Guidelines for the Application of Crossing Solutions at Roundabouts and Channelized Turn Lanes for Pedestrians with Vision Disabilities (2016)

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.

NCHRP 3-78b: Final Project Report April 2016 6 CONCLUSIONS This document represents the final project report for NCHRP 03-78b: Guidelines for the Application of Crossing Solutions at Roundabouts and Channelized Turn Lanes for Pedestrians with Vision Disabilities. The final report is a compendium to the primary deliverable of this research, the Guidebook for the Application of Crossing Solutions at Roundabouts and Channelized Turn Lanes for Pedestrians with Vision Disabilities. The guidebook is geared at providing useful and implementable guidance during the planning and preliminary design stage of a modern roundabout or channelized turn lane site. The guidebook aims to be a resource for engineers and planners developing and/or reviewing plans of designs that are about 30% to completion. Before that stage, designs may be too premature to have the necessary detail on proposed geometry and traffic characteristics; but after that stage, designs quickly become too fixed, including near- final placement of sidewalks, curb cuts, and signal poles. While the developed guidance is useful for existing sites and retrofit treatment installations, the focus is on establishing the accessibility of new sites due to the large variability of existing conditions at existing sites. This final project report provides the documentation and added details for the research that resulted in the guidebook. As such, this report documents the literature review, data collection methodology, analysis results, modeling efforts, and other aspects of the project. 6.1 Implications for Practice Roundabouts are increasingly being adopted by the transportation community in the US, due to their ability to process balanced and unbalanced traffic patterns, their aesthetic appeal, relatively low operating costs, and most importantly, their documented safety benefits (e.g., Rodegerdts et al., 2007; FHWA, 2000; Persaud et al., 2000). Similar to CTLs, there are concerns about the accessibility of roundabouts, particularly for pedestrians who are blind (US Access Board, 2003; American Council of the Blind, 2002). These accessibility challenges have been documented through extensive research by Guth et al., 2005; Ashmead et al., 2005; Schroeder et al., 2010; and Guth et al., 2013, among others. Channelized turn lanes (CTLs) are a common treatment at signalized intersections, intended to allow heavy right-turning movements to bypass the main intersection. Crosswalks at CTLs are often unsignalized in the US, and pedestrians must therefore make crossing decisions based on their perception of an adequate gap or presence of a yielding vehicle. Accessible pedestrian signals (APS) or other audible devices are not available at most CTLs because they are unsignalized, but could be retrofitted and tied to the signal hardware at the main intersection. Accessibility challenges at intersections with CTLs have been documented by Schroeder et al., 2006, and Schroeder et al., 2010, among others. Both intersection types can be associated with substantial benefits for both motorized and non- motorized traffic. CTLs can improve vehicle operations and safety at intersections, by providing added right turn capacity, by separating turning conflicts, or by accommodating larger design vehicles. Some CTL designs can also help pedestrians by shortening crossing distances and simplifying signal phasing, and help bicyclist, by reducing the potential for a “right hook” conflict between a right-turning vehicle and a through movement cyclist. As such, it is in the broader interest of the engineering and planning community to arrive at CTL designs that are accessible to all users, so that general CTL benefits can be achieved without compromising other road users. Similarly, roundabouts have a safety track record with proven reductions in fatal and injury crashes over signalized intersections or two-way stop-controlled intersections. Roundabouts also bring operational benefits, especially in a 24-hour context where they can adapt to fluctuations in traffic volumes and changing traffic patterns. Roundabouts can also be very pedestrian and bicycle friendly, if traffic speeds are slowed down to levels comparable to cyclists, pedestrian crossings are placed in highly-visible and low- 95

NCHRP 3-78b: Final Project Report April 2016 speed locations, and pedestrian decision-making is simplified to dealing with one direction of traffic at a time. Just as with CTL, roundabouts represent an intersection treatment that can enhance safety and serve all road users adequately if designed with all those users in mind. The materials presented in this final report, and more importantly in the companion guidebook, are intended to provide engineers and planners with the tools and resources to consider pedestrians with vision disabilities in the design process, and understand the tools and principles for creating an accessible environment without losing the other benefits of a roundabout or CTL. The guidebook developed in this research is not a prescriptive approach, nor is it a checklist of minimum requirements needed to meet ADA requirements. Rather, it is a principles-based approach that provides engineers and planners with an understanding of why certain characteristics of intersections can pose challenges to certain road users. The guidance is also performance-based, in that it produces measurable metrics and performance measures that can help quantify the accessibility performance, and allow analysts to weigh it against other performance measures. With a principle and performance-based design philosophy, there is no “cookie cutter” solution, as each site brings unique challenges. But this philosophy also provides opportunities to arrive at accessible designs that are uniquely suited for each crossing. The design process is iterative, and a novice designer may take several iterations to arrive at a design that meets all objectives. But with experience, we are certain that the accessibility principles put forth in this research will become second nature, and equally familiar to designers as existing design principles and performance checks in documents like the FHWA Roundabout Guide. Finally, this document and the guidebook are not standalone documents, in that they rely on the established practice and resources in publications like the FHWA Roundabout Guide, the AASHTO Green Book, the MUTCD, the Highway Capacity Manual, and others. This guidance is intended to supplement those resources, rather than replace them. Where applicable, this research borrows established methods and principles, but designers should always refer back to those documents. Orientation and mobility specialists, and the consumers they serve, can be an excellent resource for guidance about the challenges travelers without vision or with impaired vision face at roundabouts, CTL’s, and other complex intersections without traffic controls. Designers, engineers and others in the transportation community are encouraged to use the many resources available to them for ensuring that roundabouts and CTL’s are accessible to all road users, including those with visual impairments. 6.2 Summary of Findings This research collected accessibility data at 12 channelized turn lane locations, and eight approaches to modern roundabouts, which each included separate study of the entry and exit legs. This resulted in a total of 28 crossings that were evaluated for their crossing and wayfinding performance. In addition, this team gathered information in various prior research efforts, including NCHRP Project 03-78a, the Road Commission for Oakland County PHB and RRFB study, and FHWA Task Order project TOPR34 on “Evaluation of RRFBs at multi-lane roundabout approaches”. The final results in this report and the guidance in the companion guidebook document combine results from all these research efforts, to get a comprehensive look at accessibility performance as a function of various design attributes. In total, this and prior research evaluated 16 CTL and 58 roundabout sites. A “site” in this case, refers to one round of study at one approach to an intersection. A roundabout approach typically delivers two sites (one entry and one exit), and a before-and-after study similarly contributes to “sites”. The studied sites were - 16 approaches to channelized turn lanes in five different cities in four states. The CTL studies included various treatments, including raised crosswalks (5 sites), sound strips (2 sites), a stop sign (1 site), and a flashing beacon (1 site). The sites further varied with most having deceleration lanes 96

NCHRP 3-78b: Final Project Report April 2016 (15 sites), and some also having an acceleration lane (9 sites). - 6 single-lane roundabout entries in 6 different cities in 4 states. Treatments tested at single-lane roundabout entries included raised crosswalks (1 site), and rumble strips (1 site). - 6 single-lane roundabout exits in 6 different cities in 4 states. Treatments tested at single-lane roundabout exits included rumble strips (1 site), and RRFBs (1 site). - 20 two-lane roundabout entries in 11 different cities in 11 states. Treatments tested at two-lane roundabout entries included pedestrian hybrid beacons (2 sites), RRFBs (13), raised crosswalks (1 site), and rumble strips (1 site). - 20 two-lane roundabout exits in 11 different cities in 1 state. Treatments tested at two-lane roundabout exits included pedestrian hybrid beacons (2 sites), RRFBs (12), raised crosswalks (2 sites), and rumble strips (1 site). - 1 three-lane roundabout entry and 1 three-lane roundabout exits, each studied three times with various treatments including no treatment (base case), RRFB, and RRFB plus a raised crosswalk. The findings for the sites studied in this research are documented in Appendix D, with previously studied sites being summarized in NCHRP Report 674, the Oakland County Study Report, and the FHWA TOPR 34 Final Project Report on the RRFB Evaluation. Based on the combined findings from the cited research, the following summary observations can be drawn. These findings were first noted in the FHWA TOPR34 final report, but were confirmed in observations in this research: - An increase in the degree of curvature (smaller radii and shorter curves) correlates with a decrease in frequency of interventions. This trend is true for both roundabout entries and exits, as well as CTLs with the strongest correlation for exit legs. Roundabout exit legs, on average, tend to have smaller degrees of curvature (larger radii) than entry legs, and are on average associated with greater frequency of interventions. Predictive models for interventions were developed in this research as a function of vehicle free-flow speed, which in turn is a function of the degree of curvature as discussed below. - An increase in the degree of curvature at the exit can be associated with improved visibility, and can facilitate the placement and visibility requirements for traffic signals, if those are part of the treatment solution for the roundabout. Providing clear line of sight between the pedestrian waiting area and approaching vehicles has further proven to be a very critical criterion for making a site accessible, which can for example be achieved by moving the exit portion of the crosswalk further away from the circulating lane. - An increase in the degree of curvature (smaller radii and shorter curves) correlates with a decrease in the free-flow speed at the crosswalk. Free-flow speeds were generally higher at roundabout exit than at entry, which often correlates with general design and radius between the two. For CTLs, the free-flow speed is similarly correlated with radius. As curvature at entries and exits increase, free-flow speeds are expected to decrease. A predictive model to estimate FFS as a function of radius is found in the FHWA Roundabout Guide, and was found in this research to correlate well with field-measured free-flows speeds at crosswalks for both roundabouts and CTLs. - A decrease in free-flow speed correlates with an increase in the probability that drivers yield to pedestrians. With generally higher free-flow speeds at roundabout exits than entries, the associated yield probabilities are also are lower at exits than entries. For CTLs, higher speeds are similarly linked to lower yielding. A predictive model was developed to predict yielding as the function of speed at the crosswalk, with the presence of an RRFB further increasing the yield probability. From research done for FHWA, a study of the percent interventions suggests that a threshold may exist 97

NCHRP 3-78b: Final Project Report April 2016 at a roundabout entry and exit radius of around 91.4 m (300 ft). Note that this radius is not the same as the central island radius/diameter, but rather the radius that controls the speed at the crosswalk. For roundabout entries, the R1 radius most closely describes this parameter, and at roundabout exits, the R3 term is a reasonable approximation (although speeds for vehicles may be limited by the speed in the roundabout and an acceleration constraint). For CTLs, the radius of curvature of the turn lane itself typically controls the speed at the crosswalk. At entry crosswalks in the FHWA study of two-lane roundabouts, where all approaches had a radius of less than 91.4 m (300 ft), all percent interventions were less than 10 percent, and 9 out of 11 approaches had less than 5 percent intervention. This finding does not imply that all crosswalks with a controlling vehicle path radius of greater than 91.4 m (300 ft) are assured to be less accessible, nor that all crosswalks with a controlling vehicle path radius of less than 91.4 m (300 ft) are assured to be more accessible. But the findings may suggest, that 300 feet is a potential classifying threshold for distinguishing “small” and “large” radii. A 300-foot radius should generally be sufficient to accommodate most design vehicles, and is much larger than turns at traditional signalized intersections. But some multi-lane roundabouts and CTLs were found with radii significantly above this threshold, which is believed to contribute to accessibility challenges (particularly at roundabout exits). From the same research, a threshold is also evident in the relationship between vehicular free-flow speed at the crosswalk and percent intervention. The observed percent interventions changes noticeably at a vehicular free-flow speed of around 35 km/h (22 mph). In this research, free-flow speeds were measured directly in the field with a radar device. For sites with free-flow speed below 35 km/h (22 mph), all but one location had less than 10 percent intervention, and 12 out of 14 had less than 5 percent intervention. For sites with free-flow speeds greater than 35 km/h (22 mph), 5 out of 7 had more than 10 percent intervention, and six out of seven had more than five percent intervention. This finding does not imply that all crosswalks with free-flow vehicular speeds greater than 35 km/h (22 mph) are inaccessible, nor that all crosswalks with free-flow speeds less than this value are accessible. But the findings may suggest, that 22 mph is a potential classifying threshold for distinguishing “low” and “high” speeds. In addition to radius and speed, this research strongly suggests that factors such as ambient noise, pedestrian and crosswalk visibility, and vehicular lane utilization may have a greater influence in some cases than vehicular path radii and free-flow speeds. As such, sites with high ambient noise, poor visibility, and/or highly imbalanced lane utilization may prove to be inaccessible to blind pedestrians, even if vehicle speeds are low and the roundabout geometry features small radii. Nonetheless it appears roundabout designs with vehicle path radii less than 91.4 m (300 ft) and free-flow speeds at the crosswalk less than 35 km/h (22 mph) have a higher likelihood of being accessible than those with higher vehicular path radii and faster free-flow speeds. From observations in this research, it was also evident that it is useful design practice to separate the decisions points for drivers, in terms of interacting with other drivers and pedestrians. At roundabout entries, general design practice places the crosswalk one vehicle length back from the circulating lane to provide that separation. In other words, the crosswalk should be placed where drivers can focus on interacting with pedestrians, and before having to judge gaps in the conflicting vehicle stream. The same has been observed for CTLs where sites with the crosswalk located too close to the downstream yield/merge point showed greater accessibility challenges. The team reasons that for sites with these overlapping decision points drivers tend to focus more on the conflicting traffic stream (to their left), and may not pay attention or react to a pedestrian waiting to cross the street (from the right). FHWA research in TOPR34 using eye-tracker data at roundabouts suggests that drivers tend to look more for gaps in the conflicting vehicle stream, and are less likely to see a pedestrian waiting on the right sidewalk, especially if the crosswalk is close to the circulating lane. For roundabout lanes, a similar trend has been observed, where drivers who are still concerned with navigating the roundabout and the exit, are less likely to yield to pedestrians when the crosswalk is close to 98

NCHRP 3-78b: Final Project Report April 2016 the circulating lane (FHWA TOPR34). This finding was based on observational field studies with pedestrians crossing at roundabouts, and was confirmed in eye-tracker experiments in the same research. The notion of separating driver-to-driver and driver-to-pedestrian decision points is a key design principle gleaned from this research. While none of the trends above should be interpreted as being a direct causal factor, the correlations paint a useful picture of the types of variables associated with increased percent interventions at the studied sites. Conceptually, an increase in the degree of curvature (smaller radius), a reduction in free-flow speed, and an increase in yielding, are all associated with a reduction in percent intervention. The effect of traffic volumes is mixed, with higher traffic volumes at roundabouts being associated with a decrease in interventions at exit legs, but an increase for entry legs, likely due to an increased chance for multiple threat events with more vehicles. For CTLs, no effect of traffic volumes was found conclusively. In addition, other factors such as the ambient noise, visibility, lane utilization, and the local driver culture also appear to be important factors affecting percent interventions. 6.3 Treatment Effects The various accessibility projects that form the basis of this work evaluated a range of treatments that are geared at enhancing accessibility, which fall in four broad categories: 1. Treatments geared at reducing vehicle speeds through geometric modifications, which includes speed humps, raised crosswalk, or geometric changes; 2. Treatments geared at enhancing the visibility of the crosswalk and alerting drivers, which includes RRFBs and other beacons; 3. Treatments geared at providing additional audible information to blind pedestrians, which includes the sound and rumble strips tested at CTLs; and 4. Treatments geared at stopping traffic and creating crossing opportunities, which includes pedestrian hybrid beacons and other pedestrian signals. Some treatments fall into multiple categories, and therefore have combined effects. Sound strips for example may help alert drivers in addition to pedestrians and have a speed-reducing effect. Similarly, beacons and signals have to be outfitted with audible devices, which can enhance the available information to pedestrians, as well as assist with wayfinding. In general, the treatments tested in this and prior research should be evaluated in the context of the accessibility evaluation framework. Rather than being thought of as a “one size fits all” or “cookie cutter” solution, each treatment serves a very specific purpose and should be targeted to address a specific shortcoming of a site. For example, if vehicle speeds are found to be too high (above 22 mi/h as summarized above), a raised crosswalk may be a suitable option to reduce that speed. This by now is the standard design treatment for CTLs in Boulder, CO, for example, and has also proven effective at multiple roundabouts tested in this research. Similarly, beacon treatments may specifically target locations where visibility is poor and drivers may not be aware of the presence of pedestrians wanting to cross. However, as was found in the FHWA RRFB evaluation, a beacon like the RRFB alone is unlikely to greatly enhance accessibility for a site with very high design speeds. In those cases, a combination with a raised crosswalk, or a treatment that stops drivers with a red indication may be more appropriate. All treatment effects are evaluated as part of the evaluation process outlined in the accompanying guidebook document. Their effectiveness is measured in terms of their effect on one or more of the inputs in the accessibility evaluation, which in turn may impact the final outcome and performance. For example, rather than describing a raised crosswalk in terms of its ability to reduce interventions or delay, it is 99

NCHRP 3-78b: Final Project Report April 2016 evaluated first and foremost as a treatment that reduces vehicle speeds. Those reduced speeds in turn result in increased yielding and increased crossing opportunities, which are linked to reduced delay. Speed is also an input in the safety performance, and as such a raised crosswalk is expected to improve delay and risk as a consequence of its speed-reducing effects. Finally, there are a host of other treatments that have not specifically been evaluated in this research, but that are likely to have beneficial effect. In general, considering treatments in light of the four categories and the effects in the overall evaluation process can help designers screen and weigh treatment options. As such, treatments like an advanced yield line at multi-lane approaches to reduce multiple threat events is likely beneficial at roundabouts, based on its effectiveness at midblock locations. Similarly, in-pavement flashing lights may be an alternative to an RRFB, or a speed bump may be considered as an alternative to a raised crosswalk with the same end result. So in addition to treatments tested in this research, designers should look to other literature and resources like the PEDSAFE tool for other ideas on treatments. 6.4 Wayfinding The majority of the roundabout and CTL crossings under study were not found to reliably provide wayfinding information that could be expected to result in routinely safe and efficient crossing by blind pedestrians. While these findings are tentative, the diversity of the roundabouts and CTL’s at which this was the case suggests that this is not something that can be ignored by designers who wish to design accessible intersections. Observations of the techniques and strategies used and the errors made by participants provide information about features and strategies that appear to provide useful wayfinding information. Three principal wayfinding tasks were explored in this research: 1) Determining the Crossing Location, including detecting the street, 2) Aligning to Cross (establishing a correct heading), and 3) Maintaining Correct Heading While Crossing (staying within the crosswalk). The same tasks are necessary tasks for crossing in either direction from the island. Wayfinding on the islands was challenging for blind participants. There are many possible configurations, and design features are not consistent. Islands that are only delineated by pavement markings are not recognized at all by pedestrians who are blind. Typical strategies for locating a crosswalk include continuing to the curb in the direction traveled on the approach sidewalk and using traffic stopped at the stop line as a cue to the crosswalk location. Because of the curvature of the sidewalk at roundabouts and CTLs, and the fact that the crossing point is often to one’s side, even very experienced blind individuals such as many of our participants can find it difficult to recognize the desired crossing point. Features that seemed to help in determining the crossing location (locating the crosswalk) included: (1) grass or other landscape strip between sidewalk and curb, and (2) grass or gravel outside the crosswalk area, particularly on islands. Features that didn’t seem to provide adequate information to pedestrians who were blind in locating the crosswalk included paved or hardscape surfaces. This research did not examine other cues that have been used for aiding nonvisual crosswalk location, such as guidance surfaces, as is used in Australia and other countries, and which pilot research at midblock crossings in Raleigh, NC showed to have a significant benefit to locating crosswalks. More research may provide additional treatments that could assist with this task. Another important aspect of locating the crosswalk is recognizing the street edge. At most locations, detectable warning surfaces were installed to indicate the edge of the street at curb ramps. The detectable warning surface must extend the full width of area that is level with the street to provide an adequate warning. At the roundabout and CTL locations where detectable warnings were not installed or did not 100

NCHRP 3-78b: Final Project Report April 2016 extend the full width of the level area, some research participants continued into the street without recognizing that they had done so. When aligning to cross, blind participants were observed to, and reported themselves to, use a combination of cues, including underfoot surfaces and traffic movement. The alignment, relevant to the direction of travel on the associated crosswalk, of the approach sidewalk, the orientation of the detectable warning surface, the slope of the ramp, and the orientation of gutter/edge of the street are all cues that may be used by individuals who are blind. While these may not be consistent, they all contribute to the alignment decisions. Features that seemed to help in aligning to cross included having all wayfinding cues (approach direction, landscaping or edge of the sidewalk, the curb ramp slope, detectable warning surface, and edge of the street/gutter) aligned with the crosswalk direction or perpendicular to the crosswalk direction. Features that appeared to provide inadequate or confusing alignment information included installing the various features mentioned so that they are not aligned with the crosswalk direction. Without traffic traveling parallel to the crosswalk as they cross, blind pedestrians have few cues to help them maintain the correct heading and stay within the crosswalk. Short crossings allow less opportunity to veer from the beginning line of direction, so shorter distances (i.e. narrow lanes) may be an advantage. Finally, pushbuttons for pedestrian signals, accessible pedestrian signals, RRFBs or other devices can provide very useful wayfinding information for all three wayfinding tasks. However, the devices must be outfitted with an audible output, and must be located close to the crosswalk they control in order to be useful to pedestrians who are blind. 6.5 Limitations and Future Research Several limitations exist in this research, which are summarized below. - Despite a much larger sample size compared to prior research, and especially NCHRP 03-78a, the sample of sites is still limited. Sites used were locations where treatments had been installed by local agencies and there was variability in how those treatments were installed. Most performance measures showed high variability, which made it difficult to clearly isolate effects of treatments and design characteristics. An in-depth study of one intersection type and one treatment at a time (similar to the FHWA RRFB study) is recommended for any future research looking to supplement these results. - Additional testing with a red signal display (PHB or standard signal) and evaluation of more low-cost traffic calming treatments, such as raised crosswalks at two-lane roundabouts, are needed to increase sample size and build confidence in treatment effectiveness. - From the data, it is evident that local context is very important for any accessibility evaluation. Driver culture, the level of expectation of pedestrians, enforcement, and education all appear to be key factors that impact driver behavior, yielding, and ultimately accessibility. As such, any generalization of the study results should be considered carefully within the local context of a site and region of the country. - The analysis process developed in this research needs to be verified at additional sites not part of this study. Ideally, future research would build upon feedback and lessons learned from agencies and practitioners using the guidebook and final report to further enhance future documents. There is a need for training and technology transfer of the research results, to assure the guidance is applied correctly, and to help improve future revisions of these documents. - The current deliverables are presented as standalone documents and not yet integrated with other guidance documents, including the FHWA Roundabout Guide. 101

NCHRP 3-78b: Final Project Report April 2016 - The focus of the guidebook is on the design of new facilities, and not on the execution and field installation of specific treatments. As such, there is a need to train field crews and inspectors to assure that intersections designed in accordance with this guidance are properly constructed. - This report provides only limited information on education and training measures to assist blind travelers in successfully navigating unknown geometries at roundabout and CTL intersections. More work is needed in this area to try and standardize instruction and guidance provided to the travelers themselves. However, it is clear that training cannot resolve the problems of a design that does not provide adequate cues and information to an individual who cannot see. 102