Pedestrian actuation affected the ability of participants to begin crossing during the walk interval. At the two intersections in Portland and Charlotte, it was necessary to push a button to actuate a pedestrian phase for all crossings of the major street, and pushing a button was also required for the right turn lane in Portland. However, participants looked for, found, and pushed the button on only 16.3% of these crossings in Portland, and no crossings in Charlotte. For these pedestrian actuated crossings in both cities, participants began crossing during the walk interval on only 19.5% of crossings. However, for the crossings for which the pedestrian phase was on recall or was pre-timed, in all three cities, participants began crossing during the walk interval on 63.8% of crossings.

　

Orientation: Collapsing across all three cities, participants began crossing from within the crosswalk on 71.7% of crossings; they began crossing from an aligned heading on 73.4% of crossings; and they ended within the crosswalk on 57.7% of crossings. For each city, he proportion of trials on which participants completed crossings within the crosswalk was considerably less than the proportion on which participants started within the crosswalk or started in an aligned heading. In Charlotte, participants completed crossings within the crosswalk on less than 50% of trials. This was probably attributable to very skewed crosswalks at one intersection and a stop line at one crossing of the other intersection that was at a considerable distance from the crosswalk. (See Table 2.)

　

　

Independence: Participants requested assistance on various tasks, and on some crossings the experimenter had to intervene for participant safety. Intervention occurred only when the experimenter who was responsible for participant safety could not be certain that all drivers in the vicinity of the participant were aware of, and yielding to the participant. Collapsing across the three cities, and combining data for requests for assistance and for intervention, there were assists or interventions in locating the crosswalk on 15.0% of crossings; there were assists or interventions on aligning to cross on 9.8% of crossings; there were assists or interventions on deciding when to begin crossing on 18.8% of crossings; and there were assists or interventions during crossing on 11.7% of crossings. (See Table 3.)

　

　

 Pedestrians who are blind have traditionally learned that the appropriate time to begin a crossing at a signalized intersection is when there is a surge of traffic on the side of the parallel street that is closest to them. At most intersections, this coincides with the onset of the walk interval. However, this is not always the case, many factors make it difficult to determine from available acoustic information just when that time occurs. If the pedestrian does not actuate the pushbutton, a walk interval may not occur-the walk signal will not come on, and the amount of time before opposing traffic is permitted to move may be timed for vehicular needs and be too short to permit pedestrian crossing. Where an intersection is very wide, has abundant turning traffic, or is very noisy, it may be very difficult for blind pedestrians to determine the onset of the walk interval using customary techniques. Where there is little parallel traffic, where there is a leading pedestrian interval, or where there is exclusive pedestrian timing, it is often impossible to determine the onset of the walk interval from available acoustic information.

　

 The results indicate that participants had difficulty determining the onset of the walk interval in all three cities. Of those who crossed independently, 51.4% did not start during the walk interval, and 18.8% either requested assistance determining the onset of the walk interval, or required intervention to prevent a possible crash. Across all three cities, participants delayed the beginning of their crossings for a mean of 6.41 seconds after the onset of the walk interval (or straight-ahead parallel traffic on the near side of the street). Factors accounting for the delay included time taken to determine that the walk interval had, in fact, begun, and time for vehicles turning across participants path to pass in front of them. Without accessible pedestrian signal information, pedestrians who are blind have no possibility of claiming the intersection at the onset of the walk interval, before turning traffic. Therefore, even when they are able to determine that the pedestrian timing has begun, blind pedestrians may not initiate their crossings until well into the pedestrian clearance time. Furthermore, they may be completely unaware that they are beginning their crossings late because traffic clearing the intersection does not permit a recognizable surge of parallel traffic. Pedestrians who are blind depend on acoustic information from moving and idling vehicles to help them locate crosswalks, align for crossing, cross directly to the destination corner, and complete their crossings within the crosswalk. Where crosswalks are very skewed, very narrow, or not located at the corner, it is common for pedestrians who are blind to begin and continue their crossings near, but not within, the crosswalk.

　

 At some locations, and at some times of the day, there is insufficient vehicular traffic for blind pedestrians to determine, by listening, the location or direction of the crosswalk, or to help them maintain a straight path of travel across the street. Cars idling at a stop line are a good predictor of the location of the crosswalk, and cars going straight ahead are a good predictor of the location of the opposite corner. Cars traveling straight ahead, in the same direction as the blind pedestrian, provide a "wall of sound" that many pedestrians who are blind can use to keep on a straight trajectory toward the opposite corner. When there are no cars idling on the street to be crossed, and no cars traveling parallel to, and at the same time as blind pedestrians, primary orientation information is not available to them.

　

 The results of this experiment confirm that pedestrians who are blind do have difficulty aligning, as previously found in a survey of Orientation and Mobility Specialists (Bentzen et al., 2000) and as documented by Bentzen et al. reporting on street crossings by pedestrians who were blind at four intersections in San Francisco. (Bentzen et al., 1999) Some blind pedestrians use the location of curb ramps to indicate the location of the crosswalk. While this may indicate the location of crosswalks, many pedestrians who are blind are reluctant to veer from their line of travel as they approach a corner. The proprioceptive information blind pedestrians gain as they travel straight on a sidewalk approaching a corner is another excellent predictor of the direction of the crosswalk. Thus, it is the observation of many blind pedestrians and mobility specialists that if blind pedestrians have to search for a curb ramp, they find it more difficult to align to face the opposite corner. Additionally, if they locate and travel down a curb ramp that is at the apex of the corner, they are more likely to veer away from the crosswalk.

　

 The results regarding requests for assistance and interventions for safety are in the same direction as found by Bentzen et al. (1999) Also, in a manner similar to that found by Bentzen et al., despite repeated instruction to participants to ask for assistance if they would normally ask for assistance in such crossing tasks, participants tended to attempt to cross independently even though they were obviously very challenged by the crossings. Participants were aware that an Orientation and Mobility Specialist was with them at all times and was responsible for their safety. This assurance may have resulted in a lower frequency of requests for assistance than would be expected in non-experimental situations. On the other hand, it is the frequent experience of pedestrians who are blind that there are no other pedestrians at a crossing when they need to cross, so a desire for assistance cannot always result in obtaining it. The need for intervention probably reflects, in some cases, that participants should have requested assistance because they simply did not have enough information to make good crossing decisions.

　

 Accessible pedestrian signals have now been installed at all six intersections, and additional data collected on pedestrians who are blind crossing the same intersections. A subsequent paper will provide comparisons between pre- and post-installation performance.

　

 Bentzen, B.L., Barlow, J.M. and Franck, L. Addressing barriers to blind pedestrians at signalized intersections. ITE Journal. Volume 70-9, September 2000, pp. 32-35.

　

 Bentzen, B.L., Crandall, W.F., and Myers, L. (1999) Wayfinding system for transportation services: Remote infrared audible signage for transit stations, surface transit, and intersections. Transportation Research Record 1671, pp. 19-26.