CHAPTER 7

Conclusions and Suggested Research

At the outset of this project in 2019, the authors identified six main research gaps for what was not known about TWSIs. Since then, through this project and parallel research efforts, progress has been made in closing several of these gaps, and new questions have been raised. This chapter walks through each of the six initial questions to share what has been learned about each, what still remains, and what further research is needed to continue to improve wayfinding for people with vision disabilities.

Discriminable Geometry of Raised Elements

What gap spacing between raised elements in relation to top width of raised elements is acceptable for TWSI discriminability? Japanese research suggests smaller raised elements more widely spaced are most detectable and discriminable (NITE 1998; NITE 2000; Sawai, Takato, and Tauchi 1998). Given the emphasis on center-to-center spacing between raised elements, gap spacing between the edges of the tops of raised elements was not mentioned in previous reports. The 2010 ADA standards permit a broad range of gap spacing between truncated domes in relation to their top surface widths for DWSs, not all of which had previously been tested and demonstrated to be reliably detectable. Due to the variability in dome top diameter and center-to-center spacing of DWSs currently used in the United States, gap spacing ranges from 0.7 to 2.0 in. (17.6 to 49.5 mm).

Parallel research from Bentzen, Scott, and Myers (2020) compared the detectability and identifiability of a narrowly spaced DWS (1.22 in./31 mm gap spacing; 0.45 in./11.4 mm top diameter) to that of wider flat-topped TDI bars (2.1 in./53.3 mm gap spacing; 0.90 in./22.9 mm bar top width) and narrow TDI bars (1.95 in./49.5 mm gap spacing; 0.45 in./11.4 mm bar top width). They found that all the surfaces were detected more than 94% of the time. Participants were more successful in identifying the flat-topped TDI bars (87.9% on average) than the DWSs or narrower TDI bars (83.1% and 80.6%, respectively), suggesting the smaller gap spacing of the latter two surfaces may impact discriminability from one another.

Research conducted through this project similarly found that the four TWSI surfaces tested in Experiment 1 were highly detectable (95% of the time). Likewise, TDIs were identified as “bars” and DWSs were identified as “domes” between 74 and 79% of the time after only a few seconds of foot contact, depending on the test surface, with no significant difference in performance.

Based on the combined results from these two studies, the geometries of DWSs currently allowed in the 2010 ADA standards, including the narrower gap spacing dimensions, are identifiable and discriminable from TDIs. Further, while Experiment 1 found no significant difference between TDI-1 and TDI-2 for identifiability, Bentzen, Scott, and Myers (2020) did find their flat-topped TDIs (similar in geometry to this project’s TDI-1) were identified accurately at a

significantly higher rate. This finding, along with feedback from participants in Experiment 1, leads to the conclusion that TDIs with a gap spacing of approximately 2.1 in. (53 to 54 mm) are discriminable from other TWSIs. Further consideration of the effects of such geometries on those using mobility aids is warranted. Bentzen, Scott, Emerson, and Barlow (2020) investigated the effects of crossing TDIs at various angles for people using a variety of mobility aids, but they tested only a single TDI geometry (this project’s TDI-1). Changing geometries to narrower raised elements of the same height, and thus increasing the vertical edges, may result in different effects on people using mobility aids. The success of blind participants detecting and using both TDIs and DWSs in the trials conducted in Experiments 2 and 3 supports that DWS-2 and TDI-1 were discriminable.

CPIs

Are CPIs needed, and if so where and what surface type should be used? Experiment 2 resolved this question with results that showed participants correctly turned or went straight through path intersections of TDIs as instructed significantly better when a CPI was used compared to when there was none and the TDI paths simply joined. This held true across a variety of conditions but was particularly pronounced at more challenging T-intersections, where participants went the correct direction only 51% of the time when starting along the TDI at the top of the T and on the side of the path away from the leg or stem of the T. Using a 3 × 3 ft area (0.9 × 0.9 m) of DWS or blank-space CPI to mark path T-intersections resulted in participants going the correct direction over 90% of the time. Additionally, in Experiment 2 and the field experiment, the tasks provided to participants only ever required them to manage a single choice point on a given trial. In real-world contexts, pedestrians with vision impairments may need to follow a path and a set of directions that involve navigating through two or more path intersections. Detecting each choice point is relevant to the likelihood of success in reaching the desired destination, or reaching it efficiently.

Given that there was no significant difference in performance between the DWS or blank-space CPI, it is suggested that agencies use the blank space. An obvious benefit to the blank-space CPI is that it will be less expensive to use because it means no TWSI material to purchase, install, or maintain. A second reason to use the blank-space CPI is to preserve the currently limited usage of DWSs and their meaning in the United States as a warning for potential hazards. The successful application of blank CPIs in Experiment 3 is somewhat muddied by the particular task where participants were asked to follow the path (with a 45-degree left turn in it) to the choice point and then make a left. Where the TDI went straight to a CPI on the upper plaza level, 88% of participants correctly turned at the CPI; where it turned 45 degrees before reaching the CPI, only 65% of participants correctly turned at the CPI. This may reflect the experimental design and not informing participants of the path’s nonintersection turn, rather than the CPI going undetected.

Following Nonintersection Turns

How can people with vision disabilities better follow nonintersection turns? Questions remain regarding nonintersection turns in TDI paths. In Experiment 2, participants were familiarized with 45-degree and 90-degree nonintersection turns in paths, but they were not reminded that paths may turn during the trials. They went the correct direction following 45-degree path turns on average 85% of the time and 90-degree path turns 75% of the time. In Experiment 3, attempts to glean anything meaningful about performance at a 90-degree turn were lost when other questions arose around how to ensure pedestrians with vision disabilities can get onto the starting point of a TDI guidance path, given where the turn was laid out in relation to the start of the TDI

path at the bottom of the ramp from the transit station where the walkway transitioned to the upper plaza. The tested 45-degree path turn on the lower plaza was made successfully by 100% of the participants, though 25% of them spent more than 10 seconds exploring the area around the turn before continuing along the TDI path. Again, this may be a result of the experimental design, where they were asked to make a left turn at the CPI, which may have caused confusion with the left-turning 45-degree angle in the path.

The limited evidence of success for people with vision disabilities navigating nonintersection turns calls for further research.

Angled vs. Curving Turns

Are angled turns in paths easier/harder to follow than curving turns? What are the installation and maintenance considerations for each? People, including blind people, tend to walk the shortest path from point A to point B. This results in people naturally walking in curved arcs to avoid obstacles when navigating their route. Anecdotally, though, agencies tend to avoid installing curving TDI paths, opting to zigzag them instead through several 90-degree or other obtuse angles to avoid making special cuts in surfaces with inset tiles or to avoid cutting the TWSI panels themselves. Maintenance concerns also arise when using prefabricated TWSI tiles or panels in a curvilinear path layout. For example, the City of Charlotte found that cutting their DWS mats to fit the narrow angle needed to install them at bicycle ramps for roundabouts resulted in the DWSs prematurely peeling back on that corner. Methyl methacrylate (MMA) is a possible material to use where topography may lend to more horizontal and vertical curving, but no research was found to understand how pedestrians with vision impairments follow along curved routes. In fact, only Seattle was found to be testing the use of MMA material so they could lay out curved TDI guide paths. Unfortunately, the consistency in applying the appropriate amount of MMA material to maintain the required height profile of the TWSI is problematic, and Savill et al. (1997) noted that thermoplastic for TWD tended to slump and lose its height profile over time.

Future Research

Points to investigate in future research on nonintersection turns in TDI guide paths include:

• Is there a maximum turning angle (e.g., range of degrees) in a TDI path before travelers cannot efficiently and successfully continue to follow the path through a turn?
• Would the use of CPIs at a nonintersection turn improve performance?
• How are turns understood by people with vision disabilities, and what messaging effectively explains the potential to encounter nonintersection turns in TDI paths through route instructions?
• As travelers become aware of and experienced in following TDI paths with turns, does that impact performance at these turns?

Delineating Sidewalk-Level Bike Lanes

What TWSI is appropriate to delineate sidewalk-level separated bicycle lanes in the United States? Unfortunately, this project was unable to identify a suitable site in Charlotte to investigate this line of research through Experiment 3. Therefore, what is known currently is based on the parallel research conducted by Bentzen, Scott, and Myers (2020). In that research, using a high-polymer concrete TWD cast in place in a thick mortar base and measuring 6.33 in. (160.8 mm) across the top, 10.08 in. (256.0 mm) across the base of the trapezoid, and 0.75 in. (19.1 mm) high, blind participants detected the TWD with their cane and/or foot in an average of 99% of

individual trials. Identification of the TWD as a trapezoid based on short-duration foot contact also averaged 99% of individual trials. That research remains the only known experimental evaluation of a TWD, and it offers great promise regarding the application of such a material to create an accessible boundary cue to delineate pedestrian-exclusive spaces from other spaces.

While the TWD was not tested as the boundary element between the pedestrian travel way and the bicycle travel way, it was tested as a “do not cross” barrier between the walkway and the railway in Experiment 3. In this research, it did not perform up to a desirable standard for a surface meant to serve as a warning not to be crossed; 23% of participants who contacted the TWD ultimately crossed or began to cross it. As explained in the results section in Chapter 6, there are some reasons its exact application in the research environment was not as ecologically valid as desired, but all the same, the performance was a bit lacking. Further research on the effectiveness of TWDs in real-world settings (e.g., delineating a bicycle lane at sidewalk level from the sidewalk) is needed.

TWSIs Indoors vs. Outdoors

Should there be different TWSI height specifications for indoor vs. outdoor use? This research gap was identified as part of the literature review process but was outside the scope of this project to directly study. Most research and TWSI standards internationally clearly support that raised elements should be at least 0.2 in. (5 mm) high to be detectable—a specification predicated on the assumption that many accessible outdoor walking surfaces may not be entirely smooth, such as walkways paved with tile, brick, or other smaller-unit patterns that may create small grooves at each joint line or allow for heaving or sinking of individual pavers. While such pavements may still technically meet Public Right-of-Way Accessibility Guidelines (PROWAG) requirements (36 CFR Part 1190; https://www.govinfo.gov/content/pkg/FR-2023-08-08/pdf/2023-16149.pdf) as “stable, firm, and slip resistant” with level changes of no more than 0.25 in. (6.4 mm), these minor elevation changes along a pedestrian access route (PAR) can make it more difficult for pedestrians with vision disabilities to discern natural variations in the pavement surface from the TWSIs. This was anecdotally observed when conducting Experiment 3 across the various walkway surfaces the TWSI test surfaces were installed on. Participants appeared to navigate along the rougher surfaces more slowly while attempting to decipher the pavement from the TWSI surface.

Other research has suggested TWSIs as short as 0.1 in. (2.5 mm) may still be detectable, but the results are mixed as far as what minimum height threshold could be considered; it likely depends on the smoothness of the adjacent surface (Nakamura et al. 2008; Nakamura et al. 2009; Nakamura et al. 2011). This has implications when considering the use of TWSIs inside buildings such as transit stations or terminals, where the adjacent flooring may be more consistently smooth. In such circumstances, shorter TWSIs may be preferred to minimize impacts to people with mobility impairments. Indeed, German practice illustrates clear distinctions of height for indoor vs. outdoor TWSI application.

TWSIs Used Together

How can different TWSIs be used together as a system for wayfinding? This project attempted to address this research gap, which was a primary aim of Experiment 2 in a controlled environment and Experiment 3 in a natural environment. By and large, participants had no difficulty following 12-in.-wide (304.8 mm) TDI guide bar paths with the bars oriented parallel to the direction of travel and stopping when they reached a DWS terminus. In Experiment 3 specifically, participants never approached a platform edge without detecting its presence, never crossed the DWS at the rail crossing before announcing they had arrived at the crossing location,

and only once did a participant step past the DWS on a curb ramp and into the street prior to establishing and announcing they were at the crossing start position.

TDI Path Beginnings and Endings

How should a TDI guidance path begin and end? The field experiment revealed that, while participants could follow TDI guide paths well, they struggled with finding the start of a path and then correctly establishing their initial heading. Research is needed to determine considerations or contexts for when TDI guide paths can simply start/end vs. when an endpoint indicator or locator TDI may be needed to ensure people with vision disabilities contact a TWSI without much searching, minimizing their likelihood of getting disoriented before finding the path and improving their likelihood of establishing the correct heading along the path.

Locator TDI Lengths

How long should locator TDIs (transit door location bars or sidewalk alert bars) be in relation to the walkway, platform, or sidewalk width? Locator or alignment TDIs were also used in a variety of ways based on different natural environmental contexts within the variety of tasks set up through the field experiment. These tasks and TDI applications tested usage at light rail station boarding platforms, a median island streetcar boarding platform, pedestrian rail crossings, and pedestrian street crossings. In each of these applications the TDI was oriented with the bars perpendicular to the direction of travel to board or cross and was 24 in. wide.

Where the TDI was used to locate a noncorner street or rail crossing (e.g., sidewalk alert bars) or to locate where transit doors would open (e.g., transit door location bars), the research showed participants were more successful in contacting the TDI (and therefore detecting and using it) on their first pass down the walkway, sidewalk, or platform when the TDI extended across much or all of the width of the walkway, sidewalk, or platform. Further research may help to tease out a range or minimum/maximum length for locator TDIs, which may be proportionate to the walkway width. For example, while the transit door location bars were each 3 ft long (0.9 m) for three tasks, participants were most successful in contacting them on their first pass down the narrow (approximately 6 ft/1.8 m of effective width) streetcar boarding platform (100%) compared to the wider light rail boarding platform (59%) or the sidewalk at the bus stop (38%). Varying lengths of sidewalk alert bars were not tested in relation to sidewalk widths in Experiment 3 but were installed to extend across the full width of the sidewalk at two midblock crossings, a decision that was based on previous research where the same installation configuration was used and tested (Bentzen et al. 2017; Bentzen et al. 2022).

Improving Use of Alignment Squares through Practice

Does the use of TDI alignment squares improve through practice? When considering the use of TDIs for alignment specifically, the field experiment showed that participants who used the TDI to align to board or cross were successful in their alignment at least 86% of the time. This reinforces the use of TDIs for alignment found in previous research (when TDIs were present and participants attempted to use them, 77.3% aligned correctly in Bentzen et al. 2017 and 73.4% aligned correctly in Bentzen et al. 2022). Where the 2 × 2 ft (0.6 × 0.6 m) alignment square of TDI was used at the skewed street crossing in the field experiment, participants did not seem to have difficulty using the square so much as finding it to use. Thirteen of 17 participants successfully contacted and moved onto the alignment square to prepare to cross. The 4 who struggled seemed to have difficulty finding the square or possibly distinguishing it from the adjacent DWS. It is possible that ease of finding and using the novel alignment square could improve with practice and experience, which could be an area of research to explore further.

Effectiveness of TWSIs with Guide Dogs

What considerations impact the performance or use of TWSIs as a system for people using guide dogs as their travel aid? Due to budget constraints, this project was unable to robustly evaluate the successful completion of tasks laid out in the field experiment with people who use guide dogs. That said, several participants from each experiment were guide dog users, and they were asked for qualitative feedback and opinions on how different tasks might or might not work if they were using their dog rather than a long cane.

For Experiment 2, one participant thought dogs could be trained to follow TDI guide paths and specifically felt that guide dog users should use the bars on their right. Participants in the natural environment experiment did not give any thoughts specific to using guide dogs while following TDI guide paths; however, they did provide feedback on using locator TDIs and alignment squares.

One participant thought finding the bars to locate a crossing would be more difficult using a dog. Two participants said they would use an alignment square and found it helpful to have an initial heading, with one explaining, “alignment is my job.” Another felt they could use the alignment square to orient and the dog could then follow the established heading, but eventually the dog was likely to switch to its own heading/alignment. Several participants thought the dogs could be trained to target TDIs. A couple of participants conveyed that while they would use their dog to guide them to an open transit door, they still thought the locator TDI would be helpful as a confirmation for where to wait.

Indeed, two participants in the field experiment also tried some of the tasks with their dog after completing their session using a long cane. The informal assessments and interviews with these two guide dog users hint at the possibility that locator TDIs (transit door location bars, sidewalk alert bars) running the width of a walkway would be detected underfoot on approach. Further research is needed, given people’s variability in walking speeds, gait, and stride length, to determine how consistently successful that would be and to ensure guide dog users do not step over the TDIs without contacting and/or detecting them.

Interestingly, one of the participants began training his dog to target the locator TDI made of yellow plastic polymer panels while walking through the course. The dog did not slow or seem to identify the next locator TDI encountered that used the rubber mat material (bars were primarily black or blue on this TDI), however, when approaching the third locator TDI also made of the rubber mat panel with yellow bars, the dog slowed on approach suggesting it had accurately identified the TDI. This experience warrants further exploration of whether the type or color of material may be more readily targeted by guide dogs, but it anecdotally suggests dogs can be patterned.

Durability

What is the durability of different TWSI materials, how should they be maintained, and what operational impacts should be considered when installed in different contexts and weather environments? Poorly maintained or badly installed TWSIs can lead to confusion or harm to those who rely on them. The material used for TWSIs, how they are installed (surface applied vs. embedded), and how they are maintained may impact their detectability and discriminability, and therefore their use. Evaluating these differences was beyond the scope of this project, but insights may be gleaned from previous research (see Appendix A). For example, information currently known about durability performance and maintenance needs for certain materials is largely based on testing done on DWSs and should be generally transferrable to the same considerations and uses of these materials for TDIs. Further research is needed, however, to discern any

concerns about maintenance of raised-bar TDIs, which may have different orientations depending on their application, and trapezoidal TWDs that may require different operational techniques or knowledge to ensure their continued effectiveness as a TWSI over time. In fact, Elliot et al. (2017) pointed out that this may be particularly true in shared-street environments where TWSIs could be driven over by motor vehicles or plowed when clearing snow. Agencies interviewed when compiling the state of practice in the United States (see Appendix B) indicate that practitioners may still have questions around maintenance issues pertaining to rainwater management and drainage, as well as power washing or sweeping to clear out debris in order to ensure TWSIs remain detectable. As the use of TDIs expands, other questions are likely to arise around how best to maintain them. Additionally, only the trapezoidal TWD is novel in the United States, and only one vendor is currently known that produces it, so many questions remain about what materials are most appropriate for durability, conspicuity, and ease of maintenance.