi rst located the incorrect button and subsequently located and pushed the correct button after waiting and listening through one or more cycles. Th e San Diego research also identifi ed problem with locating the pushbuttons and poles (Szeto et al. 1991a). Gallagher and Montes de Oca (1998) also noted this in research on vibrotactile-only APS. In the NEI-2 cities results, at crossings where pushbutton-actuation was required, participants looked for, found, and pushed the button on only 16.3% of these crossings in Portland, and none (0%) of the crossings in Charlotte. In research under NCHRP 3-62 (see NCHRP 3-62 Final Report, , and Bentzen, Scott and Barlow, in press), blind participants crossing at signalized intersections in Tucson, having APS, but without any instruction in use or purpose of the APS, failed to even search for pushbuttons on 29% of crossings. Identifying the onset of the WALK interval In the absence of an APS, pedestrians who are visually impaired rely on the stopping of perpendicular traffi c followed by the onset of parallel traffi c, to indicate the onset of the WALK interval. Pedestrians who do not quickly and accurately identify the onset of the WALK interval are likely to be delayed in starting to cross, may miss crossing on the fi rst pedestrian phase, or may begin crossing when the WALK

Accessible Pedestrian Signals: A Guide to Best Practice C- 267 interval is not in eff ect. Complex geometry and intersection signalization make the use of traditional clues to the onset of the WALK interval unreliable. Many respondents to the surveys indicated that they or their students sometimes had diffi culty knowing when to begin crossing: ACB – 91%; AER – 98%. Th e following reasons were identifi ed. • Th e surge of traffi c was masked by right turning traffi c. • Traffi c fl ow was intermittent. • Th e intersection was too noisy. • Th e surge of traffi c was too far away. In the AER survey, 79% of respondents indicated that blind students sometimes had diffi culty determining the onset of the WALK interval at intersections having exclusive pedestrian phasing. In the Japanese survey, 46% of respondents stated that “to take a timing to start” was diffi cult without APS. Uslan found that, at the control intersection without APS, which was considered the easiest to cross without APS, 4 out of 15 participants began crossing during DONT WALK. In SKERI research on 24% of trials when APS information was not available, blind pedestrians requested assistance in knowing when to start crossing. On 34% of trials on which they independently initiated crossings, participants began crossing during the fl ashing or steady DONT WALK. Th e NEI-3 cities research found that pedestrians who were blind independently began crossing during the WALK interval on only 48.6% of crossings and completed crossings after the onset of perpendicular traffi c on 26.9% of crossings. Th e need for pushbutton-actuation of the WALK interval aff ected the likelihood that participants would begin crossing during the WALK interval. On the pedestrian-actuated crossings, participants began crossing during the WALK interval on only 19.5% of the crossings compared to 71.7% of crossings where the pedestrian phase was on recall (the pedestrian phase was included in every cycle). Mean latency in beginning crossing (the time between onset of the WALK signal or near-side parallel traffi c and the participant beginning to cross) was 6.41 seconds. Problems with APS Th e ACB and AER surveys were not limited to information about crossings that did not have APS; they had questions about APS volume, and about confusion of tones. Volume Th e 1998 ACB and AER surveys reported the experience of pedestrians with visual impairments in using APS that had “bird call” signals, bells and buzzers. Th ere

C - 268 Appendix C: Research on APS were problems both with APS being considered too quiet and too loud. 45% of participants in ACB survey considered signals to be too loud while 71% considered them too quiet. 24% of AER participants considered signals to be too loud, while 52% reported that APS were too quiet. When APS are too loud, and are at intersections that are close together, the APS for one intersection may be heard from another, leading some pedestrians to incorrectly think they have the WALK interval. Th e surveys noted this problem: ACB – 19%; AER 25%. In addition, complex phasing can contribute to problems with APS volume. Uslan et al. (1988) found that, at one intersection with split phase timing, where the bird call signals for parallel crosswalks had separate timing, three of 15 blind participants initiated their crossings with the signal for the parallel crosswalk, walking into the paths of left-turning vehicles. Problems with APS — knowing which street has the WALK signal Researchers particularly evaluated data on the AER and ACB surveys from blind pedestrians and O&M specialists from California, whose experience with APS is almost exclusively with “bird call” signals that are intended to provide unambiguous information regarding which street has the WALK interval. Many respondents indicated that they or their students sometimes did not know which crosswalk had the WALK interval (ACB – 68%; AER – 72%). Reasons were the following. • Users forgot which signal was associated with which crossing direction; • Users didn’t know which direction they were traveling; and • Th e intersection was not aligned with primary compass directions. Th e San Diego researchalso indicated that blind pedestrians had diffi culty remembering which sound was for which crossing direction. Recent research in the NEI project and under NCHRP 3-62 addressed this issue. (See Chapter 3.) Problems with APS — Confusion of tones with other sounds Anecdotal evidence has existed for some time that some birds in the U.S. have calls that are like the cheep sound of some APS, and that other birds may mimic this sound. Th is has led to confusion between APS sounds, particularly the cheep sound, and the sounds of birds. AER and ACB surveys confi rmed that blind pedestrians really do confuse the sounds of birds with APS sounds. In the AER survey, 11% of respondents stated that their students had crossed the street with an actual bird and 10% stated that their students had not crossed because they thought the signal was a bird. Uslan et al. (1988) and the San Diego research also noted this as a problem.

Accessible Pedestrian Signals: A Guide to Best Practice C- 269 Pedestrian crashes To obtain a rough measure of the incidence of intersection crashes and near crashes for pedestrians who are visually impaired, the ACB survey asked respondents whether they had been struck by a car at an intersection, and whether they had had their long canes run over. In the ACB survey, 12 of 158 (8%) of respondents had been struck by a car at an intersection, and 45 (28%) had had their long canes run over.

C - 270 Appendix C: Research on APS Effects of APS on specifi c crossing tasks INTRODUCTION Since it is known that pedestrians with visual impairments have diffi culty with many of the tasks that, taken together, comprise street crossing, it would be desirable to think that APS improve all measures of crossing at signalized intersections. To determine the extent to which this is true, a number of research projects have obtained objective data comparing street crossing with and without APS on one or more of the following measures. • Locating the crosswalk • Aligning to cross and maintaining alignment while crossing • Use of pushbuttons • Identifying the WALK interval • Delay before beginning to cross • Independence in any or all crossings tasks Projects that have already been mentioned are: SKERI research; NEI Portland pre- post; Uslan et al, (1988); Marston and Golledge, (2000) and Williams et al. (2005). Th ere are several additional studies that shed light on this topic. As part of NEI research, in three diff erent experiments Wall, Ashmead, Bentzen and Barlow (2004), blind and blindfolded sighted participants made crossings at a simulated intersection, in the presence of recorded traffi c sound to determine the eff ect on crossing accuracy with signals comprised of bird calls, percussive “toks,” a click train, or a female voice. Th e signals, all of which were mounted at a height of eight feet, came either simultaneously from both ends of a crosswalk (typical practice), alternated from one end to the other, or came from the far end of the crosswalk only. Th ese researchers also compared crossing accuracy when signals came from two parallel crosswalks (typical practice), or from a single crosswalk. A further comparison was between WALK signals with a typical 7sec. duration versus a seven sec. WALK signal followed by a locator tone. Hulscher (1976) reported an estimate by Leith (personal communication) comparing starting delay pre- and post-APS installation. Wilson conducted a pre- and post-APS installation study of behavior of adult, non- disabled pedestrians at one intersection. Further, the research undertaken for NCHRP 3-62 (REF fi nal report chapters) contributed to the understanding of the eff ects of APS on specifi c crossing tasks.

Accessible Pedestrian Signals: A Guide to Best Practice C- 271 RESULTS OF RESEARCH Locating the crosswalk SKERI research found that starting crossing from within the crosswalk increased from 70% to 97% with use of the APS. NEI Portland pre-post research found signifi cant increases in participants’ ability to begin crossings from within the crosswalk at locations where pushbutton use was required, and pushbutton locator tones were installed. Pre-installation, 77% of crossings began from within the crosswalk; post-installation, 97% of crossings began within the crosswalk, indicating that locating the crosswalk was signifi cantly improved by the presence of pushbutton-integrated APS. (Barlow, Bentzen, Bond and Gubbe, 2006) Orientation (Aligning to cross and maintaining alignment while crossing) APS have some aff ect on aligning to cross, but results have not been as positive as desired. On 48% of crossings in SKERI research, where APS information was not available, blind pedestrians were not facing directly toward the opposite corner when they started their crossing; they were facing somewhat toward or away from the center of the intersection. With the use of receiver-based APS, participants were well aligned when beginning 80% of crossings. However, in NEI Portland pre-post research, alignment showed only a very small trend toward improvement, with 70% of independent crossings starting from an aligned position pre-installation and 84% post-installation. NEI research (Wall, et al, 2004) found the presence of a locator tone during the second half of the crossing had a positive eff ect on alignment. In the NCHRP 3-62 research on device features, statistical analyses revealed no diff erences in orientation attributable to device in Tucson, and only a few minor diff erences in Charlotte. Th ese diff erences are easily explained by unique characteristics of the given intersection or pole placements, and taken together with the lack of fi ndings in Tucson, suggests that none of the diff erences in APS device features on diff erent manufacturer’s APS had an important impact on pedestrian orientation. (See NCHRP 3-62 Final Report, and Bentzen et al, in press.) Various means of beaconing and providing directional information to blind pedestrians are being investigated in the NEI research, including increasing the volume of the locator tone and providing it from a directional speaker, and an orientation tone which involves increasing the volume of the locator tone on the opposite end of the crosswalk, before the crossing interval, in response to a request. Data collection and analysis of results from these experiments are not complete.

C - 272 Appendix C: Research on APS Using pushbuttons Data reported in NEI pre-APS installation testing (NEI-2 cities, NEI-3 cities) indicated that blind pedestrians typically didn’t search for and use pedestrian pushbuttons at unfamiliar intersections. In Portland, after installation of APS with locator tones (NEI Portland pre-post), participants were not instructed to use the pedestrian pushbuttons, nor were they provided with instruction about the fact that pushbuttons were needed to call the WALK interval at some of the crossings. However, the addition of pushbutton locator tones and the knowledge that an APS might be installed resulted in participants looking for the pushbuttons on over 98% of crossings, although looking for the pushbutton did not always result in fi nding and using the correct pushbutton. Participants independently looked for and used the pushbuttons on 93% of the crossings after pushbutton locator tones were installed. In the pretest situation, pedestrians had looked for and used pushbuttons on only 16% of crossings. NCHRP 3-62 research (See NCHRP 3-62 Final Report and Bentzen et al. in press) reported that while pedestrians who are blind do not typically use pushbuttons when making unfamiliar crossings, after being made aware of the function of the pushbuttons to call the WALK interval, and hearing the demonstrator locator tones, 100% of participants in Tucson and Charlotte searched for the pushbutton. Participants in both cities said they used the locator tone when locating the pushbutton on most trials, regardless of device. However, presence of a pushbutton locator tone did not guarantee that participants would fi nd and push the correct pushbutton. Errors in fi nding the correct pushbutton were reported, leading to recommendations for instruction in understanding and using tactile arrows, as well as maintaining orientation while looking for pedestrian pushbuttons. Initiating the crossing APS have been found to positively aff ect measures of delay in beginning to cross, starting crossing during the WALK interval, and completing crossings before the onset of perpendicular traffi c. Delay in beginning to cross Hulscher (1976) cites a personal communication from Leith (1975) in which Leith estimated that, following APS installation, delay in beginning crossings was reduced an average of 2-3 seconds for all pedestrians. Wilson (1980) in a pre- and post-APS installation study of adult non-disabled pedestrian behavior at one intersection., found the following results. • For pedestrians using the pushbutton, delay in beginning crossings was reduced by 20%, from 2.7 sec to 2.1 sec. • Th e time taken to cross by persons who started during the WALK interval decreased by about 5%; crossing time for other pedestrians was unchanged.

Accessible Pedestrian Signals: A Guide to Best Practice C- 273 • For pedestrians who arrived at the crossing during the fl ashing or steady DONT WALK and who waited to cross until the onset of the WALK interval, the proportion who failed to complete their crossings before the onset of opposing traffi c was reduced by one-half, from 22% to 11%. Williams et al. (2005) found that mean latency in beginning crossing without APS was more than 5 seconds, which was reduced to 2.2 seconds with a receiver-based APS device with a tone WALK indication and 3.8 seconds with a pushbutton- integrated APS using speech messages. In NEI Portland pre- post research, in 144 crossings pre- and post-APS installation at two intersections in Portland, Oregon, the weighted mean starting delay for blind pedestrians without APS was 5.1 seconds, and after APS installation, the delay was reduced to 2.9 seconds. Uslan et al, (1988) also found signifi cant diff erences between speed of crossing at control intersection and intersections where APS were installed; crossings at locations with APS were completed faster. Williams et al. (2005) assessed participants on total number of signal cycles missed before crossing. Without APS, mean wait time was almost 2 full cycles, while with either type of APS, the mean wait time was just over a half a cycle. In NCHRP research diff erent WALK indications have been found to aff ect latency to begin crossing. (See NCHRP 3-62 Final Report, Chapter 2,). Starting during WALK, and completing crossing before the onset of perpendicular traffi c In SKERI research, participants began crossing during the WALK interval on only 66% of crossings without APS, but on 99% of crossings with APS. In NEI Portland pre-post research, without APS the pedestrian-actuated crossings were highly problematic for pedestrians who are blind. Pre-installation, participants began crossing during WALK on only 25% of crossings. Post-installation, there was dramatic improvement in participants correctly determining the appropriate time to start crossing with 84% of crossings initiated during WALK. Similarly, participants completed crossing after the onset of perpendicular green on 50% of crossings pre-installation with a signifi cant decrease, post-installation, to 12% of crossings completed after the onset of perpendicular green. Furthermore, only 77% of decisions about when to start crossing were made independently pre-installation as opposed to 95% post-installation. Pre-installation, the total number of crossings where the individual independently determined a start time and actually began crossing during the WALK interval was less than a quarter of crossings. Post-installation, with the addition of the APS, there was a signifi cant increase both in independence and in beginning to cross at the appropriate interval. For crosswalks where the pedestrian phase was on recall, the APS sounded at the beginning of the WALK interval, regardless of whether the pushbutton was used or not. Th e WALK indication only sounded for the fi rst seven seconds of the

C - 274 Appendix C: Research on APS WALK interval, unless the pushbutton was pushed again. Pre-installation, 70% of independent crossings began during the WALK interval; post-installation, this increased to 100%. Marston and Golledge (2000) found that at crossings without APS, almost half (48%) of the participants attempted to cross during the steady DONT WALK interval, a time recorded as unsafe by the researcher. With access to the pedestrian signal information provided by APS, no participant started crossing at an unsafe time.

Accessible Pedestrian Signals: A Guide to Best Practice C- 275 Effect of APS on independence and confi dence Both independence and confi dence aff ect the likelihood that people who are blind will cross streets independently. Lack of independence and low confi dence in ability to cross safely result in lack of participation in normal community life. INDEPENDENCE Both the NEI research and earlier SKERI research on which the method was based measured independence on three street-crossing tasks both with and without APS: locating the crosswalk; starting to cross within the WALK interval; and completing the crossing. Th e NEI research also measured independence on aligning to cross. Th e percent of crossings on which participants were independent on each task is as follows. TASK WITHOUT APS WITH APS SKERI NEI SKERI NEI Locating crosswalk 81% 81% 99% 95% Aligning to cross NA 94.5% NA 97% Starting to cross during WALK 76% 79% 100% 92% Completing the crossing 81% 86% 97% 96% Confi dence Marston and Golledge (2000) measured confi dence in street crossing with and without APS. Th e range of responses for the no APS condition, by street crossing task, was 2.7-3.5 (5 pt. scale; 1=no confi dence, 5=very confi dent), while the range of responses by task for the APS condition was 4.8-5.0. Table C-1. Percent of crossing tasks on which participants were independent.

C - 276 Appendix C: Research on APS Effects of WALK signal characteristics INTRODUCTION Access Board Draft Guidelines for Accessible Public Rights-of-Way (revised 2005) and this APS Guide recommend using a rapid tick WALK indication. Research on WALK signal characteristics, described in this section, is the basis for that recommendation. It is important that any audible WALK signal be detectable and localizable, and, if it is a speech message, it must also be intelligible. Two factors make it diffi cult to satisfy these requirements. First, both detection and localization are infl uenced by the ambient sound, especially the sounds of vehicles. Second, any degree of hearing impairment makes it more diffi cult to detect and localize sounds and to understand speech. One might think that simply making the WALK signal loud enough would overcome both problems. However, the problems are much more complex. Not only volume, but the nature of the sound, that is, the spectral and temporal characteristics must also be taken into account. A body of research, reported below, describes research to identify the characteristics of sounds or messages that make them detectable, localizable, and intelligible. Some of this research particularly addresses detectability, localizability and intelligibility for people with hearing impairments. Th ere are also important limits to the amount of WALK signal noise that will be tolerated in neighborhoods, and limits to signal volume that are based on OSHA requirements. Furthermore, if the WALK signal is too loud, it may make it diffi cult for pedestrians who are blind to hear the sounds of vehicles that are about to cross their path. APS available today respond to ambient sound, providing WALK signals that are louder when the ambient sound is high, such as when a large truck accelerates, and quieter when the ambient sound is low, such as at night. Th is greatly increases the public acceptance of APS. APS users must not only be able to detect the WALK signal and localize it, they must be able to quickly and unerringly determine which crosswalk the signal is for. Th e nature of the signal and its location are critical in the performance of this important part of the street-crossing task. RESULTS OF RESEARCH Hearing and blind pedestrians A majority of persons who are severely visually impaired are age 65 or older, and typically have some amount of upper frequency, age-related, hearing loss. In addition, the incidence of hearing loss in people with visual impairments is higher than for the general population because a number of causes of blindness also result in hearing loss.

Accessible Pedestrian Signals: A Guide to Best Practice C- 277 Upper frequency hearing loss results in a decrease in the ability to localize sound and to understand speech, particularly in noisy environments (Wiener & Lawson, 1997). Characteristics of traffi c sound Th e sound produced by vehicular traffi c is concentrated in the low frequencies, especially for vehicles that are accelerating from a stop. Th e noise produced by accelerating vehicles is approximately 10 dB louder than that of vehicles traveling at a constant rate of speed. Th e mean intensity of accelerating traffi c, measured from the position of a pedestrian waiting to cross streets in residential and small business areas, was found by Wiener and Lawson (1997) to be 89 dB, equal to the maximum APS volume currently permitted by the MUTCD. (Th e 89 dB maximum in the MUTCD was based on OSHA 8 hour exposure limits). Th is means that signals at their maximum permitted volume will sometimes be diffi cult to hear. Detectability Hulscher (1976) found that, because of the masking of high frequency signals by predominantly low frequency traffi c noise, and because a majority of blind pedestrians have some upper frequency hearing loss, the optimal fundamental frequency of the WALK tone should be between 300 Hz and 1000 Hz, and the tone should be comprised of multiple short bursts of sound to aid localization. Staff eldt (1968), in research cited by Hulscher (1976), conducted extensive testing of APS at crossings where they were mounted on the pedestrian signal head, and found that an 880 Hz signal was most detectable in a background of traffi c noise. Hulscher’s recommendation and Staff eldt’s result was supported by Poulsen (1982) who compared the noise spectrum of traffi c as attenuated by windows to arrive at a recommended signal frequency (880Hz) that would not be largely masked by traffi c noise, but would also not transmit through windows and become a public annoyance. In San Diego research, laboratory measurements of “birdcall” signals from Nagoya Electric Works of Japan found that neither signal was highly directional. However, the cheep was more detectable than the cuckoo. Th e cheep was produced by a continuous frequency variation with a fundamental frequency base of 2800 Hz and the cuckoo consisted of two frequencies with a combined frequency base of 1100 Hz. (currently available “cuckoo/cheep” signals may vary from this manufacturer’s standard.) Hall, Rabelle & Zabihaylo (1996) worked with audiologists to develop a signal that provided the most localizable melody for an accessible signal and recommended a melody that was composed of fundamental frequencies between 300 Hz and 1000 Hz, but including harmonics extending to 7000 Hz. In NEI research (unpublished data), Wall, Ashmead, Barlow, and Bentzen carried out a series of experiments on detectability of WALK signal indications in a laboratory setting. Experiment 1 evaluated the detectability of signals in white noise, experiment 2 evaluated the detectability in traffi c noise, and experiment 3 evaluated detectability in

C - 278 Appendix C: Research on APS traffi c noise for subjects with age related hearing loss. Signals evaluated were an 880Hz square wave, a bird chirp, a cuckoo, two click trains, two percussive signals (“bink” and “tok”), a 4-tone melody, and female and male voice signals. Results indicated that audible signals of a more percussive nature with predominantly lower frequencies were best heard in background traffi c noise. In addition to one of the percussive signals, participants with age related hearing loss were better able to detect male voice signals. Signals with a simple percussive nature tended to need less gain to be heard in noise, relative to the levels necessary to be heard in quiet. In other words, percussive signals were more detectable at lower volumes. When asked their preference, most participants liked voice signals best, but all of the voice signals needed more gain to be heard. Note that the measurement in these studies was the point at which the signal was detected, not the point at which the message was easily understandable. Intelligibility of voice messages was not evaluated or assessed by participants. Th e cuckoo and chirp, most commonly used in the U.S. as an audible signal at the time of the study, required the most gain to be detectable amidst background noise. Localizability of WALK signal An additional factor, if audible pedestrian signals are to be used as beacons to guide pedestrian with visual impairments across a street, is how well signals can provide directional information. Laroche, Giguère and Poirier (1999) compared localization of cuckoo and cheep signals to localization of four four-note melodies varying in fundamental frequencies, harmonics, note duration, and temporal separation between notes. In combined objective and subjective testing, the cheep and a melody with minimal harmonics were found to be less localizable than the cuckoo and the other three melodies. In a follow-up study, Laroche, Giguère and Leroux (2000) compared the typical cuckoo-cheep sounds used in Canada with a cuckoo having a lower fundamental frequency, and the melody that was recommended as a result of their 1999 research. In both studies, the signal was 36 seconds long (much longer than typical U.S. WALK indications) and the measurements were in a simulated pedestrian corridor in a quiet environment. In situations with actual traffi c sounds, the cheep was found to result in signifi cantly greater veering and longer crossing time than any of the other signals, which did not diff er from each other. In NEI research, Wall, Ashmead, Bentzen and Barlow (2004) found no signifi cant diff erences in localizability among several disparate signals including cuckoo, chirp, “tok” and voice messages when tested in research that involved multiple crossings of a simulated street, in the presence of recorded vehicular sound. Th e fi ve signals used were representative of signals in wide use, or which showed promise for directional beaconing. None of the analyses indicated any systematic diff erences between the fi ve signals. Further experiments focused on presentation mode and signal location, rather than signal sound characteristics. Speech WALK indicatons Several APS systems in the U.S. are capable of producing directly audible speech messages, either from a speaker that is integrated into the pushbutton housing,

Accessible Pedestrian Signals: A Guide to Best Practice C- 279 or from a speaker at the pedestrian signal head. APS with speech messages are considered by many people to be especially user-friendly, when demonstrations are given indoors, to an audience for whom English is the predominant fi rst language. However if messages are not correctly understood by users, APS with speech messages, especially speech WALK indications, can lead to catastrophe. Intelligibility of speech messages is infl uenced not only by the relationship between signal volume and ambient sound, but also by the nature of the message, how familiar the hearer is with the English language, and any kind of hearing impairment that the user may have. Understanding speech in noise Listeners with normal hearing require that speech be 15 dB louder than background noise for intelligibility to reach 90% (Killion, 1999). Th is means that, in order to be intelligible, speech messages should be louder than tone indications. Th e eff ect of that louder sound level on the ability of blind pedestrians to hear other sounds in the intersection, or on near neighbors, may limit the acceptability of speech messages. At this time, MUTCD and Draft PROWAG limit the output of APS to 5 dB above ambient sound, except when special actuation requests a louder beaconing signal for a single pedestrian phase. As noted in the section on detectability, speech messages can be detected in traffi c sounds, but that does not guarantee that they can be understood. Speech WALK message structure and wording Bentzen, Barlow and Franck (2002) conducted research to obtain information from stakeholders regarding the structure and content of speech messages for APS WALK messages and for “pushbutton information messages” that are available during fl ashing and steady DONT WALK only. WALK messages convey that the WALK signal is on, and provide the name of the street being crossed. Pushbutton messages provide intersection and crosswalk identifi cation information, and may also provide information about unusual intersection signalization and geometry. Th e research utilized an expert panel of stakeholders, who prepared a survey comprised of sample messages to rate, and items to determine respondent understanding of messages. Th e survey was administered to people who are visually impaired, O&M specialists, transportation engineers, and APS manufacturers. Speech WALK messages should provide information to pedestrians who are blind that is similar to the information being provided to pedestrians who are sighted. Th e message should not be worded in a way that seems to provide a ‘command’ to the pedestrian. For example, ‘Cross Howard Street now’ would not be an appropriate message. Messages also should not tell users that it is “Safe to cross.” Research resulted in recommended messages for the WALK interval, which include begining with the name of the street being crossed, for example, “Broadway, Walk sign is on to cross Broadway”. Recommended messages begin on page 4 – 8 of this APS Guide.

C - 280 Appendix C: Research on APS Eff ect of Speech Messages on all pedestrians Van Houten, Malenfant, Van Houten and Retting (1997) found that redundant information conveyed by audible pedestrian signals increases the attention of all pedestrians to turning traffi c and may contribute to a reduction in pedestrian- vehicular confl icts and crashes at signalized intersections. Th eir research in Clearwater, Florida used prototype speech message technology in which speech messages were broadcast from the pedestrian signal head. When the pedestrian push button was pressed, the message was “Please wait for WALK signal.” Th e message “Look for turning vehicles while crossing [street name]” began 200 msec before WALK signals were illuminated. Th e signal also gave participants who were blind precise information about the onset of the WALK interval and which street had the WALK interval.

Accessible Pedestrian Signals: A Guide to Best Practice C- 281 Accuracy and speed in identifying WALK indication for a given crossing INTRODUCTION Access Board Draft PROWAG and the APS Guide emphasize the importance of installation location for APS. Installation location is critical for accurate and fast identifi cation of which crosswalk has the WALK signal. Research on WALK signal location and tone, described in this section, was the basis for that recommendation. APS must provide unambiguous information regarding which crosswalk has the WALK indication. A pedestrian who mistakes the signal for one crossing from a corner, for the signal for the other crossing, is at risk of making a crossing when vehicular traffi c has the right-of-way and pedestrian crossing is not permitted. A two- tone system such as cuckoo/cheep has long been assumed to provide unambiguous information, however San Diego surveys and Uslan et al. noted problems with this system. ACB and AER surveys confi rmed that there were numerous problems with it. Speed in identifying the WALK signal for the desired crossing is related to delay in starting to cross. Th e most desirable signal from this perspective is one that is quickly identifi ed, enabling users to initiate crossings promptly, and to fi nish crossing before the onset of perpendicular traffi c. RESULTS OF RESEARCH In NEI research, Ashmead, Wall, Bentzen and Barlow (2004) investigated the eff ects on accuracy and speed of identifying the correct crosswalk, of location of APS speakers located in diff erent positions relative to the crosswalk (see fi gure C-1 below for APS positioning used in research), using typical placements seen in the U.S. Under the most typical signal mode, simultaneous presentation from both ends of the crosswalk, the most accurate performance occurred when signals were placed close to the curb line, near the side of the crosswalk that was furthest from the center of the intersection (see corner 1 in fi gure C-1). Th e pedestrian could easily tell which of the two loudspeakers at the corner was active because each loudspeaker was close to the position of the pedestrian waiting to cross at the associated crosswalk. Note that this was true despite the fact that both signals on the corner had the same sound. Other arrangements of loudspeakers resulted in somewhat worse performance. Accuracy at corner 3 was poor, with participants answering correctly on only 25% of trials in the simultaneous presentation condition and 50% of trials in the alternating presentation condition. Th ere was no evidence that response time diff ered across the four corners, that is, for diff erent speaker arrangements. Th is suggests that the inaccurate judgments made from corner 3 about which crosswalk had the WALK signal did not refl ect uncertainty, but rather were mistakes of which the participants were largely unaware.

C - 282 Appendix C: Research on APS Th e above graphic illustrates the recommended placement of APS devices in relation to the crosswalk provided in the APS Synthesis in 2003. Th is arrangement was based on the NEI research described above. However, the Panel for the NCHRP Project 3-62 asked for additional research to determine if use of two poles at each corner was necessary, or if installation of two devices on one pole could result in accurate and fast identifi cation of which crosswalk had the WALK indication. Both the location of pushbutton-integrated APS, and various associated signals were investigated under NCHRP 3-62 at an intersection in Portland, Oregon. (See Figure C-3 for the locations of APS on each corner, and the sounds associated with each.). Ninety participants who were blind, who had low vision, or who had cognitive disabilities made all eight approaches to a 4-way intersection, were asked to push the button for to cross the street in front of them, and then to indicate when the WALK signal came on for the street in front of them. (Participants did not cross the street except as they were guided from one corner to the next). (See 3-62 Final Report and Scott, Myers, Barlow and Bentzen, 2006.) Figure C-1. Positions and headings of loudspeakers and pedestrians at each corner in NEI research. (Figure from Ashmead, et al., 2004) Note: The question to the pedestrian always was, “Which has the WALK signal, the crosswalk straight in front of you or the one to your right?” Corner 1: loudspeakers near curb, on outside of crosswalk line. Corner 2: loudspeakers near back edge of sidewalk, on outside of crosswalk line. Corner 3: loudspeakers near curb but facing across pedestrian’s position. Corner 4: loudspeakers near back edge of sidewalk mounted on the same pole. The fi gure is not drawn to scale.

Accessible Pedestrian Signals: A Guide to Best Practice C- 283 Results indicated that where pushbutton-integrated APS were mounted on separate poles, near the crosswalk line furthest from the center of the intersection, approximately 3 feet from the curb line, and approximately 10 feet apart, accuracy in judging when the correct crosswalk had the WALK signal was signifi cantly better than when APS were located according to other criteria (see corner C, Figure C-3). On Corner C, participants in the totally blind and legally blind subgroups indicated that the street in front of them has the WALK signal, when in fact the WALK signal was to cross the street beside them, on only 7.55% of trials. For all other tested arrangements, the same participants made this type of error on at least 26.9% of trials. Responses to the onset of the correct WALK signal were also signifi cantly faster at corner C. For the two subgroups with the least vision, use of two diff erent tones, when APS speakers were located on the same pole, resulted in errors in determining which street had the WALK interval on 50% of trials. Even when APS were located on two separated poles, accuracy in identifying the onset of the correct WALK signal was signifi cantly greater when both APS on the same corner used the same tone (rapid tick) than when the two APS used two diff erent tones (cuckoo and rapid tick). Speech WALK messages, when two APS were located on the same pole, were also evaluated in this research. Speech messages resulted in a much lower error rate than two tones, 19% vs 50%, however locating the APS on two separated poles and using the same tone on each for the WALK indication resulted in an error rate of less than 4%. Locating the APS speakers close to the crosswalk being signaled resulted in much better accuracy in identifying which street had the WALK interval Figure C-2. Optimal location of pushbutton- integrated APS.

C - 284 Appendix C: Research on APS than variation in WALK indications. Th e research recommended use of the rapid tick WALK indication because it produced the fastest and most accurate responses regarding which crosswalk has the WALK indication. Th is recommendation was in conjunction with recommendations for specifi c locations for the APS. Participants who had enough vision to see the pedestrian signal (identifi ed in the study as low vision or cognitively impaired) reported that they used the visual signal. However, the results for these groups are in the same direction as the results for the two subgroups with the least vision. It thus appears that the participants who could see the visual pedestrian signals might nonetheless have been infl uenced by the APS pole arrangement and signal sound. In research comparing devices having various features, conducted in Tucson and Charlotte under NCHRP 3-62, (see NCHRP 3-62 Final Report; Barlow et al., 2005; Bentzen et al., in press), there were three types of WALK indications used in the devices, speech messages, birdcalls (cuckoo-cheep), and the rapidly repeating tone indication (rapid tick). Once pedestrians understood the crossing signal, the rapid tick provided the best cue in terms of starting to cross quickly in both cities. Th e faster response to the rapid tick signal confi rms results of previous research on pushbutton location and nature of WALK signal, in which responses were faster to the tick than to two diff erent tones or to speech messages. Figure C-3. Locations of pushbutton-integrated APS and associated WALK signals in NCHRP 3-62 research

Accessible Pedestrian Signals: A Guide to Best Practice C- 285 Research on source of WALK signal INTRODUCTION Signals that have been typically installed in the U.S. have provided a loud beaconing WALK indication simultaneously from both ends of the crosswalk, and usually from two parallel crosswalks at the same time. As shown in the ACB and AER surveys (refs), pedestrians who are blind often have diffi culty if they try to use the APS to indicate the direction of travel on the associated crosswalk. Research has taken place in Canada, Japan, and the U.S. on the eff ect of source of the WALK signal on accuracy of aligning to cross and of making actual crossings. RESULTS OF RESEARCH Presentation mode — simultaneous, alternating, and far side Stevens (1993) and Tauchi, Sawai, Takato, Yoshiura and Takeuchi (1998) tackled the problem of improving localization of WALK signals (beacons) by varying the source of the sound. Th ey found that blind pedestrians could cross more quickly and with less veering when the WALK signal alternated back and forth from one end of the crosswalk to the other. Laroche et al. (2000) confi rmed the superiority and subjective preference for an alternating signal for beaconing at a simulated intersection, but found no advantage of the alternating signal when data were collected at an intersection with steady traffi c on both streets. Th is was true for all tones tested (chirp, cuckoo, low cuckoo, and melody). It may have been that blind participants had good directional information from vehicular sound at the intersection or it may have been related to the shorter duration of the APS sound when installed at the intersection. In the previous testing at a simulated intersection (in quiet), the WALK signal continued for 36 seconds, more than the time required for participants to complete the entire crossing. Testing at the simulated intersection was also in a quiet environment. Poulsen (1982) reported more accurate walking in a simulated crossing setting with a far end signal, than when a signal came simultaneously from both ends, and that far end signal was regarded favorably by a group of blind pedestrians in a fi eld test at a real intersection. Tauchi, Takami, Suzuki, Kai, Takahara, and Jajima (2001) examined the eff ects of alternating WALK signals in which the sounds from both ends of a 60 ft. long crosswalk at the top of a “T” intersection with alternating WALK signals were diff erent. Participants were better aligned, and maintained alignment better with the APS with diff erent tones at the end of the crosswalk than with the APS having the same tone at the end of the crosswalk.

C - 286 Appendix C: Research on APS Wall et al. (2004) compared the usefulness of auditory signals in guiding crossing behavior in three signal presentation modes, simultaneous, alternating and far side only. In several experiments, blind adults and blindfolded sighted adults crossed a simulated crossing with recorded traffi c noise to approximate street sounds. Audible signals were presented simultaneously from both ends of the crosswalk, alternating from one end to the other, or from the far end of the crosswalk only. Th e signals continued only for the typical U.S. WALK interval of seven seconds, stopping when participants were approximately halfway across the simulated street. Th e principal fi ndings were the same for blind and sighted participants and applied across a range of specifi c signals (e.g. chirps, clicks, voices). Crossing was more accurate when audible signals came only from the far end of the crossing, rather than the typical practice of presenting signals simultaneously from both ends. Alternating the signal between ends of the crossing was not helpful. However, providing a locator tone at the end of the crossing during the pedestrian clearance interval improved crossing accuracy. Th ese fi ndings off er less promise for the usefulness of the alternating signal mode, especially when the fi ndings for single vs dual crosswalks are considered. In previous studies, only one crosswalk was signaled, but this research compared signaling a single crosswalk versus two parallel crosswalks in two experiments. Errors were lower in the alternating mode than the simultaneous mode, but only when a single crossing was signaled. Th e customary practice of signaling two parallel crossings at the same time seemed to draw participants somewhat toward the center of the intersection.

Accessible Pedestrian Signals: A Guide to Best Practice C- 287 Research on other APS Features Previously described research has also looked at other specifi c features of APS including the pushbutton locator tone. Th e following sources are referred to: NCHRP 3-62 research, Barlow et al. 2005 and Bentzen et al., in press; Poulsen, 1982; San Diego research; and Williams et al., 2005. PUSHBUTTON LOCATOR TONES Pushbutton locator tones are a standard feature of almost all pushbutton-integrated APS in use worldwide. In the U.S. they are standardized to repeat once per second, and are to be audible only 6-12 feet from the pushbutton unless there is special actuation to raise the volume during the following pedestrian phase [should it be pedestrian change interval rather than ped phase; it really has nothing to do with completion of the crossing; using that language makes me think of some kind of pedestrian detection] (Manual on Uniform Traffi c Control Devices 2003 4E.09; Guidelines for Accessible Public Rights-of-Way, revised 2005) . Pushbutton locator tones inform blind pedestrians that they need to push a button to actuate a WALK signal and/or pedestrian timing. Because the sound comes from the pushbutton, it indicates the location of the pushbutton. Bentzen, Barlow, & Gubbé (2000), compared the speed of blind pedestrians on locating and walking to an APS with a pushbutton locator tone (880 Hz square wave, with multiple harmonics, 3 ms attack time, 15 ms sustained tone) at three repetition rates and three loudness levels relative to traffi c sound along an eight lane artery in Los Angeles. Best performance was with a repetition rate of once per second and loudness of 2 – 5 dBA above ambient sound measured at 1 meter from the locator tone speaker. TACTILE ARROW Tactile arrows aligned in the direction of travel on the associated crosswalk are features of all known pushbutton-integrated APS in use worldwide. Arrows vary in size and location on the APS. Some are on the pushbutton, some are on the vertical face of the housing and some are on the top (horizontal) surface. Th e length of the arrow varies from approximately 1.5 inches to 2.5 inches; stroke thickness varies from approximately 3/32 inch to ¼ inch; and height above the surface varies from approximately 1/6 inch to ¼ inch. Th e only research on usefulness of the tactile arrow for establishing crossing alignment was done in Denmark by Poulsen (1982). Th e “arrow” tested was a rod on top of the APS that was approximately 2.5 inches in length, ¼ inch in width, and ¼ inch high, with a bump indicating the far side of the crossing and additional bumps indicating the presence of islands or medians, if any. Alignment was equal with or without use of the arrow. Th e size and graspability of this unique arrow, as well as its location on the top of the APS, are thought to make it a better indicator of direction

C - 288 Appendix C: Research on APS than smaller, non-graspable, arrows and those mounted on the vertical face of the APS. Th e failure to fi nd any positive eff ect on alignment indicates that such an arrow (or probably any arrow) should not be considered a primary cue for alignment. Nonetheless, tactile arrows do serve the important purpose of indicating the crosswalk with which a particular APS is is associated. Research under NCHRP 3-62 (See Final Report and Bentzen et al. in press) found that increased familiarity with tactile arrows in Tucson and Charlotte resulted in an increase in use of tactile arrows, with arrows actually on the pushbutton being used more frequently than arrows that were not on the pushbutton itself. While participants looked at the incorrect (not desired crossing) arrows on 28 trials in Tucson, after extensive familiarization, they then found and pushed the correct pushbutton on all of these trials. In Charlotte, while there was a decrease in use of the wrong pushbutton following extensive familiarization, use of the wrong pushbutton on some trials remained. Under the same research, subjective responses indicating preference for various features indicated strong support for use of a tactile arrow to identify the correct pushbutton. VIBROTACTILE INDICATIONS Most pushbutton-integrated APS worldwide have vibrating tactile arrows or other surfaces that vibrate during the WALK interval. Th e vibration is required by pedestrians who are deaf-blind to inform them that the WALK signal is on. It is also used by some pedestrians without hearing loss to confi rm which crosswalk has the WALK signal, especially in very noisy conditions. Because it is necessary for pedestrians who are deaf-blind, as well as helpful for blind pedestrians in some situations, a vibrotactile WALK indication is required along with an audible WALK indication in the Draft Guidelines for Accessible Public Rights-of-Way. A signal having vibrotactile indication only is not permitted. An APS that is vibratory only gives no indication of whether there is a pushbutton, or where the pushbutton is located, and it gives no audible directional guidance. Gallagher and Montes de Oca (1998) surveyed blind pedestrians who were familiar with vibrotactile signals that did not have audible WALK indications, and on which a vibrating arrow was located on a horizontal surface above the pushbutton. Th ey found the vibrotactile signal to be well-liked. In fi eld research, they also found that use of the vibrotactile indication resulted in accurate crossing timing. TACTILE MAP OF THE CROSSWALK Maps of the crosswalk are standard features of pushbutton-integrated APS in Sweden, are in wide use around the world where Swedish devices are used, and are required in Austria for all APS regardless of the equipment manufacturer (see Chapter 4 for photos and more information]. Th ere does not appear to be any research on the legibility or eff ectiveness of these maps of the crosswalk, but they do

Accessible Pedestrian Signals: A Guide to Best Practice C- 289 have the potential to enable users who are unfamiliar with a particular crosswalk to anticipate such characteristics as the number of vehicular lanes in each direction, and the presence of islands or medians, rail tracks, and bicycle lanes. In NCHRP 3-62 research (see and Bentzen et al., in press) one of four devices compared had a tactile map of the crosswalk. Mean ratings of participant agreement in Tucson and Charlotte to the statement “Th e crossing map was useful and easy to understand,” were 4.00 and 4.17, re4spectively, on a 5-point scale (5 = strongly agree). However, even when they were thoroughly familiar with the map, only 9 of 40 participants used it across the two cities. PUSHBUTTON INFORMATION MESSAGE On some pushbutton-integrated APS, an optional feature is a speech message that comes from the pushbutton either whenever the button is pushed, or whenever the button is pushed and held in for an extended time (see Extended Button Press, above). Th is message always begins with the word “Wait”, as it comes on only during the fl ashing and steady DONT WALK intervals. Th e next information identifi es the intersection and the street to be crossed. Th e recommended format for this message is “Wait, to cross Howard at Grand.” Additional information may be provided regarding unusual signalization (e.g. split phasing) or geometry (e.g. narrow median in the roadway). In research conducted under NCHRP 3-62 (NCHRP 3-62 Final Report, Chapter 3; Bentzen et al., in press), three of four devices used in fi eld research in Tucson and Charlotte had pushbutton information messages. On all three devices, the pushbutton had to be depressed for at least three seconds to actuate the full pushbutton information message. One device did not have the standard pushbutton information message when used in Tucson. Th e objective measure most closely related to the pushbutton information message was only whether participants used the extended button press which actuated the pushbutton information message; it was not possible to observe whether participants actually understood or used the information provided by the pushbutton information message. However, when asked to rate the extent of their agreement with the statement “Th e pushbutton information message was easy to understand,” the mean ratings for each city were above 4.0 on a 5-point scale (5 = strongly agree), indicating that the message was usually understood. EXTENDED BUTTON PRESS Additional features on pushbutton-integrated APS may be actuated by an extended button press. Th ese features include a pushbutton information message, a louder (beaconing) signal, and extended crossing time. Noyce and Bentzen (2005) found that it was unusual for pedestrians to push pushbuttons for as long as one second. Th erefore an extended button-press of only one second is being standardized to actuate any optional features that are available at an APS.

C - 290 Appendix C: Research on APS In NCHRP 3-62 research (NCHRP 3-62, Final Report, Chapter 3; Bentzen et al, in press) the extended button press feature was included on three of four types of APS. Th e extended button press was little used except following familiarization with each device. Th e extended button press was used on 65-85% of crossings in Tucson and Charlotte following familiarization to device features. Th is may indicate that specifi c information and training are necessary for pedestrians who are blind, if use of the extended button press is expected. Desirability of a pushbutton information message as well as beaconing, both of which were actuated by an extended button press, was supported by subjective data. AUDIBLE BEACONING An optional feature on some currently available APS is audible beaconing, which is usually actuated only by an extended button press. Beaconing is provided by a louder signal during the next pedestrian phase. Th e beacon is intended to aid initial alignment for crossing and crossing within the crosswalk. In NCHRP 3-62 research (NCHRP 3-62 Final Report, Chapter 3; Bentzen et al, in press), one of four devices tested had the audible beaconing feature. On this device, the WALK signal and subsequent locator tone increased in volume for the next pedestrian phase following a button press of at least three seconds. No objective measure of the use of audible beaconing could be made. Th e only measure possible was use of the extended button press feature that also actuated a pushbutton information message. However, when asked to rate the extent of their agreement with the statement “Th e louder signal was helpful,” the mean responses for Tucson and Charlotte were 3.86 and 4.00, respectively, on a 5-point scale (5 = strongly agree), indicating that this feature is desirable. Additional research on the usefulness of APS for alignment and crossing within the crosswalk is provided under Eff ect of APS on specifi c crossing tasks: Results of research: Orientation, above.

Accessible Pedestrian Signals: A Guide to Best Practice C- 291 Other concerns and needs ENGINEERING CONCERNS Noyce & Barlow (2003) investigated problems reported with the interface between APS devices and signal controllers to determine whether there were systemic problems with the APS/controller interface. Most problems were found to be installer errors, or wiring problems that had already been corrected by the manufacturer by the time research was conducted. While the adjustment of the volume of the APS is critical for neighborhood acceptance and for usability by pedestrians who are blind, it continues to be an issue in many jurisdictions. In NCHRP 3-62 research on eff ect of device features (NCHRP 3-62 Final Report, Chapter ?) researchers found that it was not possible to adjust all devices to have the same perceived loudness, despite extensive eff orts and adjustments and involvement by manufacturers’ representatives. Perceived loudness is not amenable to objective measurement, and is infl uenced by conditions such as wind, humidity, precipitation, and nearby refl ective surfaces. APS devices, as with much new technology, have continued to generate maintenance and engineering concerns. As part of NCHRP 3-62 research, devices that were installed for human factors testing were monitored for a year by signal maintenance staff (see NCHRP 3-62 Final Report, Chapter 7). Concerns were expressed about more failures than expected, particularly of the vibratory feature of the devices. Manufacturers were said to be responsive to concerns and were continuously modifying devices to provide better durability. Case studies of devices installed in cold weather areas were also developed as part of Project 3-62 (see NCHRP 3-62 Final Report, Chapter 6). TRAINING NEEDS In NCHRP 3-62 Research in Tucson and Charlotte, experimenters observed that some participants did not have adequate information or techniques for using pushbuttons and APS devices in crossing streets and many did not have good information and understanding about the complexity of intersection signalization. Of concern to researchers were comments from some participants (prior to any explanation or training in the use of APS) who seemed to be trying to use the fact that the locator tone was sometimes louder in response to ambient sound as a WALK indication. Th ese participants generally did not look for or push the pushbutton, but heard the audible locator tone and without knowledge of locator tones and their function, and made an assumption that it was some kind of WALK indication. Th is misunderstanding of ambient sound adjustment and the locator tone could lead to dangerous crossings. Where APS with locator tones are installed, it may be necessary to make a concerted eff ort to provide information about the devices to individuals in the community.

C - 292 Appendix C: Research on APS Th e major reason for the tactile arrow is to enable users to identify which street the pushbutton controls. Many participants in NCHRP 3-62 research and in NEI research needed to be shown the arrows on the devices to understand which way the arrow pointed. Use of the tactile arrow and the incorrect pushbutton presented a very diff erent picture in Charlotte than in Tucson. In Tucson, participants who looked at the arrow on the incorrect pushbutton always correctly rejected the incorrect pushbutton. However, in Charlotte, even following familiarization, participants looked at the arrow on the incorrect pushbutton and failed to reject it on 20% of trials. Th is confusion in using the correct pushbutton seemed to be related to participants’ lack of strategies to maintain their orientation; some would completely turn to face the street parallel to their travel path, while examining the arrow on the device, then push the button and line up to cross the parallel street. Strategies for looking for pushbuttons and aligning to cross need to be taught.

Accessible Pedestrian Signals: A Guide to Best Practice C- 293 Need for additional research At a meeting in September, 2005, the project panel for NCHRP 3-62 developed the following list of issues for additional research: • Guidance on the need for pushbutton APS on side streets. • Wayfi nding/beaconing (NEI project may address this need) • Location/size/color of tactile arrow • Speech message clarity (consistency, programming and downloading messages, standardized library) • Tactile map (consistency, availability, info on median refuge, guidelines) • Cold weather issues (future follow-up on case studies, comparison study) • Tone types (acceptability to nearby residences, which are appealing) • APS eff ect on general pedestrian population (tie-in to tone acceptability study?) • Sound volume (best response to ambient sound, range and speed of response, range of decibel levels, resource is FHWA noise study offi ce) • Beaconing (volume, duration after press) • Combining speech message and tones • Audible message during fl ashing DONT WALK (tone vs. speech, countdown) • Design of curb ramps (use in orientation/wayfi nding, integration with signal design, eff ects on pole location)

C - 294 Appendix C: Research on APS References Adams, P.F., Hendershot, G.E., and Marano, M.A. Current estimates from the National Health Interview Survey, 1996. National Center for Health Statistics. Vital Health Statistics, 10, 200, 1999. Ashmead, D.H., Wall, R.S., Bentzen, B.L., & Barlow, J. M. (2004). Which crosswalk? Eff ects of accessible pedestrian signal characteristics. ITE Journal, 74(9), 26 – 31: Barlow, J. M., Bentzen, B.L., Bond, T. and Gubbe, D. (2006) Accessible Pedestrian Signals: Eff ect on Safety and Independence of Pedestrians who are Blind. Transportation Research Board 85th annual meeting compendium of papers. CD-Rom, Transportation Research Board, Washington, D.C.. [NEI Portland pre-post] Barlow, J.M., Bentzen, B.L. and Bond, T. (2005) Blind pedestrians and the changing technology and geometry of signalized intersections: Safety, orientation and independence. Journal of Visual Impairment and Blindness. 99(10), 587 – 598. [NEI 3-cities] Barlow, J.M. and Franck, L. (2005). Crossroads: Modern interactive intersections and accessible pedestrian signals. Journal of Visual Impairment and Blindness 99(10),599-610. Bentzen, B.L., Barlow, J.M. & Bond, T. (2004). Challenges of Unfamiliar Signalized Intersections for Pedestrians who are Blind: Research on Safety. Transportation Research Record: Journal of the Transportation Research Board, 1878, 51 – 57. [NEI 2-cities] Bentzen, B.L., Barlow, J.M., and Franck, L. (2000) Addressing barriers to blind pedestrians at signalized intersections. ITE Journal, 70(9), 32 – 35. [AER survey] Bentzen, B.L., Barlow, J.M. and Franck, L. (2004). Speech messages for accessible pedestrian signals. ITE Journal, 74(9), 20 – 24. Bentzen, B.L., Barlow, J.M. & Gubbe, D. (2000). Locator tones for pedestrian signals. Transportation Research Record 1705, 40 – 42. Bentzen, B.L., Crandall, W.F., and Myers, L. (1999) Wayfi nding system for transportation services: Remote infrared audible signage for transit stations, surface transit, and intersections. Transportation Research Record 1671, 19 – 26. [SKERI research] Bentzen, B.L., Scott, A.C., and Barlow, J.M. (in press) Accessible Pedestrian Signals: Eff ect of Device Features. Transportation Research Record: Journal of the Transportation Research Board.

Accessible Pedestrian Signals: A Guide to Best Practice C- 295 Brabyn, J.A., Haegerström-Portnoy, G, Schneck, M.E., and Lott, L.A. Visual impairments in elderly people under everyday viewing conditions. Journal of Visual Impairment and Blindness, 94, 741 – 755, 2000. Carroll, J. & Bentzen, B.L. (1999). American Council of the Blind survey of intersection accessibility. Th e Braille Forum, 38(7), 11 – 15. [ACB survey] Carter, D.L., Harkey, D.L., Bentzen, B.L., and Barlow, J.M. (in press). Development of an intersection prioritization tool for accessible pedestrian signal installation. Transportation Research Record: Journal of the Transportation Research Board. Centers for Disease Control and Prevention (CDC). “Blindness and Vision Impairment,” Online fact sheet, last update shown as 2003, available at http://www.cdc.gov/communication/tips/blindness.htm Crandall, W., Bentzen, B.L. and Myers, L. (1998). Remote signage development to address current and emerging access problems for blind individuals. Part I. Smith- Kettlewell research on the use of Talking Signs®at light-controlled street crossingss. Final report, Washington, DC, National Institute on Disability and Rehabilitation Research. [SKERI research] Crandall, W., Bentzen, B.L., Myers, L., and Brabyn, J. (2001) New orientation and accessibility option for persons with visual impairment: transportation applications for remote infrared audible signage. Clinical and Experimental Optometry. 84, 120- 131 [SKERI research] Crandall, W., Brabyn, J, Bentzen, B.L. and Myers, L. (1999). Remote infrared signage evaluation for transit stations and intersections. Journal of Rehabilitation Research and Development 36:341-355. [SKERI research] Draft Guidelines for Accessible Public Rights-of-Way (Draft PROWAG). U.S. Architectural and Transportation Barriers Compliance Board, Washington, D.C., 2002, revised November 2005. Available at http://www.access-board.gov/prowac/draft.htm Gallagher, B., Montes de Oca, P. (1998) Guidelines for Assessing the Need for Adaptive Devices for Visually Impaired Pedestrians at Signalized Intersections. Journal of Visual Impairment and Blindness, 92, 633-646. Harkey, D., Carter, D., Barlow, J.M., Bentzen, B.L., Scott, A.F. and Myers, L. (2006) Guidelines for Accessible Pedestrian Signals. Final Report on NCHRP 3-62 to National Cooperative Highway Research Program, Transportation Research Board, Th e National Academies. [NCHRP 3-62 research] Hall, G., Rabelle, A. & Zabihaylo, C. (1996). Audible traffi c signals: A new defi nition. Montreal: Montreal Association for the Blind. Health and Activity Limitations post-censal survey (HALS), Statistics Canada, Catalogue No. 89-542-XPE, 1995.

C - 296 Appendix C: Research on APS Hulscher, F. (1976). Traffi c signal facilities for blind pedestrians. Australian Road Research Board Proceedings 8, 13 – 26. Isler, Frederick. U.S. Access Board, Memorandum to Federal Highway Administration entitled “INFORMATION: Public Rights-of-Way Access Advisory,” January 23, 2006. Available at http://www.fhwa.dot.gov/environment/bikeped/prwaa.htm Killion, M.C. (1999). Guilt-free quick SIN [speech in noise]: When to give up on 4000Hz. International Hearing Aid conference V, University of Iowa. Larouche, C. Giguere, C. and Poirier, P. (1999) Evaluation of Audible Traffi c Signals for Visually-Impaired Pedestrians. Final Report, Institut Nazareth et Louis-Braille. Larouche, C. Leroux, T., Giguere, C. and Poirier, P. (2000) Field Evaluation of Audible Traffi c Signals for Blind Pedestrians. San Diego, Triennial Congress of the International Ergonomics Association. McMillen, Barbara. Handout at ProWalk/ProBike Conference (unpublished), Minneapolis, MN, September 2002. Manual on Uniform Traffi c Control Devices for Streets and Highways (MUTCD). U.S. Department of Transportation, Federal Highway Administration. Washington, D.C., Revision 1, 2003. Available at http://www.mutcd.fhwa.dot.gov Marston, J.R. and Golledge, R.G. (2000) Towards an accessible city: Removing functional barriers for the blind and vision impaired: A Case for Auditory Signs. Final Report submitted to the University of California Transportation Center. University of California Berkeley: University of California Transportation Center. Murakami, T., Ishikawa, M., Ohkura, M., Sawai, H., Takato, J. and Tauchi, M. (1998). Identifi cation of diffi culties of the independent blind travelers to cross intersection with/without audible traffi c signals. Th e 9th International Mobility Conference Proceedings. Rehabilitation Research and Development Center, Veterans Administration Medical Center, Decatur, GA. Noyce, D.A. and Barlow, J.M. (2003). Interfacing Accessible Pedestrian Signals with Traffi c Signal Control Equipment. Washington, D.C.: U.S. Access Board. Noyce, D.A. and Bentzen, B.L. (2005)). Determination of pedestrian pushbutton activation duration at typical signalized intersections. Transportation Research Record: Journal of the Transportation Research Board No. 1939, pp 63 – 68 Poulsen, T. (1982). Acoustic traffi c signal for blind pedestrians. Applied Acoustics 15:363-376. Public Rights-of-Way Access Advisory Committee (PROWAAC). Building a True Community: Final Report, Access Board, Washington, DC, January 2001. Rehabilitation Act of 1973, U.S. Code Title 29 § 794, 1973.

Accessible Pedestrian Signals: A Guide to Best Practice C- 297 San Diego Association of Governments. (1988) Evaluation of audible pedestrian traffi c signals. San Diego Association of Governments, San Diego, CA. [San Diego research] Schmeidler, E, & Halfmann, D. Distribution of people with visual impairment by community type, prevalence of disability, and growth of the older population. Journal of Visual Impairment and Blindness, 92, 380 – 381., 1998. Scott, A. C., Bentzen, B.L., Myers, L. and Barlow, J.M. (2005) Experimental Trials on Pushbutton Location and WALK Indicator. NCHRP 3-62, Guidelines for Accessible Pedestrian Signals, Task 5 Report, Chapter 3. Scott, A.C., Myers, L., Barlow, J.M., and Bentzen, B.L. (2006) Accessible pedestrian signals: Th e eff ect of pushbutton location and audible WALK indications on pedestrian behavior. Transportation Research Record: Journal of the Transportation Research Board, No. 1939, pp 69 – 76. Stevens, A. (1993) A comparative study of the ability of totally blind adults to align and cross the street at an off set intersection using an alternating versus a non- alternating audible traffi c signal. M.Ed. research report, University of Sherbrooke. Szeto, A.Y.H., Valerio, N.C., and Novak, R.E. (1991a). Audible pedestrian signals: Part I. Analysis of sounds emitted. Journal of rehabilitation research 28(2):57 – 64. [San Diego research] Szeto, A.Y.H., Valerio, N.C., and Novak, R.E. (1991b). Audible pedestrian signals: Part II. Prevalence and impact. Journal of rehabilitation research 28(2):65 – 70. [San Diego research] Szeto, A.Y.H., Valerio, N.C., and Novak, R.E. (1991c). Audible pedestrian signals: Part III. Detectability. Journal of rehabilitation research 28(2):71-78. [San Diego research]. Tauchi, M., H. Sawai, J. Takato, T. Yoshiura, and K. Takeuchi. (1998) Development and Evaluation of a Novel Type of Audible Traffi c Signal for the Blind Pedestrians, Th e 9th International Mobility Conference Proceedings. Rehabilitation Research and Development Center, Veterans Administration Medical Center, Decatur, GA, pp 108 – 109. Tauchi, M., Takami, R., Suzuki, S., Kai, T., Takahara, S., and Tajima, T. (2001). Comparison of disorientation and walking tendency of the visually impaired pedestrians under diff erent types of alternating audible traffi c signals. Proceedings of World Congress on Intelligent Transport Systems. Th e Lighthouse National Survey on Vision Loss: Th e Experience, Attitudes, and Knowledge of Middle-Aged and Older Americans, New York: Th e Lighthouse Inc., 1995. Transportation Equity Act for the 21st Century (TEA-21), Public Law 105-178, U.S. Congress, June 1998, Revised July 1998. Available at http://www.fhwa.dot.gov/tea21 Uslan, M. M., Peck, A. F., & Waddell, W. (1988). Audible traffi c signals: How useful are they? ITE Journal, 58 (9), 37-43.

C - 298 Appendix C: Research on APS Van Houten, R., Malenfant, J., Van Houten, J. & Retting, R. (1997). Using auditory pedestrian signals to reduce pedestrian and vehicle confl icts. Transportation Research Record No. 1578. Washington, DC: National Academy Press. Wall, R.S., Ashmead, D.H., Bentzen, B.L., & Barlow, J. (2004). Directional guidance from audible pedestrian signals for street crossing. Ergonomics. Vol. 47, (12), 1318 – 1338. Wall, R.S., Ashmead, D. H., Barlow, J.M., & Bentzen, B.L. (unpublished manuscript) Detectability of Audible Pedestrian Signals. Wiener, W. R., Lawson, G., Naghshineh, K., Brown, J., Bischoff , A., & Toth, A. (1997). Th e use of traffi c sounds to make street crossings by persons who are visually impaired. Journal of Visual Impairment & Blindness, 91, 435 – 445. Williams, M. D., Van Houten, R., Ferraro, J., and Blasch, B. (2005). Field comparison of two types of accessible pedestrian signals. Transportation Research Board 84th annual meeting compendium of papers. Washington, D.C.: Transportation Research Board. Wilson, D.G. (1980). Th e eff ects of installing an audible signal for pedestrians at a light controlled junction. Transport and Road Research Laboratory, Department of the Environment, Department of Transport, U.K.