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.
13 Facility Design Current roundabout and CTL design criteria are presented in documents such as the AASHTO Policy on Geometric Design of Highways and Streets (AASHTO 2004); the Inter- section Channelization Design Guide (Neuman 1985); the FHWAâs Manual on Uniform Traffic Control Devices (2009); the AASHTO Guide for Planning, Design, and Operation of Pedestrian Facilities (2004); the FHWAâs Pedestrian Facilities Users Guide (Zegeer et al. 2002); the FHWAâs Signalized Inter- sections: Informational Guide (2004); the FHWAâs Round- abouts: An Informational Guide (2000); an updated version of the FHWA roundabout guide available as NCHRP Report 672: Roundabouts: An Informational Guide, Second Edition (Rodegerdts et al. 2010); and the research results and synthesis to come from NCHRP Project 3-72, âLane Widths, Channelized Right Turns, and Right-Turn Deceleration Lanes in Urban and Suburban Areasâ (Midwest Research Institute 2011), which will be available in 2011. These documents include pro- visions for determining the placement of crosswalks, signage, and other aspects of roundabout and CTL design. A key issue is that existing designs are intended to accommodate the major- ity of pedestrians, who have normal vision. Current designs were not developed specifically to support unassisted cross- ing of streets by pedestrians who are blind. Geometric Design for Pedestrian Crossings at Roundabouts Current practice in the United States (FHWA 2000) locates the pedestrian crosswalk approximately one car length back from the circulating lane, although this varies. The crosswalk is generally perpendicular to the travel lane and passes through an approach splitter island. This island is designed to sepa- rate opposing traffic streams, reduce wrong-way movements around the central island, and provide refuge to pedestrians before they cross the second leg of the approach. The presence of the splitter island serves to divide the pedestrian crossing task into two separate segments. Under low traffic volumes, a pedestrian may be able to cross in a single movement. Under higher traffic volumes, pedestrians may wait on the splitter island until a crossable opportunity is detected on the second leg of the crossing. In either case, the pedestrian crossing task is typically focused to one direction at a time. Figure 1 shows a schematic drawing of typical roundabout crosswalk geometry. The actual alignment of the crosswalk can vary. Often there is no deviation in the orientation of the crosswalk, and the crosswalk proceeds straight from curb to curb. However, some crosswalks are designed with a bend at the splitter island, which may pose wayfinding challenges for blind travelers in the absence of additional tactile cues. There are a few crosswalks that use an offset or zigzag design that deflects pedestrian traf- fic onto an elongated splitter island before the second part of the crossing. The intent of this treatment is to reinforce two- stage crossing behavior and, to some extent, increase the distance between the crosswalk and the circulating lane. For all crosswalks, pedestrian ramps at either curb are sup- posed to be perpendicular to the curb/gutter line. Due to the radius of the curve, they may not be in line with the direction of travel on the crosswalk and therefore may cause alignment difficulties for blind pedestrians. Curb ramps built after 2001 are supposed to have truncated dome detectable warning sur- faces at the bottom of the ramp to alert the pedestrian who is blind that he or she has arrived at a streetâsidewalk boundary. In the United States, few pedestrian crosswalks at round- abouts are signalized (either for pedestrian or traffic control purposes). The Geometric Design for Pedestrian Crossings at Channelized Turn Lanes Channelized turn lanes are much more prevalent in the United States than roundabouts. Despite their increased preva- lence, less attention has been given to the effects of treatments C H A P T E R 2 Synthesis of the Literature
designed to improve pedestrian safety and access. In essence, there are three typical locations for a pedestrian crosswalk associated with a CTL. The crosswalk may be located at the upstream (entering) side of the turn lane, at the downstream (or exiting) side, or at the midpoint, perpendicular to the tan- gent of the curve that defines the turn lane. Figure 2 shows a schematic drawing of a typical CTL with the crosswalk located at the midpoint. The intersection shown has a deceler- ation lane in the approach of the CTL. This feature may not be present in all designs and can vary in length. Alternate designs may also have an acceleration lane at the downstream end of the CTL (FHWA 2004). The midpoint crosswalk location presumably minimizes the crossing distance and is likely to coincide with the loca- tion of slowest vehicle speeds. When vehicles are stopped for pedestrians, this design provides some degree of storage upstream of the crosswalk before traffic in the through lane is affected. Upstream crosswalk locations require pedestrians to dis- criminate between through vehicles and vehicles that intend to turn into the channelized turn lane. This can be difficult, even for sighted pedestrians, since there may be no indication of driver intent until the vehicle is very close to the crosswalk. If it is not possible to discriminate through traffic from turning 14 This figure shows a schematic drawing of a CTL with key features highlighted. The pedestrian crosswalk is placed at the midpoint of the CTL and approximately one car length back from the downstream merge point.The CTL shows a deceleration lane that allows right-turning traffic to slow down for the turn away from through traffic at the intersection. Figure 2. Typical CTL crosswalk geometry (source: FHWA 2004). This figure shows a schematic drawing of a roundabout with key features highlighted. The pedestrian crosswalk is placed approximately one car length back from the circulating lane. The entry- and exit-leg portions of the crosswalk are separated by a raised splitter island to provide pedestrian refuge. Figure 1. Typical roundabout design features and crosswalk geometry (source: FHWA 2000).
vehicles, the blind pedestrian must wait until traffic is stopped or until there is no traffic approaching from either direction on the through street. The downstream crosswalk location creates different issues. At the downstream location, drivers of vehicles in the turn lane are more likely to be looking to their left at vehicles approach- ing on the major street and not to their right where a pedes- trian may be waiting to cross from the curb. Where volumes on the major street are low and gaps required to merge from the turn lane are readily perceived, vehicles may accelerate as they approach the downstream exit, lessening the likeli- hood that they will yield to pedestrians. However, where vol- umes on the downstream departure leg are high, there may be times in the signal cycle when vehicles in the channelized lane are regularly stopped to wait for a gap in traffic. Sighted pedestrians often cross between stopped vehicles at these times, but blind pedestrians may have difficulty determining that vehicles have stopped. Higher speeds and lower likelihoods of yielding to pedestrians are also more likely when an accel- eration lane is provided at the CTL exit. Facilities built since 2001 compliant with the American with Disabilities Act (ADA) include a curb ramp with trun- cated-dome detectable warnings that delineate the edge of the roadway. If the crosswalk and ramp are located upstream or downstream within the CTL, the ramp may terminate into the radius of the curve, which is not perpendicular to the crosswalk. Blind pedestrians may experience problems with identifying the crossing location and alignment at all cross- walks at channelized turn lanes because of the curvature. Accessibility Challenges Recent research on the crossing performance of people who are blind at complex intersections demonstrates that there are unique challenges for this population (Ashmead et al. 2005, Guth et al. 2005). Complex intersections, including round- abouts and channelized turn lanes, present some unique chal- lenges for pedestrians with vision impairments. The traffic control strategy at a roundabout entry leg is typically a yield sign, and many drivers are able to enter the circle without the requirement to come to a full stop. Similarly, traffic exiting the roundabout is free-flowing (often accelerating), resulting in largely uninterrupted traffic flow at the exit portion of the crosswalk. Traffic patterns at CTLs are similar in that the right- turning movement is largely free-flowing. Crosswalks at both types of facilities are typically not signalized, and the task of identifying crossing opportunities is thus unassisted. Depend- ing on the geometric design and the location of the crosswalk, vehicle speeds may be relatively high, and the auditory inter- pretation is complicated because vehicles are moving on a cir- cular path (Ashmead et al. 2005). At signalized intersections, the two traffic streams typically move perpendicularly to each other, presumably making it easier for someone who is blind to interpret directional traffic movements. Finally, the contin- uous flow of traffic circulating the roundabout can create a difficult auditory environment, and the listening task is com- plicated by elevated levels of ambient noise. All of these factors may contribute to accessibility challenges at roundabouts and CTLs. Research in this area and the work of this project have divided the crossing task into four princi- pal areas that guide the understanding of the challenges faced by pedestrians who are blind. Additionally, the four areas require different treatments to improve targeted aspects of pedestrian accessibility. The four crossing components are: 1. Locating the crosswalk, 2. Aligning to cross, 3. Identifying a crossing opportunity, and 4. Maintaining alignment during crossing. In the following, each component is discussed in detail. Locating the Crosswalk A pedestrian approaching an intersection needs to be able to identify the location of the crosswalk. For standard orthogonal intersections, this is a fairly simple task since the curb ramps are located in the vicinity of the corner of the intersection, usu- ally within 15 ft if the turning radius is not significant. How- ever, at roundabouts and CTLs, there is no distinct point of interest (such as the corner) where a pedestrian would expect a crosswalk to be located. Instead, the auditory environment from nonlinear traffic patterns can be difficult to navigate com- pared to the typical orthogonal intersection. For roundabouts, finding the crosswalk is highly depend- ent on the direction of travel and the side of the roundabout being approached. For instance, a pedestrian wishing to turn left with traffic approaching the roundabout from behind her would have the crosswalk on her left, similar to a standard orthogonal intersection. Even so, the circulating traffic does not provide a consistent point of reference to begin looking for the crossing location. The task of walking straight through a roundabout intersection poses further challenges. The pedes- trian must be aware that she must navigate around the circle via sidewalk cues or landscaping and then locate the crossing location a significant distance from where the typical crossing would be located at a standard intersection. Since vehicles do not move in a perpendicular fashion with other vehicles, the pedestrian may need repeated attempts to identify the crossing location, which can be time-consuming and dangerous. For similar reasons, CTLs can be challenging to a blind pedestrian. The unusual geometry associated with the large 15
turning radius does not follow the pattern of a typical inter- section. Making the locating task even more difficult is the placement of the crosswalk, which requires crossing an uncon- trolled movement at the turn lane. With non-standardized placement of the crosswalk at the upstream end, midpoint, or downstream end, it is difficult to provide blind travelers consistent guidance. Aligning to Cross Once the crosswalk has been found, the proper alignment should be determined so that the pedestrian does not start the crossing with an unsuccessful alignment. For instance, many times curb cuts at signalized intersections are installed as a single curb cut in the middle of the curb radius to eliminate the (perceived) need for two perpendicular curb cuts, saving time and costs. This single curb-cut installation method sig- nificantly hinders the ability of a blind pedestrian to align correctly. Alignment is an important task that if not done correctly, could lead to problems when maintaining that align- ment during crossing. For blind pedestrians, typical aids used for aligning are specialized tactile surfaces or the edge of curb cuts (Barlow et al. 2005). Other research efforts underway at the time of this report [National Institutes of Health (NIH) 2010] are looking at alternative treatments to further aid pedes- trians in aligning under various conditions, primarily focused on improving detectable warning surfaces using parallel and perpendicular bar tiles and far-side beacons. At the time of this report, no findings were available on the success or failure of such devices in aligning to cross. Identifying a Crossing Opportunity The focal point of this research effort was on measures describing the third component, the task of deciding when it is actually safe to cross conflicting traffic. At unsignalized cross- ings, which constitute the large majority of roundabouts and CTLs, the two crossing opportunities available to the pedes- trian are yielding drivers or large gaps in traffic. At signalized crossings, the pedestrian walk phase presents a planned cross- ing opportunity that is a function of signal phasing, which can be very useful to a blind pedestrian when negotiating very dif- ficult geometries and traffic conditions such as those posed by large two-lane roundabouts or CTL facilities. However, the walk phase does not ensure that traffic is stopped due to right turns and permissive left-turning movements. Therefore, the walk phase by itself does not identify a safe time to cross. Maintaining Alignment During Crossing Finally, the fourth component is the task of maintaining alignment during the crossing, which is greatly facilitated by perpendicular crosswalk geometry. This fourth component depends on the provision of proper aligning geometry (the second component, discussed above) and assistive devices so that the appropriate heading can be properly determined and maintained. In other words, maintaining alignment is highly unlikely if a pedestrian starts the crossing in the unintended direction. Far-side locator beacons or other treatments can be extremely helpful in properly aligning and maintaining alignment. Under a separate grant (NIH 2010), research is underway to test alternative alignment devices such as raised guiding surfaces in the pavement area, far-side audible bea- cons, and remote infrared audible signals. No definitive analy- sis of such treatments has been completed at this time on the ability of these devices to successfully aid pedestrians in main- taining alignment. The U.S. Access Board and ADA One of the responsibilities of the U.S. Access Board is to develop design guidelines for transportation facilities, ensuring that public rights-of-way are accessible to and usable by all people and are thereby in compliance with the ADA. The Access Board published the draft Public Rights-of-Way Acces- sibility Guidelines (PROWAG, U.S. Access Board 2005), out- lining requirements for making crosswalks and intersections in the public rights-of-way compliant with the ADA. Even before the draft PROWAG was published, the requirements for accessibility were outlined in the implementing regula- tions of Title II of the ADA, which specifies that any newly constructed or altered public facility shall be âreadily accessi- ble to and usable by individuals with disabilitiesâ (DOJ 1990), including those with vision loss, mobility impairments, or other disabilities. The draft PROWAG provides more specific guidance, outlining features that make a site compliant with the ADA. Specifically, the provision of a pedestrian-actuated signal with APS at two-lane roundabout approaches is dis- cussed as making the site usable by pedestrians who are blind. A pedestrian-actuated and APS-equipped signal thereby satisfies the accessibility requirement for two-lane round- about approaches (PROWAG R305.6.2). However, the draft PROWAG continues to allow the use of alternative treatments if justified. The draft PROWAG language for two-lane CTL crosswalks is very similar to two-lane roundabouts in that a pedestrian signal with APS satisfies the accessibility require- ment. In addition, the draft PROWAG also specifies the pro- vision of landscaping or barriers to delineate the crossing locations at roundabouts and CTLs, the use of APS devices at all signalized pedestrian crossings, and the provision of detectable warning surfaces on the curb ramp to demark the streetâsidewalk boundary (PROWAG R305.6.1). The draft PROWAG does not discuss signalization at single-lane round- abouts or single-lane CTLs. 16
Pedestrian Signals Pedestrian crossing control devices can be grouped into two types: pedestrian displays and vehicle displays. Pedes- trian displays in the United States typically feature a âWalkâ and a âDonât Walkâ phase, which operate in either flashing or solid state depending on the phase. Other countries use green and red walking figure symbols instead of the text description. Vehicle displays are further divided into typical green/yellow/red signals that are more commonly used at roadway intersections for vehicular traffic control, and the newer pedestrian hybrid signals (e.g., PHBs) that feature a modified arrangement of vehicle signal displays. The United States and other countries have warrants (FHWA 2009) for the installation of pedestrian signals, but the installation of other treatments is less standardized. The pedestrian signal warrants in the United States were initially intended for midblock loca- tions or conventional intersections that would otherwise be stop controlled. A warrant is generally intended to define when a treatment (in this case a signal) is justified. Warrants are not requirements to place signals. Midblock Crossings and Conventional Intersections In the United States, the Manual of Uniform Traffic Control Devices (MUTCD) provides warrants for when traffic sig- nals may be installed (FHWA 2009). There are a total of nine warrants, one of which deals with pedestrians. In the 2009 MUTCD, Warrant 4 (Pedestrian Volume) states that a traf- fic signal âat an intersection or midblock crossing shall be considered if . . . one of the following criteria is met: ⢠For each of any 4 hours of an average day, the plotted points representing the vehicles per hour on the major street (total of both approaches) and the corresponding pedestrians per hour crossing the major street (total of all crossings) all fall above the curve in [Figure 3], or ⢠For 1 hour (any four consecutive 15-minute periods) of an average day, the plotted point representing the vehicles per hour on the major street (total of both approaches) and the corresponding pedestrians per hour crossing the major street (total of all crossings) falls above the curve in [Figure 4].â Alternatives for the two conditions above are also given for speed limits (posted, statutory, or 85th percentile) that exceed 35 mph or intersections that lie within an isolated commu- nity with a population of less than 10,000. The pedestrian vol- ume signal warrant is not to be applied at locations where the distance to the nearest traffic control signal or stop sign is less than 300 ft unless the proposed traffic signal will not restrict the progressive movement of traffic (FHWA 2009). If one or both of the conditions is met for the pedestrian sig- nal installation, further guidance on placement of the pedes- trian traffic signal is provided: ⢠Traffic signals installed at traditional intersections or major driveways should control the minor street or driveway traf- fic through actuated means and should include pedestrian detection. ⢠If the pedestrian signal is installed at a non-intersection location, it should be installed no closer than 30 m (100 ft) from the nearest side street or driveway controlled by stop or yield signs and should also be pedestrian actuated. Rec- ommendations are also provided for adequate sight dis- tance, including removing obstructions within 100 ft of the signal heads on each direction of the approach and equip- ping with proper signage and pavement markings. ⢠If the pedestrian signal is installed within a signal system, the traffic control signal should be coordinated appropriately. 17 Figure 3. Warrant 4, pedestrian four-hour volume (source: FHWA 2009).
Last, traffic signals may be justified outside the bounds of the aforementioned MUTCD guidance if the 15th-percentile crossing speeds are less than the assumed 3.5 ft/s at all inter- sections. The MUTCD states that the pedestrian volume crossing the major street may be reduced as much as 50%. In looking at international literature, a variety of pedestrian- related signal warrants from the United Kingdom, Canada, and Australia were identified. In Canada, the national MUTCD and an Ontario traffic manual provide pedestrian-based warrants for the installation of full traffic signals (City of Hamilton 2005). Intersection pedestrian signals, or half signals, are also used in some locations in Canada. Warrants and guidelines for these devices largely reside at the local level (or provincial level in the case of British Columbia) and include measures such as vehicular and pedestrian volumes, pedestrian delay, avail- able gaps, roadway geometry, sight distance, speed, pedestrian demographics, and distance to the nearest adjacent signal. In the United Kingdom, pedestrian traffic signals at round- abouts are fairly common. The warrant for such a signal is based on the formula PV2 > 108, where P = pedestrian volume per hour (average of peak 4 hours) and V = vehicle volume per hour (average of peak 4 hours). Both sides of the splitter island must satisfy the criteria separately in order to meet the warrant (entry and exit volumes from roundabouts must be considered separately) (Baranowski 2004). In Australia, some states have pedestrian-based warrants for full signals at intersections and midblock signals (Queensland 2003, New South Wales 2008). These warrants are similar to U.S. warrants in that they are primarily based on traffic and pedestrian volumes, with additional emphasis placed on avail- able crash information. Additional wording is added to include warrants for signals that âcater mainly to persons with partic- ular disabilitiesâ including disabled, aged, or hearing-impaired pedestrians. Specialized pedestrian signal applications such as the PHB [HAWK (high-intensity activated crosswalk)], Pelican, Puffin, and Toucan crossings (Fitzpatrick et al. 2006) have all been designed for midblock crossings. Installation of these devices at roundabouts presents a unique set of challenges. In many cases, such an installation would create conditions that are not recommended by the MUTCD. ⢠According to Section 4D.01 of the 2009 MUTCD, âMidblock crosswalks should not be signalized if they are located within 30 m (100 ft) from side streets or driveways that are con- trolled by stop signs or yield signsâ (FHWA 2009). While not directly applicable to roundabouts, this suggests that prox- imity between a pedestrian signal and sign-controlled move- ment may cause confusion. ⢠According to Section 4D.13 of the 2009 MUTCD, signal heads should be located no closer than 40 ft and no farther than 180 ft from the stop bar, except where the width of an intersecting roadway or other conditions makes it physi- cally impractical (FHWA 2009). The minimum distance is necessary to ensure that the signal head is visible through the windshield of a car. At a typical roundabout, the cross- walk is only 25 ft from the circulatory roadway. This would result in the stop bar being placed in the circulatory road- way to achieve proper signal head placement on the exit leg. ⢠Section 4D.12 recommends that âat signalized midblock crosswalks, at least one of the signal faces should be over the traveled way for each approach.â In this case, mast arms would be needed for signalized crosswalks at roundabouts, which would increase cost. In addition to the above sections, there is a general need to ensure that the signal indications on the roundabout entry do not create confusion with the yield sign at the circulating lane. 18 Figure 4. Warrant 4, pedestrian peak hour (source: FHWA 2009).
For any signal installation, the use of APSs is essential to ensure that the signal is accessible to pedestrians who are blind. APS installations commonly feature a push-button locator tone to help pedestrians find the pedestrian push button, and an audi- ble signal or message that alerts the pedestrian when the walk phase is shown on the pedestrian signal display. Audible sig- nals include variations of nonverbal sounds (cuckoo, chirp, or click) or a verbal message saying, for example, âWalk sign is on to cross [street name].â APS installations are a requirement for all new or altered pedestrian signal installations in the draft PROWAG by the U.S. Access Board (2005). Although they are common in the United Kingdom and other countries, pedestrian crossing signals at roundabouts are rare in the United States. If signals are used at roundabouts, they are often installed as a two-stage crossing to reduce vehi- cle delays. In that case, a zigzag crossing geometry prevents the pedestrian from inadvertently crossing the entire roadway without sufficient crossing time. Currently, pedestrian signals of varying types have been installed in the United States at only a handful of roundabout intersections, including Salt Lake City, Utah; Charlotte, North Carolina; Clearwater Beach, Florida; Alpine City, Utah; and Oakland County, Michigan. Further information can be found in Appendix B. Blind Pedestrian Crossing Experiments In studies at roundabouts completed with support from a Bioengineering Research Program funded by the National Eye Institute (NIH 2010), researchers sought to document the street crossing behavior of people with total blindness at single- and multiple-lane roundabouts. In the initial study, conducted in 2000, adults who were totally blind and adults who were sighted made judgments at three Baltimore-area roundabouts about whether gaps in vehicular traffic were crossable, or were long enough to permit crossing to the splitter island before the arrival of the next vehicle without assuming any vehicular yielding. Trials were conducted at both exit-lane and entry-lane crosswalks of single-lane and two-lane roundabouts. Overall, blind participants were about 2.5 times less likely to make correct judgments than sighted participants, took significantly longer to detect crossable gaps, and were more likely to miss crossable gaps altogether. Fur- ther, the errors of blind participants were much more likely to be high risk than the errors of sighted participants at the two roundabouts that carried moderate and high volumes of traffic, in contrast to data collected at the lower volume roundabout. This research is reported in Guth et al. (2005). This research team also investigated judgments of gaps at a single-lane roundabout in Tampa, Florida, with large and predictable variations in traffic volume over the course of the day (Long et al. 2002). Blind participants made more high-risk judgments during peak hours than during off-peak hours. A similar pattern was not found for sighted participants. The judgments of the blind participants improved, although they remained risky, in a condition in which they made judgments at rush hour at simulated downstream crosswalk locations 60 ft from the actual crosswalks. The work in Baltimore and Tampa involved making judg- ments about crossing without actually crossing. This left open the possibility that participants were using different judg- ment criteria than would have been the case had they actually crossed. To address this possibility as well as to follow up on several differences found in the earlier studies, the researchers conducted a third study in Nashville at a high-volume, two- lane roundabout. Blind and sighted participants made judg- ments without actually crossing during half of their trials and crossed during the other half. The Nashville findings validated the judgment-only measure, confirmed earlier findings, and provided important new data about pedestrianâdriver inter- action (Ashmead et al. 2005, Guth et al. 2005). A new measure used in the Nashville study was the frequency of interventions by an O&M instructor who followed the participants during crossing trials. Although interventions occurred in only a small percentage of trials (6%), the authors calculated a 99% prob- ability of a serious pedestrianâvehicle conflict at this inter- section if a person who was blind crossed daily for 3 months. A conflict was defined as a situation in which a crash is likely unless the driver or pedestrian takes immediate evasive action and is synonymous with events that resulted in an O&M intervention. Also, in post-experiment questionnaires, most participants who were blind reported that they would not cross at this intersection if they had any other option. An alternative to crossing in a gap in vehicular traffic is to cross the street in front of a vehicle that has yielded upstream of the crosswalk. Geruschat and Hassan (2005) investigated the likelihood that drivers would yield to individuals holding a white cane and standing at a roundabout crosswalk, and contrasted these yielding data to data associated with indi- viduals without white canes standing in the same location. This study, completed in Annapolis, MD, revealed that yield- ing rates overall were low, and varied as a function of vehi- cle speeds. Lower speeds were associated with higher yielding rates. The presence of the long cane resulted in only a modest increase in yielding rates. Also, when drivers yielded to a blind pedestrian, the pedestrian was often unable to detect the pres- ence of the stopped (yielding) vehicle and would subsequently often fail to take the crossing opportunity provided. Yielding rates were higher at roundabout entry lanes than exit lanes. Inman, Davis, and Sauerburger (2005) reported that the mean observed time before vehicles yielded in both lanes was 63 s. Inman, Davis, and Sauerburger (2005) reported the results of two studies related to the effectiveness of pavement treat- ments. Their first study, conducted on a closed course, was to 19
evaluate the feasibility of a pavement treatment designed to alert blind pedestrians when vehicles have yielded to them. The second study examined driversâ yielding behavior at a two-lane roundabout, along with an evaluation of the effec- tiveness of the roadway treatment identical to that used on the closed course study. In the first study, there were two experi- mental conditions: a control condition and a treatment con- dition in which devices similar to rumble strips were placed on the roadway surface. Seven individuals with severe visual impairments participated. Participants stood at a crosswalk and used hand signals to indicate when they detected vehicles stopping or departing after a stop. Compared to the control condition, the sound-strips treatment increased the probabil- ity of detecting stopped vehicles and decreased by more than a second the amount of time needed to make a detection; however, the treatment did not reduce the number of false detections. The authors noted that false detections could result in the pedestrian crossing when moving vehicles are approaching the crosswalk. The second study was an experi- ment conducted at a double-lane roundabout. In that envi- ronment the rumble-strip-like treatment was not effective in increasing detection of yielded vehicles. The authors attributed this to the fact that the majority of vehicles stopped before crossing over the rumble strips. A âYield to Pedestrians â State Lawâ sign that was placed in between the two travel lanes resulted in an increase in driversâ yielding from 11% of vehi- cles in the control condition to 16% in the experimental con- dition. It was concluded that the treatments explored in these studies do not appear promising for double-lane roundabouts but should be explored further to see if they might work at single-lane crossings. The Long List of Pedestrian Crossing Treatments Many different types of pedestrian treatments are available to aid engineers in designing safe crosswalks. Although there is limited guidance for choosing when a certain treatment should be implemented, there are resources that can assist in making good judgments. For instance, the Pedestrian Safety Guide and Countermeasure Selection System (Zegeer et al. 2002) is an Internet site dedicated to providing practitioners with up- to-date information on engineering, education, and enforce- ment to improve pedestrian safety and mobility. Another important source of information is TCRP Report 112/NCHRP Report 562: Improving Pedestrian Safety at Unsignalized Cross- ings (Fitzpatrick et al. 2006). This resource contains a detailed overview of pedestrian crossing treatments at unsignalized intersections and midblock locations and presents field study results on their effects on driver yielding behavior. The treatment selection process for NCHRP Project 3-78A started with a long list of treatments that were believed to have potential in improving specific aspects of pedestrian accessibility. In the base condition, it is assumed that the sites meet current design and basic accessibility standards, includ- ing appropriate curb ramps, detectable warnings, and marked crosswalks. Under these baseline conditions, it is recognized that sighted pedestrians likely have better yield- and gap- detection capabilities and that delay and risk are higher for blind pedestrians. However, any treatment tested is hypoth- esized to also improve crossing conditions for sighted pedes- trians, and for that matter any other special pedestrian popu- lations, including children, wheelchair users, and the elderly. The research team identified 28 candidate treatments from various literature sources that showed initial potential to improve (blind) pedestrian accessibility by improving gap and yield utilization, minimizing risk, and reducing delay dur- ing the crossing task. In this research, treatments are grouped into six basic categories: 1. Driver information treatments, 2. Traffic calming treatments, 3. Pedestrian information treatments, 4. Crosswalk geometry modification, 5. Signalization treatments with APS, and 6. Grade-separated crossings. Each category is intended to categorize treatments based on their principal intended effect on vehicle operations. They are discussed in more detail below, and Appendix B provides more detail as well as photographs of most of these treatments. Driver Information Treatments There is evidence (Fitzpatrick et al. 2006) that the use of static pedestrian crossing signs that are uncorrelated with actual pedestrian presence is unlikely to generate predictably high levels of driver yielding. This is not to say that all signing is ineffective or that signing is not required. However, several improvements over static roadside warning signs are possible and are summarized below. ⢠Continuous flasher: A continuous flashing beacon can be added to any static sign to make it more visible. The continuous flasher is a static device in that it will continue to flash whether a pedestrian is actually at the crosswalk or not. This type of treatment can become ineffective, espe- cially if the available pedestrian traffic is not sufficient to provide feedback to the traveling public that the crosswalk is actually used on a frequent basis. ⢠In-roadway warning sign: In-street âYield to Pedestriansâ signs are placed in the roadway between travel lanes to increase the visibility of the crosswalk. The signs typically post messages such as âState Law â Yield to Pedestrians.â 20
The signs give a third dimension to the usual crosswalk striping. Speed reductions associated with slight increases in driver compliance are expected with this type of treatment. ⢠Active-when-present flasher: This treatment looks similar to the continuous flasher; however, it is operated dynami- cally by activation of a pedestrian push button or by passive pedestrian detection. The dynamic push button activated beacon serves to increase the conspicuity of the static pedes- trian sign. The treatment typically takes the form of a flash- ing beacon at the roadside, mounted overhead, or imbed- ded in the pavement. Traffic Calming Treatments Traffic calming is a method of designing streets using visual or physical cues to encourage drivers to reduce speeds. Traffic calming is largely self-enforcing in that the design of the road- way should result in the desired outcome of speed reduction. Traffic calming can be a very effective tool for reducing the severity and frequency of crashes and even noise levels. In addition, studies suggest that drivers are more likely to yield to pedestrians when traveling at slower speeds (Geruschat and Hassan 2005). Two possible treatment alternatives aimed at reducing vehicle speeds are described below. ⢠Posting lower speed (15 and 25 mph): This treatment considers posting regulatory reduced speed limit signs at 15 mph or 25 mph. This treatment represents the lowest- cost traffic calming treatment; however, the desired out- come is highly dependent on compliance of the driver. If the design of the roundabout does not encourage slower speeds (i.e., the deflection is not properly designed), then expected compliance of the posted lower speed is small unless heavily enforced. A lower regulatory speed limit for a CTL is impractical since it would also apply to the main line. ⢠Raised crosswalk: A raised crosswalk will reduce vehicle speeds as a function of its height relative to pavement sur- face and the transitional slope. A low and a gently sloping raised crosswalk would likely have higher speeds since vehi- cles easily maneuver over the crosswalk. Likewise, a steep incline to a high raised crosswalk could result in significant speed reductions; however, the reduced lane capacity may outweigh the benefit of the reduction in speed. Raised cross- walks also introduce vertical obstructions for ambulances and snow plows that need to be considered. Pedestrian Information Treatments This category includes treatments that provide pedestri- ans with audible information that can be used to make more informed decisions about when to safely cross using available yields and/or gaps. It should be noted that some treatments in this functional category have not been fully developed at this time but were still considered as a possibility as the team developed the research plan. The four possible treatment categories are: ⢠Surface alterations/rumble strips: Roadway surface alter- ations, such as rumble strips, generate auditory cues of approaching and/or yielding vehicles (Inman, Davis, and Sauerburger 2005). The treatment can also have the added benefit of providing information on the availability of cross- able gaps. As an added benefit, the driver may be more cau- tious when approaching the crosswalk due to the additional sound cue provided by the treatment. ⢠Yield-detection system: The use of in-road sensors or video image processing to detect whether vehicles have yielded (stopped or slowly rolling) has shown promise in initial tests completed under a related NIH grant (NIH 2010). An audi- tory signal provides a speech message to the pedestrian indicating when a vehicle has yielded. The functional prob- lems of such a system are primarily based on reliability of detecting vehicles that roll very slowly, queued vehicles stopped over the crosswalk (at the entry for instance), and in the event of a yielded vehicle that begins moving again. ⢠Gap-detection system: It is possible to use in-road sensors or video image processing to detect if there is an approach- ing vehicle (or no vehicle) within some predetermined safe crossing time or distance from the crosswalk. As with yield detection, the use of an auditory signal via an audible device is imperative to provide a speech message to the pedestrian indicating when it is safe to cross. The ability to sufficiently or accurately detect such gaps at roundabouts (especially the exit approach) and CTLs is not known at this time but is under development (NIH 2010). ⢠Yield- and gap-detection system: This treatment would combine the two previous treatments to take advantage of the yield- and gap-detection capability that could ultimately be possible. It is not known at this time whether such a sys- tem is even plausible since there has been no development of gap detection for pedestrian crossing treatments com- pleted at this time. Crosswalk Geometric Modification There is the possibility of a modified crosswalk location or an alternative crossing location at roundabouts. This approach would place all or parts of the crosswalk further away from the circulating lane to separate pedestrianâvehicle interaction from vehicleâvehicle interaction at the roundabout. Supple- mental treatments such as static signing, pedestrian-activated signs, and traffic calming techniques can all be applied in the distal crosswalk situation to further enhance accessibility. 21
Four variations of the concept of a relocated crosswalk are presented: ⢠Distal crosswalk: This treatment would relocate the cross- walk to a distance of approximately 100 ft from the circu- lating lane of the roundabout. The (presumed) benefit is the lower level of ambient noise at the crosswalk that is associ- ated with moving the crosswalk further from the circulatory roadway. Driver benefits include reduced queue spillback issues in the roundabout with added storage capacity for the exit lane(s). Drawbacks of this treatment include poten- tially longer pedestrian walking distances, depending on the originâdestination patterns at the site. An additional drawback is that sighted pedestrians may ignore the distal crosswalk and cross closer to the roundabout unless phys- ically restricted from doing so. ⢠Traffic calming at distal location: The distal crosswalk can be combined with other treatments to provide some traf- fic calming measures to reduce speeds, increase the likeli- hood of drivers yielding, and reduce the risk of collisions. Potential treatments include lowering regulatory speeds and the installation of a raised crosswalk. ⢠Median island at distal location: The distal crossing location would no longer have the benefit of a pedestrian refuge island since the roadway at that point is most likely undivided. Therefore, a distal crossing would require a one-stage cross- ing of both directions of vehicular traffic. A median island would provide pedestrian refuge and re-establish a two- stage crossing. ⢠Offset exit crossing: The potential effectiveness of this treatment rests on the premise that pedestrians (in partic- ular, blind pedestrians) experience more difficulty crossing exit lanes than entry lanes. By offsetting the exit-lane por- tion of the crosswalk and creating a zigzag crossing, gap selection ability may be facilitated if ambient noise levels are in fact reduced relative to the typical crosswalk loca- tion. The zigzag configuration would further maintain and even enforce a two-stage crossing strategy and would pro- vide supplemental queue storage at the exit lane. The crosswalk modification treatments primarily apply to roundabout crossings. Some special considerations for geo- metric design at CTLs include: ⢠Add deceleration lanes: The use of deceleration lanes for traffic using the CTL has several potential advantages: (1) if vehicles, in fact, slow down in the deceleration lane, slower vehicle speeds can increase the likelihood of drivers yielding to pedestrians, and (2) when used in conjunction with some type of audible surface treatment, such a cue may facili- tate crossing decision-making. ⢠Remove acceleration lanes: While facilitating the move- ment of traffic exiting the CTL, acceleration lanes are often associated with higher vehicle speeds. Higher vehicle speeds are associated with a decreased likelihood of drivers yield- ing and an increased injury rate in the event of a collision. Signalization Treatments with APS Signals at roundabouts and channelized turn lanes rep- resent a more costly and intrusive treatment for providing a safe crossing environment for pedestrians. Traffic signals may introduce delays to both pedestrians and vehicles. Addi- tionally, depending on signal timing and placements, vehi- cle queues can spill back on the roundabout exit from the signal to affect roundabout circulating flow or CTL through movements. However, APS-equipped signals can be effective at stopping traffic and at providing the pedestrian with visual and auditory cues of when the crossing phase is active. CTL signal impacts can be reduced through coordination with phasing at the main intersection and to avoid the likelihood of queue spillbacks onto the through lanes. Pedestrian sig- nals with a walk indication can and should be outfitted with an APS to provide auditory cues in addition to the visual sig- nal display. ⢠Pedestrian scramble phase: This signal strategy stops all vehicular traffic at the roundabout intersection to allow pedestrian movements in any and all directions (along marked crosswalks). Pedestrian activation at any approach of the facility would (following some minimum green time for vehicles) produce a red signal at all entry lanes. Follow- ing a clearance interval designed to allow all vehicles in the circulatory lane to exit the roundabout, a pedestrian walk signal would be presented to all pedestrians waiting to cross. This treatment alternative enables pedestrians to cross in a single stage. Following the pedestrian walk phase, vehicles at all entry lanes would be given a green signal to proceed. While simple in concept and in operation, the effectiveness of such a signalization strategy has yet to be determined. This strategy likely has little application to CTLs. ⢠Pedestrian-actuated traditional signal â one or two stage: This treatment uses a traditional traffic signal for pedestri- ans at (typically) unsignalized locations such as a round- about or CTL. The signals are standard red-yellow-green traffic signal heads that rest in green when no push-button activations are in place. The treatment is particularly useful for blind pedestrians because the signal provides auditory information about phase indication via APS, much like they are accustomed to from a conventional intersection. In areas with high traffic and/or pedestrian volumes, delay and queue spillback at roundabouts could be problematic, 22
especially with false (unused) pedestrian actuations. Also, because the signals rest in green the majority of the time, it is possible that drivers may react slowly (or not at all) to the red stop indication. ⢠Pedestrian hybrid beacon â one or two stage: The PHB or HAWK signal aims to be more efficient than a conven- tional signal by allowing vehicular traffic to move during the pedestrian âFlashing Donât Walkâ phase. During that phase a flashing red indication for drivers allows traffic to proceed after stopping if no pedestrian is in the crosswalk. This phas- ing scheme allows for less vehicular delay while providing similar pedestrian-related benefits to a regular signal. ⢠Distal pedestrian-actuated signal â one or two stage: Entry-lane and exit-lane pedestrian-activated signals used at a distal crosswalk location or in a zigzag configuration could be used to establish a one- or two-phased pedestrian crossing that maximizes the storage capacity of the exit lane during a vehicle red phase. If a two-phase crossing is used, a median refuge island would be necessary. Depend- ing on pedestrian route patterns, these configurations may result in an increase in the travel time for pedestrians com- pared to a crossing at the traditional splitter island. ⢠Distal/zigzag PHB â one or two stage: The PHB could also be used at a distal location or in a zigzag arrangement, com- bining the advantage of the extra queue storage capacity at the exiting approach of the roundabout with more efficient signal phasing. Depending on pedestrian route patterns, these configurations may result in an increase in the travel time for pedestrians compared to a crossing at the tradi- tional splitter island. The location of the distal crosswalk requires a median refuge island to be utilized if a two-stage crossing is necessary. Grade-Separated Crossing Grade separation allows pedestrians to cross the road with- out affecting the movement of vehicles. Grade-separated facil- ities must accommodate all persons, including those with vision and mobility impairments. To accommodate all users, these treatments may require ramps or elevators. Grade sep- aration is typically used in cases where pedestrians must cross very busy streets or freeways, and where pedestrian volumes are extraordinarily high. Grade-separated facilities should not be considered where opportunities for crossing at the street level are available on a regular basis because this discourages use of the facility. ⢠Pedestrian overpass: Overpasses are the more common form of grade separation in the United States and are acces- sible via stairs, ramps, or elevators. Overpasses should be designed so that they provide the ability for multiple users to pass by or around each other. ⢠Pedestrian underpass: Pedestrian underpasses are most easily accounted for and installed during the design and construction process. Underpasses installed as a retrofit require costly underground construction. Underpasses may be difficult to keep clean and safe, but with proper design and lighting these challenges can be overcome. 23
