David Guth, Richard Long, Paul Ponchillia
Western Michigan University
Dan Ashmead, Robert Wall
Support for this work was provided by the US Access Board, the National Eye Institute of the National Institutes of Health, and the the American Council of the Blind
This report is a summary of a research project conducted in the Baltimore, Maryland metropolitan area in April, 2000. The study was the first of a series of research projects to be conducted to evaluate access to modern roundabouts by pedestrians who are blind. The long-term goals of the research program are to determine where interventions may be needed at roundabouts to facilitate safe and efficient street crossings, to identify potential interventions, and to evaluate the effectiveness of these interventions.
The Americans with Disabilities act and its implementing regulations require that public rights-of-way be accessible to all users, including pedestrians with disabilities. Some roundabouts may pose barriers to safe and efficient independent travel by some individuals with blindness and low vision. The research described in this report provides information about the ability of blind pedestrians to use their hearing to distinguish "crossable" gaps in traffic at roundabouts from gaps that are too short to afford safe crossing. "Crossable gaps" were defined as those that would have allowed pedestrians sufficient time to cross from a curb line to a splitter island before the arrival of the next vehicle at the crosswalk.
The study was conducted at three modern roundabouts in metropolitan Baltimore, MD with assistance from the Maryland Department of Transportation. The three roundabouts included a large, high-volume, urban, two-lane roundabout (the Towson Roundabout); an intermediate volume, urban, two-lane roundabout (the Annapolis Gateway Roundabout); and a low-volume, single-lane roundabout (the University of Maryland – Baltimore County Roundabout, "UMBC").
Six blind individuals and four sighted individuals participated in the study. All participants routinely traveled in urban areas; most were familiar with roundabouts; and all were familiarized with each roundabout before making their judgments. Participants stood at roundabout crosswalks and observed traffic continuously during 2-minute periods. Whenever they believed they could complete a crossing before the arrival of the next vehicle at the crosswalk, they pressed a button, and button presses were recorded on a laptop computer. The criterion used to judge whether a gap was acceptable or not was a conservative one that does not rest on the assumption that all approaching vehicles would yield. In addition to the buttons used by participants (typically, an individual who was blind and an individual with typical vision participated at the same time), a third sighted observer used a button to record the arrival of vehicles at the crosswalk. This made it possible to measure the durations of the gaps that occurred at the crosswalk during each 2-minute trial as well as to relate participants’ judgments of appropriate times to cross to the arrival of vehicles at the crosswalk. Participants completed at least six, two-minute trials at an entry and an exit leg of each of the three roundabouts.
The average duration of the gaps between vehicles was 5 seconds at Towson, 7 seconds at Annapolis, and 20 seconds at UMBC. At both Annapolis and Towson, most gaps were so brief that pedestrians walking at 4 feet per second would not have been able to fully cross before the next vehicle arrived at the crosswalk. At UMBC, it was very common to have gaps longer than 10 seconds, and there were frequent periods of "all quiet." Listening for such periods appeared to be an effective strategy for identifying acceptable gaps at UMBC, where blind and sighted participants rarely indicated that it was appropriate to cross when there was not enough time to reach the splitter island before the arrival of the next vehicle. However, it was rarely "all quiet" at Annapolis or Towson, thus requiring participants to judge when vehicles that they could hear and/or see would arrive at the crosswalk. A vehicle traveling at 20 miles per hour travels 29.3 feet per second. Pedestrians typically need at least 5 seconds to cross to or from a splitter island, so to be entirely safe, a pedestrian needs to know about vehicles that are about 150 feet away.
A "risky" crossing judgment was defined as one for which there would not have been sufficient time to reach the splitter island before the arrival of the next vehicle. By this definition, "risky" crossings would require either that all vehicles yield to pedestrians who were crossing the street and/or that the pedestrian were able to monitor the speed and trajectory of nearby vehicles and take evasive action as necessary. A "safe" judgment was one in which there would have been sufficient time to reach the splitter island.
At UMBC, "risky" judgments were few and not significantly different across the two groups. At Towson and Annapolis, however, blind participants averaged more than twice as many "risky" judgments as sighted participants. Thus the degree of risk would have been much greater for the blind participants than the sighted participants had they actually initiated a crossing at the time they indicated that it was appropriate to do so. That is, the blind participants were much more likely than the sighted participants to indicate that there was an acceptable gap when approaching vehicles were only 2 or 3 seconds away. Towson’s exit lanes stood out as particularly problematic for blind participants, with 70% of the judgments falling in the "risky" category. Participants reported having difficulty in this condition due to a combination of the high level of traffic noise at this roundabout and the need to monitor both (a) vehicles on adjacent legs that could quickly reach the exit lane crosswalk and (b) vehicles on the circulatory roadway that might or might not exit.
LATENCY AND DELAY
In addition to investigating the accuracy of participants’ judgments about whether gaps were acceptable or not, the research team also investigated differences in the latency of these judgments. Considering only the acceptable gaps that were correctly detected, blind participants made their judgments approximately 3 seconds later than sighted participants. This finding was consistent at all three roundabouts. This appears to be because vehicles that have just cleared the crosswalk produce engine and tire noise that masks other sounds, interfering with blind pedestrians’ ability to hear whether or not there is another the vehicle approaching the crosswalk. Sighted pedestrians can typically identify the onset of an acceptable gap just as a vehicle clears the crosswalk, but blind pedestrians must let the vehicle pass and wait several seconds for its sound to recede into the distance so that any approaching traffic can be heard
A three-second delay in detecting an acceptable gap probably is of little consequence at low volume, long-gap roundabouts such as UMBC. However, a delay of three seconds would appear to be important at higher volume roundabouts, considering that many of the gaps at such roundabouts may be just long enough to afford safe crossing. At the two higher volume roundabouts studied, delays of three seconds resulted in the inability of blind pedestrians to detect many of the acceptable gaps, or resulted in the detection of gaps too late to make a safe crossing.
These data indicate that there are significant differences in pedestrians’ abilities to determine whether it is safe to initiate a crossing at some roundabouts, depending on whether they are using vision and hearing or hearing alone. These differences have important practical implications for traffic engineers who wish to design roundabouts that are accessible by blind users and to blindness researchers and O&M instructors who wish to find ways to reduce these differences through training.
More work is needed to build upon these early data. Existing pedestrian signals now in place should be analyzed for cost, usability, and accessibility. Additional specialized research and development is necessary to investigate ways of providing useful auditory and tactile cues to roundabout use that can be pilot-tested at roundabouts now in planning or under construction. New roundabouts research should routinely include pedestrian and accessibility considerations in addition to vehicle studies.