By Kay Fitzpatrick, Ph.D., P.E., Ann Do, P.E., and Bruce Friedman, P.E.

Transcription

1 feature shutterstock.com/dieter Hawlan Rapid-Flashing Beacons for Pedestrian Treatments By Kay Fitzpatrick, Ph.D., P.E., Ann Do, P.E., and Bruce Friedman, P.E. The rectangular rapid-flashing beacon has received national attention and interim Federal Highway Administration (FHWA) approval; however, before including it in the Manual on Uniform Traffic Control Devices, additional information is needed on several issues. To help answer these questions, FHWA conducted studies on the effects of shape (circular or rectangular), flash pattern, and beacon location (above or below the sign). December

2 Source: Fitzpatrick, K., et al. 2 Various traffic control devices and pavement markings have been used at uncontrolled pedestrian crosswalks to increase driver awareness and to improve driver yielding to pedestrians. A device that has received national attention is the rectangular rapid-flashing beacon (RRFB). On July 16, 2008, the Federal Highway Administration (FHWA) issued interim approval (IA-11) for the optional use of the RRFB. 1 FHWA approved the use of this device at pedestrian and school crosswalks across uncontrolled intersection approaches and at midblock crossings. An RRFB consists of two rapidly flashing rectangular yellow indications with light-emitting diode (LED) array-based pulsing light sources. Figure 1 shows an example of an RRFB. The RRFB is activated by a pedestrian, and previous studies have found driver yielding ranging from 34 to 98 percent. 2 The Signals Technical Committee (STC) of the National Committee on Uniform Traffic Control Devices makes recommendations to FHWA regarding the contents of chapter 4 of the Manual on Uniform Traffic Control Devices (MUTCD). 3 The initial research studies did not address certain issues that the STC believes to be important in crafting language suitably generic for the MUTCD. For example, when the interim approval for the RRFB was issued, the only flash pattern that had been tested had two pulses on the left-hand side of a light bar followed by five pulses on the right-hand side (commonly called the 2-5 pattern). Because the 2-5 pattern usually appears to the human eye to be a 2-3 flash pattern, several devices were installed with the 2-3 pattern rather than the 2-5 pattern. The 2-3 flash pattern has two pulses on the left-hand side of the light bar followed by three pulses on the right-hand side. Therefore, the STC is seeking advice on several issues, including the following: Do the housings have to be rectangular, or will circular-shaped housings achieve the same effect? and Since the initial studies were only conducted with one rapid-flashing pattern, are there more effective rapid-flashing Figure 1. Photo of a Rectangular Rapid-Flashing Beacon Included in the Field Study 40 December 2014 ite journal patterns available? FHWA sponsored work to assist in the investigation of these questions. 2,4-7 This paper provides information on the findings to date from those efforts. Rapid-Flashing Yellow LEDs Impact Detecting Pedestrians in a Closed-Course Setting 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, LED brightness or other characteristics 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). Disability glare impairs a driver s ability to detect hazards near the device even in situations where the driver is not experiencing discomfort glare. On the other hand, discomfort glare is the perceived discomfort of the light source and might result in drivers looking away from a device. Either condition disability glare or discomfort glare might result in drivers missing hazards located near the source of the glare. And in the case of LEDs used at pedestrian crossings, this might affect drivers ability to see the sign s legend or to detect pedestrians, especially during nighttime conditions. A 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. During both daytime and nighttime conditions, participants drove the study vehicle to the starting location, where they parked the vehicle 200 feet from the sign assemblies, which consisted of a pedestrian crossing sign with LEDs within the sign face and LEDs in rectangular beacons above and below the sign. For the beacons above the sign, the bottom edge of the beacon housing was approximately 11.6 ft. from the pavement. For the beacons below the sign, the bottom edge of the beacon housing was approximately 7.0 ft. from the pavement. When the LEDs were embedded within the sign, the height to the middle of the sign was approximately 9.5 ft. After the drivers placed their vehicles in park, the drivers were asked to wear occlusion glasses. The occlusion glasses obscure the participant s vision by going opaque when there is no power supplied to them and going clear when power is supplied. Once the driver s vision was occluded, technicians placed a static cutout photo of a pedestrian (either 4.5 ft. high to represent a child or 5.75 ft. high to represent an adult) within the crosswalk located near the sign assemblies (see Figure 2). A technician then restored the driver s vision, and the participant was 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 driver pressed the button, the glasses turned opaque again. Following the driver s identification of the pedestrian s direction, the researcher asked the participant to rate the intensity of the LEDs (comfortable, irritating, or unbearable) before

3 Source: Fitzpatrick, K., R. Avelar, and J. Robertson. 5 Figure 2. Researcher Removing Short Cutout Pedestrian after Placing Tall Cutout Pedestrian 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. The two key measures examined in this study were detection time (i.e., the time for the participant to identify which direction the pedestrian was walking) and discomfort glare as measured by the driver s rating of comfortable, irritating, and unbearable. The following is an overview of the important findings from the closedcourse research study: Average nighttime detection time for the participants to search and determine which direction a cutout pedestrian was walking was and seconds for older (55 years old and older) and younger (less than 55 years old) participants, respectively. Average daytime detection time for the participants was, as expected, faster (1.281 and seconds for older and younger participants, respectively). LED intensity had a measurable adverse impact on detection time at night but not during the day. Under nighttime conditions and using 0 candelas as the base condition, detection time increased 8.5 percent when 2,000 candelas was present. 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 the sign compared to LEDs within the sign, and 12 percent longer for LEDs below compared to LEDs above the sign; likewise, detection times with LEDs within were 6 percent longer than for LEDs above. Discomfort glare differed by LED position at night, with a 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 pattern had a significantly larger detection time (5.2 percent longer) than no flash pattern. At night, both 2-5 and wig-wag patterns were found to delay detection compared to no pattern (increases of 6.0 and 13.7 percent, respectively). The wig-wag flash pattern has two signal indications that flash alternately, with each signal indication being illuminated for approximately the same length of time, such that the total time of illumination of the pair of signal indications is the entire operating time. Pedestrian position had an impact on detection time during both day and night. Under both conditions, detection was faster when the pedestrian was located at the center of the crosswalk. For both conditions, detection times for a pedestrian at the left or at right Key Findings from FHWA Studies on Pedestrian-Activated Rapid-Flashing Beacons: Previous studies have found driver yielding to an installed RRFB ranging from 34 to 98 percent. A closed-course study revealed that LED intensity (brightness) had a measurable adverse impact on detection time of a cutout pedestrian photo and that certain flash patterns had statistically longer detection times. An open-road study found that brighter beacons were associated with higher yielding. The shape of the rapid-flashing beacon (circular or rectangular) does not have a significant impact on whether a driver decides to yield to pedestrians. Three rapid-flash patterns were used with rectangular beacons at eight sites, and no significant difference in yielding was found for those three patterns. FHWA released an official interpretation that states a preference for using the wig-wag and simultaneous flash (WW+S) pattern on future installations of the RRFB (see mutcd.fhwa.dot.gov/resources/interpretations/4_09_41.htm). December

4 were not statistically different from each other. 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. Circular or Rectangular Beacon Shape The open-road study was conducted at 12 pedestrian crossings in four cities (Austin, TX; College Station, TX; Flagstaff, AZ; and Milwaukee, WI). Two beacon shapes were tested: rectangular and circular. Each city identified two to five locations where they had installed or were planning to install a pedestrian treatment and where they would be willing to switch between the RRFB and the circular rapid-flashing beacon (CRFB) at the site. This enabled the research team to evaluate both devices at the same site and, therefore, have similar pedestrian and driver populations. At half of the sites, the RRFB was installed first, followed by the CRFB; at the other half of the sites, the CRFB was installed first, followed by the RRFB. Examples of two study assemblies are shown in Figure 1 for the RRFB and Figure 3 for the CRFB. The data collection which included a data collection period following installation of the first device and a data collection period following installation of the second device was conducted between November 2012 and April Data were collected when vehicles were free flowing. Since the characteristics of the beacons and sites might have different impacts on drivers yielding at night, nighttime data were collected for one of the sites within each city. The research team used a staged pedestrian protocol to collect driver yielding data to ensure that oncoming drivers receive a consistent presentation of approaching pedestrians. Under this protocol, a member of the research team acted as a pedestrian using the crosswalk to stage the conditions under which driver yielding would be observed. Each staged pedestrian wore similar clothing (gray T-shirt, blue jeans, and gray tennis shoes) and followed specific instructions for crossing the roadway. The staged pedestrian was accompanied by a second researcher, who observed and recorded the yielding data from a concealed position. Typically, a minimum of 60 staged pedestrian crossings were obtained at each site during daytime conditions, and a minimum of 40 staged pedestrian crossings were obtained at night. The range of driver yielding to staged pedestrians at sites with either the RRFB or CRFB went from a low of 22 percent to a high of 98 percent. At night, driver yielding to staged pedestrians averaged 72 percent for the RRFB and 69 percent for the CRFB. During the day, driver yielding to staged pedestrians averaged 59 percent for the RRFB and 63 percent for the CRFB. From the preliminary review of the findings, it appears that there are only minor, if any, differences between the CRFB and the RRFB. The results from the generalized linear mixed model indicate that there are no significant differences between the two beacon shapes. For Figure 3. Photo of a Circular Rapid-Flashing Beacon Used in the Field Study a subset of the sites, the brightness of the beacons was measured. For those sites, there is clear evidence of an increasing yielding rate with increasing intensity at night. The trend is in the same direction during the day but with a smaller magnitude, which the analysis found statistically insignificant. In conclusion, the shape of the yellow rapid-flashing beacon does not have a significant impact on whether a driver decides to yield to pedestrians. Variables that did have an impact on driver yielding include beacon intensity (for nighttime) and city (yielding was higher in Flagstaff compared to the other three cities). Rapid-Flashing Patterns An open-road study format was also used to examine different flash patterns for RRFBs. The measure of effectiveness was the percentage of drivers who yielded to or stopped for a staged pedestrian who activated the RRFB and was attempting to cross the roadway. The study included eight sites located in College Station and Garland, TX. Seven of the eight sites had four lanes with a posted speed limit of 40 or 45 miles per hour (mph). The remaining site had two lanes and a posted speed limit of 30 mph. Figure 4 is a photo of one of the sites. A temporary light bar and controller were developed to give the research team control over several of the beacon s characteristics, such as the flash pattern and the brightness. The light bar was designed so that it was not obvious that the beacons being observed during the staged pedestrian crossings were any different from the permanent RRFB light bars to which they were mounted. Table 1 illustrates the three flash patterns selected for testing in the field using the temporary light bars. The patterns examined in this study included the following: Source: Fitzpatrick, K., et al December 2014 ite journal

5 Table 1. Three flash patterns selected for testing in the field using the temporary light bars Pattern Blocks WW+S 2-5 Cumulative Time in Milliseconds (ms) Left a Right b Left a Right b Left a Right b (ms) (ms) (ms) (ms) (ms) (ms) On time (ms) Percent of cycle for a given beacon with the beacon on 38% 38% 25% 25% 31% 38% On ratio = percent of cycle where at least one of the beacons is on 56% 37% 69% Off ratio = percent of cycle where both beacons are dark 44% 63% 31% Yellow cell = beacon is on for 25 ms Gray cell = beacon is off a Left time beacon is on (ms). b Right time beacon is on (ms). Source: Fitzpatrick, K., R. Avelar, J. Robertson, and J. Miles. Driver Yielding Results for Three Rectangular Rapid-Flash Patterns. Executive Summary, Texas A&M Transportation Institute website, Available: (Accessed on October 7, 2014.) December

6 Source: Fitzpatrick, K., R. Avelar, J. Robertson, and J. Miles. 7 Figure 4. Study Site with Installed Temporary Light Bars and Staged Pedestrian Crossing A proposed pattern using a combination of long and short flashes (called Blocks); A proposed pattern using a combination of wig-wag and simultaneous flashes (called WW+S); and The pattern that was specified in FHWA s original interim approval (called 2-5). The research team used a staged pedestrian approach to evaluate driver yielding for the different patterns. Data were collected for a minimum of 40 crossings for each pattern at each site during February and March The brightness of the flashing beacons was the same for all three flash patterns and for all eight sites. Logistic regression was used to model the yielding and not-yielding data for each individual crossing. The results from the generalized linear mixed model indicate that there is no significant difference between the 2-5 pattern and the WW+S pattern or between the 2-5 pattern and the Blocks pattern. To provide an overview of the yielding, the overall average driver yielding for each pattern at the eight sites was also calculated. The overall average driver yielding was 80 percent for the WW+S and the Blocks patterns and 78 percent for the 2-5 pattern. Thus the WW+S and Blocks patterns developed as part of this research study were equally as effective as the 2-5 pattern. Discussion In the closed-course study, the brightness intensity of the LEDs used ranged from 0 (i.e., the LEDs were not on) to 2,200 candelas. 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 (statistically significant). The brighter the LEDs, the longer it took for the participants to determine which direction the pedestrian was facing. Brightness was also measured as part of the open-road study comparing beacon shape. In that study, the brighter beacons were associated with higher driver yielding results. The results from the open-road and closed-course studies provide a mixed message. The open-road study indicates brighter is better (higher yielding), while the closed-course study indicates brighter is worse (longer time to detect pedestrian). Future research is needed to identify the optimal brightness levels needed for daytime and nighttime conditions. Some of the flash patterns used with the devices in the closedcourse study were associated with longer detection times. Of the six flash patterns tested, only two flash patterns the 2-5 and the wig-wag were associated with statistically significantly longer detection times when compared to the no flash pattern (i.e., beacons are not on) condition. Both of these patterns have longer on times (the 2-5 is on 69 percent of the cycle and the wig-wag is on 100 percent of the cycle) as compared to the other patterns tested (range of 10 to 38 percent on time). The LEDs being constantly on might cause the participants to look away from the device. In addition, the lack of sufficient dark period(s) between the flashes might be limiting the participant s ability to adequately search for the pedestrian. A better flash pattern than the current 2-5 pattern should retain multiple pulses (because survey results found that 44 December 2014 ite journal

7 participants felt patterns with multiple pulses are associated with greater urgency), more dark periods (because the closed-course study found longer detection times for patterns with fewer 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). 5 The findings for pedestrian position and LED location in the closed-course study indicate that the distance between the pedestrian and the light source affect 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. Another phase of the FHWA study anticipated for spring 2015 will investigate this theory in an open-road setting. The findings from these research efforts were presented to the STC during its June 2014 meeting. The STC recommended that the WW+S pattern be used with future rapid-flashing pedestrian treatments. They also recommended that the beacon shape used with a rapid-flashing beacon at a pedestrian treatment could be either circular or rectangular. 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. 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 because they are usually powered by solar energy. itej References 1. Furst, A. MUTCD- Interim Approval for Optional Use of Rectangular Rapid Flashing Beacons (IA-11). U.S. Department of Transportation Federal Highway Administration Memorandum, July 16, 2008, Washington, DC. 2. Fitzpatrick, K., et.al. Investigating Improvements to Pedestrian Crossings with an Emphasis on the Rectangular Rapid-Flashing Beacon. Anticipated Federal Highway Administration. Manual on Uniform Traffic Control Devices for Streets and Highways 2009 Edition. U.S. Department of Transportation Federal Highway Administration, Washington, DC, Fitzpatrick, K., J. Robertson, and R. Avelar. Closed-Course Study of Driver Detection of Pedestrians beyond Flashing Beacons within a Sign Assembly. to be published in the Journal of Transportation Research Record. Anticipated Fitzpatrick, K., R. Avelar, and J. Robertson. Rapid-Flash Yellow LEDs Impact on Detecting Pedestrians in a Closed-Course Setting draft Technical Memorandum to the Federal Highway Administration, Fitzpatrick, K., R. Avelar, J. Robertson Comparison of Driver Yielding to Staged Pedestrians for Three Rectangular Rapid-Flash Patterns draft Technical Memorandum to the Federal Highway Administration, Fitzpatrick, K., R. Avelar, J. Robertson, and J. Miles. Driver Yielding Results for Three Rectangular Rapid-Flash Patterns Executive Summary, Texas A&M Transportation Institute website, Accessed from: d2dtl5nnlpfr0r.cloudfront.net/tti.tamu.edu/documents/tti pdf. Accessed on: October 7, Kay Fitzpatrick, Ph.D., P.E., was recently honored with the Burton Marsh award for distinguished service to ITE. She is currently the president of the ITE Brazos Valley Section and the Local Arrangement Chair for the 2015 Spring TexITE Meeting. She was a member of the executive committee and then the chair of the ITE Traffic Engineering Council. She has written chapters in the ITE Traffic Engineering Handbook and the Urban Street Geometric Design Handbook and was one of the assistant editors for the 2000 edition of the ITE Traffic Control Devices Handbook. She is the co-author of several ITE briefing sheets, ITE Compendium articles, and ITE Journal papers. She is a fellow of ITE. Ann H. Do, P.E., currently serves as Research Highway Engineer for Turner-Fairbanks Highway Research Center, McLean, Virginia. Ann has been the program manager for FHWA Pedestrian and Bicycle Safety Research since She is responsible for designing and managing the research study and providing technical assistance, guidance, and support to other FHWA offices and to state and local transportation agencies in areas related to pedestrian and bicycle safety. She received a bachelor of science in civil engineering from Virginia Polytechnic Institute and State University, Blacksburg, VA, in June Bruce E. Friedman, P.E., is a transportation specialist with the Federal Highway Administration s Manual on Uniform Traffic Control Devices (MUTCD) team in Washington, DC. He is responsible for Parts 4 and 8 of the MUTCD, which includes highway traffic signals and beacons. Prior to joining FHWA in 2008, he was a member of the National Committee on Uniform Traffic Control Devices for 25 years. He received a bachelor of science degree and a master of science degree in civil engineering, both from the Georgia Institute of Technology in June 1972 and August 1973, respectively. He is a fellow of ITE. December