Evaluation of Pedestrian Hybrid Beacons and Rapid Flashing Beacons

CHAPTER 7. SUMMARY/CONCLUSIONS, DISCUSSION, AND FUTURE RESEARCH NEEDS

OVERVIEW

This chapter provides summaries and conclusions along with a discussion of the implications of the findings or each of the studies. The chapter concludes with a list of future research needs.

CLOSED-COURSE STUDY

Summary/Conclusions

The closed-course study was designed to quantify drivers' ability to detect pedestrians within and around a crosswalk (a measure of disability glare) and quantify discomfort glare ratings associated with LEDs in traffic control devices. Participants drove the study vehicle to the starting location where they parked the vehicle 200 ft from sign assemblies that consisted of a pedestrian crossing sign with LEDs within the sign face and LEDs in rectangular beacons above and below the sign. After the driver placed the vehicle into park, they were asked to set occlusion glasses on their faces. The occlusion glasses obscured the participants' vision by going opaque when there was no power supplied to them and going clear when power was supplied.

Once the participants' vision was occluded, technicians placed a static cutout photo of a pedestrian (either 54 inches tall to represent a child or 70 inches tall to represent an adult) within the crosswalk located near the sign assemblies. An experimenter then restored the participants' vision, and the participants were asked to identify the direction the pedestrian was traveling (to the left, to the right, or not present) as quickly as possible using a button box. When the participants pressed the button on the button box, the glasses turned opaque again. Following the participants' identification of the pedestrian's direction, the researcher asked the participants to rate the intensity of the LED (comfortable, irritating, or unbearable) before asking the field crew to set up the next condition. This process was repeated for various combinations of LED brightness, LED locations, pedestrian positions, and flash patterns. This portion of the study was stationary, and after completion, the participants drove to the check-in location and completed a laptop survey that asked a series of questions to obtain the participants' opinions regarding flash patterns for LEDs used with signs.

To increase the number of flash patterns tested in the study but to keep within a reasonable testing period, data were collected within two sets. Within each set, two flash patterns were tested for the LEDs in rectangular beacons and two flash patterns for the LEDs within sign. For set I, the study was conducted during both daytime and nighttime. For set II, the study was only conducted during the nighttime. During the testing of set I, it was determined that night was the more critical condition, which is why only nighttime data were collected during set II.

A summary of the findings for LED intensity and location along with flash pattern is provided in table 76. Table 77 summarizes the findings for pedestrian height, pedestrian position, and participant age. Following is an overview of the key findings from this research study.

Note: Bold text indicates the results were statistically significant.

Note: Bold text indicates the results were statistically significant.

Average nighttime detection time for the participants to search and determine which direction a cutout pedestrian was walking was 1.473 and 1.292 s for older and younger participants, respectively. Average daytime detection time for the participants was, as expected, faster (1.281 and 0.971 s for older and younger participants, respectively, during the day).

LED intensity had a measurable adverse impact on detection time at night but not during the day. Under nighttime conditions, detection time increased 8.5 percent for 2,000 candelas compared to 0 candelas (no LEDs). Similar to detection time, LED intensity adversely affected accuracy at night but not during the day. Regarding discomfort glare, LED intensity had an adverse impact under both daytime and nighttime conditions.

LED location affected nighttime detection times but had no detectable daytime effect. At night, detection time was 6 percent longer for LEDs below compared to LEDs within (or 12 percent longer for LEDs below compared to LEDs above). Likewise, detection times with LEDs within were 6 percent longer than for LEDs above. Discomfort glare was different by LED position at night with higher discomfort level with LEDs below compared to LEDs above.

Flash pattern affected detection times during both nighttime and daytime conditions. During the day, only the 2-5 flash pattern had a significantly larger detection time (5.2 percent longer) than no flash pattern. At night, both the 2-5 and wig-wag flash patterns were found to delay detection compared to no pattern (increases of 6.0 and 13.7 percent, respectively). For accuracy and discomfort glare, no significant differences among flash patterns were found under both daytime and nighttime conditions.

Pedestrian height impacted detection time both day and night. During the day, detection time for the short pedestrian increased by 3.9 percent during the day and by 3.6 percent at night compared to detection time for the tall pedestrian. Similarly, accuracy was higher when the experiment involved the tall pedestrian instead of the short pedestrian.

Pedestrian position had an impact on detection time, both during the day and at night. Under both conditions, detection was faster when the pedestrian was located at the center of the crosswalk. Also, for both light conditions, detection times for pedestrian at left or at right were not statistically different from each other. Accuracy trends by pedestrian position were similar to detection time trends, though only at night were these trends statistically significant. Pedestrian position was found to influence discomfort glare at night with higher discomfort when searching for the pedestrian at either side of the crosswalk as compared to when the pedestrian was at the center.

Age of participants drew a clear gradient of increasing detection times, both during the daytime and nighttime. Accuracy of detection decreased by age, both during the day and at night.

The survey found that multiple flashes within a short time period were better at communicating the need to stop for a pedestrian at a crosswalk as compared to few or no flashes such as the wig-wag or no LED illuminated conditions.

The survey also found that when observing close-up views of a sign assembly consisting of a pedestrian crossing sign and LEDs either embedded or below the sign, the patterns that used multiple pulses communicated greater urgency in needing to yield to a pedestrian. The participants indicated that LEDs below communicated more urgency than the LEDs within.

When asked to count the number of pulses in a light bar with the 2-5 flash pattern, the majority of the participants (77 percent) correctly counted two pulses; however, almost none of the participants correctly counted the five faster pulses. Only four participants provided the correct answer. The majority of the participants (55 percent) saw three pulses when five pulses were presented.

Discussion

The flash pattern along with the brightness of LEDs, whether used within beacons or embedded in a sign, can help draw drivers' attention to a device and the area around the device. However, characteristics of the LEDs, such as brightness or flash pattern, can also make it more difficult for drivers to see objects around a device (disability glare) or result in drivers looking away from a device (discomfort glare). This study used several measures to gain an understanding of how brightness and flash pattern affect driver's ability to detect a pedestrian within a crosswalk. These measures included time to correctly identify pedestrian walking direction and participant's rating of discomfort glare.

The brightness intensity of the LEDs used in this study ranged from 0 candelas (i.e., the LEDs were not on) to 2,200 candelas. In another FHWA study, devices installed in the field were measured with higher brightness intensity; the range used in this closed-course study did not reflect the wider range currently being used in on-road installations.⁽⁵⁾ The brightness of LEDs in the field appears to be highly variable. Part of the reason could be that current requirements only specify a minimum intensity.The minimum intensity is defined within SAE Standard J595; the minimum measured at a horizontal angle of 0 degrees and vertical angle of 0 degrees for class I yellow peak luminous intensity is 600 candelas.⁽¹⁵⁾

For this study, brightness intensity did not have a significant impact on detection time for daytime conditions while being significant for nighttime conditions. Nighttime detection time increased by 8.5 percent at 2,200 candelas (the maximum used in the study) as compared to when the LEDs were off. The brighter the LEDs, the longer it took for the participants to determine which direction the pedestrian was facing. In other words, lower brightness was associated with reduced disability glare.

Some of the flash patterns used with the devices were associated with longer detection times. Of the six flash patterns tested, only two flash patterns—the 2-5 and the wig-wag flash patterns—were associated with statistically significant longer detection times as compared to the no flash pattern condition. Both of these patterns had longer on times (the 2-5 flash patter was on 69 percent of the cycle, and the wig-wag pattern was on 100 percent of the cycle) as compared to the other patterns (range of 10 to 38 percent on time). The LEDs being constantly on may have caused the participants to look away from the LEDs. In addition, the lack of sufficient dark period(s) between the flashes may have limited the participants' ability to adequately search for the pedestrian. A better flash pattern than the current 2-5 flash pattern should retain multiple pulses (since the survey results found that participants felt patterns with multiple pulses were associated with greater urgency), more dark periods (since the study found longer detection time for patterns with less dark periods), and a maximum intensity that limits discomfort when attempting to detect objects while still commanding driver attention (i.e., resulting in high driver yielding).

The findings for pedestrian position and LED location indicate that the distance between the pedestrian and the light source affected the ability to quickly detect the pedestrian. When the pedestrians were located at the edge of the crosswalk (i.e., next to the assembly) and when the LEDs were located below the sign (i.e., closer to the pedestrian), detection time was longer and detection accuracy was lower. These findings support the idea of placing the LEDs above rather than below the sign and investigating the benefits of locating the LEDs over the roadway rather than on the roadside.

The shorter height pedestrian required more time to detect and had lower detection accuracy, which were expected findings. The smaller target provided by a child-sized pedestrian was a known concern for pedestrian crosswalks.

This study found strong evidence that there was potential value in mounting the LEDs above the sign instead of below. Nighttime detection time was fastest when LEDs were above the signs, with a 6 percent increase when LEDs were within the sign and a 12 percent increase when LEDs were below the sign. Both of these findings were statistically significant. This finding supports the idea that separating the pedestrian from the light source may benefit the driver's ability to search and identify the location of the pedestrian.

ABOVE-BELOW (OPEN-ROAD) STUDY

Summary/Conclusion

Based on the findings from the closed-course study, the following combination was examined in open-road settings: beacons located above the sign as compared to when the beacons were located below the sign.

The RRFB in positions above and below the pedestrian crossing sign were installed at 13 sites located in 4 States (Aurora, IL; Douglas County, CO; Marshall, TX; and Phoenix, AZ). At all 13 sites, after collecting data for the initial beacon position, the beacons were moved to the opposite position. The same flash pattern was used regardless of beacon position. The research team used a staged pedestrian protocol to collect driver yielding data to ensure that oncoming drivers received a consistent presentation of approaching pedestrians.

Because a previous study that included RRFBs found that posted speed limit, crossing distance, and city influenced driver yielding, the initial analyses were conducted with those variables along with the beacon shape variable.⁽³⁵⁾ An indicator variable for nighttime conditions was included in the final model to determine if the nighttime results were significantly different from daytime results. From the preliminary review of the findings (average daytime yielding was 64 percent when the beacons were above the sign and 60 percent when the beacons were below the sign), it appears that there were only minor, if any, differences between the above and below positions. The results from the GLMM indicate that there were no significant differences between the two positions (p-value = 0.1611).

In conclusion, the position of the yellow RRFB did not have an impact on whether a driver decided to yield to the waiting pedestrians. Variables that did have an impact on driver yielding include posted speed limit, intersection configuration, and city (yielding was lower in Illinois compared to the other States included in study).

Discussion

With respect to the location of the LEDs, the findings from the closed-course study for pedestrian position and LED location indicate that the distance between the pedestrian and the light source affected the ability to quickly detect the pedestrian. When the pedestrians were located at the edge of the crosswalk (i.e., next to the assembly) and when the LEDs were located below the sign (i.e., closer to the pedestrian), detection time was longer. These findings support the idea of placing the LEDs above rather than below the sign.

The open-road study found that the position of the RRFB (either above or below the sign) did not affect a driver's decision to yield. With the apparent benefits identified from the closed-course study (i.e., lower discomfort and improved ability to detect the pedestrian, as measured by identifying the direction the cutout pedestrian is traveling) and the finding that there was no difference in driver yielding due to position, locating the beacons above the sign could improve the overall effectiveness of this treatment. Based on these findings, FHWA is considering issuing an official interpretation to permit the placing of the beacons above the sign.

FLASH PATTERN (OPEN-ROAD) STUDY

Summary/Conclusions

When IA-11 was issued in July 2008 for the RRFB, the only flash pattern that had been tested was the 2-5 flash pattern in which the beacon pulses two times on one side followed by five faster pulses on the other side.⁽⁴⁾ However, because the 2-5 flash pattern appears to the human eye to be a 2-3 flash pattern, IA-11 specified a 2-3 flash pattern and, up until official interpretation 4(09)-21 (I), many devices were installed with the 2-3 flash pattern rather than the 2-5 flash pattern.⁽⁹⁾ The inability to accurately determine the number of pulses within a pattern was later confirmed in the closed-course study (see chapter 3). The same closed-course study found that certain flash patterns—those that could be characterized as having limited or no dark periods within the flash pattern—negatively influenced the amount of time participants needed to identify the direction a pedestrian is walking. Prior to developing the proposed provisions for incorporating a rapid-flashing beacon traffic control device into the MUTCD, it is important to determine which flash patterns are acceptable from the perspectives of effectiveness and simplicity.⁽¹⁾ There is a desire to know if a less complicated flash pattern or a flash pattern with different dark/light proportions would be equally or more effective than the 2-5 or 2-3 flash patterns.

An open-road study was conducted to examine different flash patterns with yellow RRFBs. The objective of the study was to determine if the use of simpler flash patterns used with RRFBs resulted in different driver yielding rates at uncontrolled crosswalks. The MOE was the number of drivers who did and did not yield for a staged pedestrian who activated the RRFBs and was attempting to cross the roadway. The study included eight sites located in either College Station, TX, or Garland, TX. Most of the sites (seven out of eight) had four lanes with a 40- or 45-mi/h posted speed limit. The remaining site had two lanes and a 30-mi/h posted speed limit.

A temporary light bar and controller were developed to permit the research team to have control over several of the beacons characteristics, such as flash pattern and brightness. The light bar was designed such that it was not obvious that the beacons being observed were any different from the permanent RRFB light bar they were mounted to. A remote control was used to activate the temporary light bar.

A flash pattern workshop along with meetings with FHWA resulted in the selection of the following patterns for testing:

• Pattern using a combination of long and short flashes (blocks).
  
  
• Pattern using a combination of WW+S flashes.
  
  
• The 2-5 flash pattern.

The research team used a staged pedestrian approach to evaluate driver yielding for the different patterns. Each staged pedestrian wore similar clothing (gray t-shirt, blue jeans, and gray tennis shoes) and followed specific instructions in crossing the roadway. A second researcher, who observed and recorded the yielding data on pre-printed datasheets, accompanied the staged pedestrian. Data were collected in February and March 2014.

The average driver yielding was 80 percent for the WW+S flash pattern, 80 percent for the blocks pattern, and 78 percent for the 2-5 flash pattern. While there was a small numeric difference of 2 percent, the statistical analysis found that this difference was not statistically significant. Logistic regression was used to model the yielding and not yielding relationships for the individual crossings. The results from the GLMM indicate that there were no significant differences between the tested flash patterns. The WW+S flash and block patterns developed as part of this research study were as effective as the 2-5 flash pattern.

Discussion

This study, combined with the closed-course study that found drivers were better at judging pedestrian direction when there were more dark periods (see chapter 3), suggest an advantage in using a flash pattern with a longer dark period during night time conditions and that this advantage was not offset by a reduction in driver yielding during the daytime conditions. This suggests the profession should consider using a flash pattern with increased dark periods when specifiying the pattern for RRFBs.

The findings from the research effort were presented to the NCUTCD STC during its June 2014 meeting. The STC recommended that the WW+S flash pattern should be used with future rapid-flashing pedestrian treatments. Based on the findings from this research, FHWA issued an official interpretation on July 25, 2014, to permit agencies to use either the previously approved 2-5 flash pattern or the optional WW+S flash pattern.⁽²⁾ Although both flash patterns are available for use, the official interpretation mentions that FHWA favors the WW+S flash pattern because it has a greater percentage of dark time when both beacons of the RRFB are off and because the beacons are on for less total time. The greater percentage of dark time is important because this will make it easier for drivers to read the sign and to see the waiting pedestrian, especially under nighttime conditions. The less total on time will make the RRFB more energy efficient, which is important since they are usually powered by solar energy.

PHB STUDY

Summary/Conclusions

The PHB has shown great potential in improving safety and driver yielding; however, questions have been asked regarding actual driver and pedestrian behaviors. A total of 20 locations in Tucson, AZ, and Austin, TX, were selected for inclusion in this study representing a range of posted speed limit, median type, and number of major roadway lanes. Data were collected using a multiple video camera setup. The final dataset reflected over 78 h of video data and included 1,979 pedestrians.

The videos were reviewed to identify each occurrence when a vehicle stopped at the crossing when the PHB was displaying a dark indication. None of the drivers who stopped at the crossing when the PHB was dark appeared to be confused regarding the device. In the cases when a queue was present during the flashing red indication, about half of the crossings included at least one driver who did not completely stop prior to entering the crosswalk. Overall, driver yielding for these 20 sites averaged 96 percent. In almost all of the crossings, drivers appropriately yielded to the crossing pedestrians.

For the pedestrian crossings observed, only 124 of the 1,689 pedestrians (7 percent) departed during a dark indication. For the majority of these pedestrians, the roadway volume was such that the pedestrian was able to find sufficient gaps to cross. The 1-min/lane volume count was less than 4 vehicles/min for the majority of these crossings. An examination of departures on the dark indication revealed that pedestrians were more likely to depart on dark at coordinated sites compared to hot-button sites (13 versus 7 percent), but departures on dark were much less frequent at the coordinated site that had pushbuttons with red lights that illuminated when the button was pressed. The coordinated site with the red-lighted buttons had 10 percent of pedestrians departing on dark, while the coordinated site with the non-lighted buttons had 20 percent of pedestrians departing on dark.

Of the 1,979 arriving pedestrians, 290 were research team members who always activated the PHB. For the remaining 1,689 general public pedestrians, 157 did not push the button because the PHB was already active. For those who arrived when the PHB was not active, 91 percent pushed the button and activated the PHB. A review of the data by site characteristics shows trends for the highest values. A greater number of pedestrians activated the device when on 45-mi/h posted speed limit roads as compared to 40 mi/h or less roads. The percentage of pedestrians pushing the button was always greater than 80 percent for the longer crossing distances (longer than 110 ft). When the hourly volume for both approaches was 1,500 vehicles/h or more, the percent of pedestrians activating the PHB was always 90 percent or more.

All occurrences of pedestrian/vehicle conflicts and erratic maneuvers were noted when observed in the video footage. The conflict rate was found to be higher for non-compliant pedestrians than for compliant pedestrians. Slightly less than half of the observed conflicts occurred during the dark beacon indication and involved a through vehicle. These conflicts usually involved pedestrians who either crossed without pushing the button or pushed the button but did not wait for their walk indication and then paused in the raised-curb median while crossing.

Notable conflict rates for both compliant and non-compliant pedestrians were observed at several sites where the PHBs were located near supermarkets and multiple bus stops. At these sites, many bus riders would walk through the supermarket parking lots or run across the major street while transferring between bus lines. The presence of bus stops near access points with significant turning vehicle volumes tended to result in higher conflict rates.

Discussion

The PHB has shown great potential in improving safety. It is also associated with less delay for the major roadway as compared to a full TCS because of the PHB's flashing red indication that permits stop-and-go operations if the pedestrians have finished crossing their half of the roadway.

Experience of city traffic engineers has indicated that drivers did not understand that they can start the stop-and-go operations once the crosswalk is clear. In response to this need, a sign was created and has been installed in several cities. FHWA now recommends that a slightly different wording be used on such a sign (see figure 58 for an example).

The results from this research have shown, however, that drivers are not always stopping on the flashing red before proceeding through the cleared crossing. In about half of the actuations where a queue existed on the approach, at least one of the drivers in the queue did not come to a complete stop before driving through the crossing. A small number of drivers (about 5 percent) was observed staying stopped on the flashing red indication, sometimes when it would have been clear to proceed, but often because pedestrians were still crossing or conflicting minor movement vehicles were occupying the intersection.

Within the 78 h of video data reviewed, conflicts were observed with most of the conflicts associated with non-compliant pedestrians. Several conflicts were observed at a site with a nearby access point (e.g., driveway), which could indicate that access points should be limited within a certain distance to the PHB, especially if they serve major traffic generators. Additional research is needed to determine the distance(s) access points should be restricted. The research should also consider the type of access point or the anticipated volume from the access point as well as proximity to bus stops where pedestrians may be making transfers between bus lines.

Most of the PHB sites included in this study were at intersections or major driveways. Including midblock sites was a priority for the study, and four locations were identified. The midblock PHBs had driveways/intersections that were within 80 ft of the PHB. All of these sites were in Austin, TX. Conflicts were not counted at two of these sites because of minimal cross street volume or restricted movements (e.g., right in/right out turns only). For the other two sites, 1 had minimal conflicts, but the remaining site had 11 conflicts. Examination of the video footage revealed that the conflicts at this site occurred when left-turning drivers departed a major traffic generator and did not have adequate space on the major street to complete the turning maneuver before encountering the midblock crosswalk. This site did not have a median, so the back-left corner of the turning vehicles were still in the opposing through lanes when the vehicles were stopped in the diagonal position. The drivers wanted to move out of their awkward position (diagonal, partly encroaching on opposing through lanes) and sometimes encroached on pedestrians in the crosswalk while doing so. Hence, guidance for the placement of PHBs and/or access points near PHBs needed to account for turning vehicle paths.

While drivers stopping at a dark PHB were observed, it did not appear that the stopping was caused by a driver being confused with the dark device. Rather, drivers stopped because of congestion from a nearby driveway or intersection or because of crossing pedestrians or stopped buses.

This study identified high driver yielding (greater than 94 percent) for the site with the widest crossing and the site with the 45-mi/h posted speed limit. This finding, along with findings from previous studies and the overall high yielding for PHBs identified in this research (overall 96 percent), supports the use of this device at a variety of locations, including on high-speed and wide roads, at residential intersections, and elsewhere.^(29,33)

FUTURE RESEARCH NEEDS

While several research studies have examined the effectiveness of the RRFB and the PHB, many research needs remain, as presented in this section. Several of these ideas were presented in a previous FHWA report but are included here for completeness.⁽⁵⁾

Based on the research conducted as part of this study, along with discussions held at professional society meetings and with other practitioners, additional research questions regarding RRFBs used at pedestrian crossings are as follows:

• Appropriate brightness level of RRFBs: The brightness of the RRFBs can help draw a driver's attention to a device and the area around the device. It can also result in drivers looking away from the device because the brightness is irritating or unbearable. When the discomfort glare is unbearable, drivers are more likely to divert their eyes away from the discomfort, which might result in drivers missing people or objects located near the glare source. Recommendations are needed for a maximum brightness for beacons used with pedestrian crossing signs and for other traffic control devices with embedded LEDs or supplemental beacons. The maximum brightness should vary between daytime and nighttime conditions.
  
  
• When RRFBs should be dimmed and by how much: Guidance is needed on whether to dim these devices during low light conditions and, if so, by how much. A study of disability glare and discomfort glare in both bright and dark conditions can be used to determine appropriate maximum nighttime and daytime brightness for RRFB. The investigation into brightness levels should consider an open-road portion to be able to associate different motorist yielding behavior with the different brightness levels.
  
  
• Appropriate use of RRFB assemblies on only one side of the roadway approach: The original IA for the RRFB requires the assembly to be located on the right-hand and left-hand sides of the roadway. There may be street widths where having two assemblies provides limited benefits. If so, the additional cost savings in purchasing and maintaining fewer devices at a site could provide additional resources to treat other locations.
  
  
• Appropriate installation of RRFB assemblies overhead rather than on the roadside: FHWA issued an interpretation in 2009 that indicated overhead mounting is appropriate and that if overhead mounting is used, a minimum of only one sign per approach is required, and it should be located over the approximate center of the lanes of the approach.⁽⁶⁾ Presence of buses and street width are two examples of site conditions where RRFBs could be installed overhead rather than roadside, but there might be other criteria that should be considered when making this decision. In addition to identifying the applicable criteria, developing numeric guidance for these criteria is also needed (e.g., at what roadway width should overhead rather than roadside installation be considered). The guidance might also need to consider additional variables beyond primary characteristics such as roadway width. For example, if the sidewalks at the site are adjacent to the face of curb, then the roadside assembly might need to be located more than 5 ft from the curb, which would place the assembly beyond the driver's cone of vision. The research would need to consider if placing the beacons above the sign would satisfy some of the concerns expressed in this research idea discussion. This research effort may also need to consider if larger beacons are needed for overhead application along with some adjustments to direct them to the approaching driver since many LEDs are directional.
  
  
• Identification of roadway and traffic variables that influence driver yielding: This research would examine the relationship between driver yielding rate and geometric and/or traffic variables. Additional research is needed to identify what characteristics are influential so that the characteristics (e.g., better enforcement or overall street design) of those communities with higher driver yielding could be reproduced.
  
  
• Identification of optimal use of signing and pavement markings at pedestrian crossings: This research would examine how signing can be used to improve a pedestrian crossing. Types of signs at some pedestrian crosswalks include warning signs used in advance of the crossing (sometimes with flashing beacon), signs at the yield or stop bar of the crossing to inform drivers of the appropriate place to stop, and signs at the crosswalk on the mast arm structure or roadside. Examples of signs being used at the crossing include regulatory signs such as the crosswalk stop on red [ball] and internally illuminated signs that say PEDESTRIAN CROSSING. Research questions could include (1) do the signs influence drivers' alertness at the crossing, (2) do the signs help to communicate the likelihood that a pedestrian will be at the crossing, or (3) how long are signs needed that provide information on stop-and-go behavior.
  
  
• How driver yielding for LED-embedded pedestrian crossing signs compares to RRFBs: Another flashing pedestrian treatment is signs with LEDs embedded into the sign face. The performance of this treatment as measured by driver yielding is needed.

Research needs associated with the PHB include the following:

• Optimal location of the PHB. At some sites, a logical place for a PHB is near a bus stop or other major traffic generator that may not be at a street intersection or a major driveway. At one of the sites in this research project, several conflicts were observed between vehicles and pedestrians with the vehicles turning in and out of a nearby driveway contributing to several of the conflicts. This research would examine tradeoffs for locating the PHB near a major access point. These tradeoffs may include changes in crosswalk utilization, pedestrian compliance, or conflict rate. A range of pedestrian volumes, access point volumes, and distances between crosswalk and access point should be considered along with identifying alternative treatments such as restricting some turning movements at the access point.
  
  
• Pedestrian compliance at coordinated PHBs: PHBs can be programmed to begin service to pedestrians instantly upon actuation (i.e., hot-button operation) or to begin service in coordination with adjacent traffic signals. Compared to instant service PHB operation, coordinated PHB operation reduces vehicular delay by preserving progression bandwidth and avoiding stopping the platoon of major street vehicles. However, a pedestrian actuating a coordinated PHB will often see delayed service and may conclude that the device is malfunctioning and then cross the street without the assistance and protection of the device. Research is needed to quantify pedestrian compliance trends as they are influenced by PHB operational mode (instant service versus coordinated), vehicular volumes, provision of visual or auditory feedback/guidance devices for waiting pedestrians (e.g., small indicator lights above the pushbutton that illuminate upon pressing), and other site characteristics. This insight would be used to formulate guidance on choosing between coordinated and instant service mode for a PHB.
  
  
• PHBs within signal system: Research is needed to determine the optimal background cycle length for a PHB (e.g., what should be the minimal major street green time between subsequent PHB activations?). The study could also investigate the minimum separation between a PHB and a signal that will allow a roadway to operate adequately. Investigate how that separation distance should change for various roadway features such as width of roadways.
  
  
• BikeHAWK: Research is needed to evaluate modifications to the PHB that can better accommodate bicyclist crossings along with pedestrians. Tucson, AZ, has developed a modified PHB called BikeHAWK, but there is a concern with the potential for late entry by bicyclists. Even though there is an R9-5 sign that instructs bicyclists to use the pedestrian signal, it is not known whether bicyclists know that there is a flashing red during the countdown. Even though the bicyclist can cross in the remaining time, a motor vehicle may be proceeding through the crossing. Other issues also exist in attempting to modify a PHB to better accommodate bicyclists along with pedestrians.

Other research needs for pedestrian treatments include the following:

• Guidance on selection of appropriate pedestrian crossing treatment for a particular location: In general, the PHB has higher yielding rates but costs more than RRFB assembly. The RRFB is more effective than many other pedestrian treatments; however, a Texas study found lower compliance for the RRFB for longer crossing distances.⁽³⁵⁾ This finding indicates that there is a crossing distance width for which a device other than the RRFB should be considered. The dataset included sites with total crossing distances that ranged between 38 and 120 ft. A research study with an objective of developing guidelines for selecting appropriate pedestrian crossing treatments would help to improve uniformity across the country. The study would also need to identify the site conditions that should be considered (e.g., roadway volume, pedestrian volume, crossing distance, posted speed limit, typical pedestrian walking speed at the site, etc.).
  
  
• Minimum number of pedestrians to justify a pedestrian treatment: There is a growing use of the PHB and the RRFB for pedestrian crossings. Establishing guidance that can be consistently applied would help to facilitate use of these devices in appropriate settings. A particular question is whether there is a minimum number of pedestrians before a device should be considered. The MUTCD contains graphs that illustrate when to consider a PHB, and these graphs include a minimum of 20pedestrians/h.⁽¹⁾ When deciding to recommend this minimum pedestrian number, the NCUTCD based its decision on a value developed through engineering judgment during an FHWA study on whether to mark crosswalks.⁽⁵¹⁾ Research is needed to more fully consider what should be the minimum pedestrian value used for selecting various traffic control devices. For example, should this minimum number be a function of crossing distance or posted speed limit? In addition, should it consider the distance to the nearest crossing? A location that is only a few hundred feet from an established crossing should have a higher minimum number compared with a crossing that is more than one-fourth or one-half of a mile from a signal on a wide high-speed arterial.
  
  
• Number of pedestrians induced as a result of installation of selected pedestrian treatments: The primary objective of this study would be to determine reasonable values for estimates of latent pedestrian crossing demand (i.e., estimated number of pedestrians that would now use the site because of the installation of a specific pedestrian treatment). The results of the research could improve the process for selecting pedestrian treatments. The research should make appropriate suggestions for changes to key reference documents, such as design manuals or the MUTCD.⁽¹⁾ Improved guidance should help to improve conditions for pedestrians by identifying appropriate devices for crossings, which should improve pedestrian mobility and reduce the number of pedestrian crashes.
  
  
• Drivers' search patterns near flashing beacons: There was evidence in this study that the closed-course drivers were more accurate in seeing objects beyond the signs with flashing beacons compared with seeing objects beyond the distractor signs. This could be an artifact of this study or it could be because the flashing beacons attracted the eye to the area. Additional research could focus on drivers' search patterns when a flashing beacon is present to test the theory that the presence of the beacons or LEDs encourages drivers to search a particular area. By varying the brightness of the beacons along with the light source (e.g., beacons or LED-embedded signs), the study could also investigate whether drivers need additional time to search an area because of the brightness of the device.
  
  
• Pedestrians' attitudes toward using treatments: Observations of pedestrians in the open-road portion of this study (and in other studies) have documented crossing pedestrians that did not activate the beacon treatments when they were provided. Some of those pedestrians were not within the treated crosswalk to be able to use the beacon, while others crossed at the crosswalk but chose not to activate the beacon. This study would explore pedestrian decisionmaking and examine why pedestrians who have the opportunity to use a treatment (such as an RRFB) to support their crossing choose not to do so. For example, at crosswalks marked as school crossings, do adult pedestrians think that the treatment is for use only by schoolchildren? Results from this study could feed into the suggested educational campaign mentioned previously, and results could also be used to support guidance on where treatments should be installed and what information (e.g., instructional plaques next to the pushbutton) should be provided to crossing pedestrians.
  
  
• Estimating pedestrian exposure: With ADT being the key predictor of vehicle crashes, there is a desire to have similar types of data for pedestrians. With limited resources for collecting counts—vehicle, bicycles, or pedestrians—researchers could study what are the most effective means for obtaining pedestrian exposure.
  
  
• Distance between crossings: How far will a pedestrian be willing to walk to reach a crossing with a pedestrian treatment? How does that distance change based on the treatment type (e.g., PHB versus crosswalk markings only), on the presence of a median, on the posted speed limit of the major street, and other factors that influence pedestrians walking behavior? These are all questions that could be investigated with further studies.
  
  
• National education campaign on the RRFBs and/or PHBs: Research is needed to determine what education campaigns have been used by cities and jurisdictions that have implemented RRFBs and whether they were successful. For example, are there common themes that could be used on a national level? The campaigns could also include other considerations of pedestrian behaviors such as the need to activate the pushbutton as well as cautions against distracted walking and walking during nighttime conditions, blind spots around commercial vehicles, and others. Education campaigns could be directed toward drivers, pedestrians, or both. The portion of the campaign could be directed to police who have to enforce the device to provide them with information on what is and is not a violation within their State laws.