Skip to contentUnited States Department of Transportation - Federal Highway Administration FHWA Home
Research Home
Report
This report is an archived publication and may contain dated technical, contact, and link information
Publication Number: FHWA-HRT-11-039
Date: April 2011

Evaluation of Pedestrian and Bicycle Engineering Countermeasures: Rectangular Rapid-Flashing Beacons, HAWKs, Sharrows, Crosswalk Markings, and the Development of an Evaluation Methods Report

CHAPTER 3. EFFECTS OF YELLOW RRFBS ON YIELDING AT MULTILANE UNCONTROLLED CROSSWALKS

This chapter summarizes the FHWA report, Effects of Yellow Rectangular Rapid-Flashing Beacons on Yielding at Multilane Uncontrolled Crosswalks, FHWA HRT 10-043.(9)

OBJECTIVE

This study included a series of experiments to examine the effects of side-mounted yellow light-emitting diode (LED) RRFBs at uncontrolled marked crosswalks. Many methods to increase driver yielding behavior to pedestrians at multilane crosswalks at uncontrolled sites with relatively high average daily traffic (ADT) were examined. In previous studies, only treatments that employed a red phase consistently produced sustained high levels of driver yielding behavior.(10) A series of 5 experiments at 22 sites in 3 cities in the United States (St. Petersburg, FL; Washington, DC; and Mundelein, IL) examined the effects of RRFBs on driver yielding behavior. Data were also collected over a 2 year follow-up period at 18 of these sites to determine any long-term effects of the RRFB treatments. Another objective of the study was to compare an RRFB with a traditional overhead yellow flashing beacon and a side-mounted traditional yellow flashing beacon. A final objective of the study was to assess ways to further increase the effectiveness of the treatments. Variants subjected to evaluation included mounting additional units on a median or pedestrian refuge island and aiming the RRFB systems to maximize brightness.

INTRODUCTION

Drivers generally do not yield the right-of-way to pedestrians in marked crosswalks at uncontrolled sites. One alternative to in-roadway signs and yellow flashing beacons is to add yellow LED RRFBs to pedestrian warning signs. These LED RRFBs are similar in operation to emergency flashers on police vehicles. Figure 3 shows an example of an RRFB mounted below a W11-2 pedestrian warning sign at a crosswalk. This system is solar powered and is linked to the unit on the other side of the street by radio frequency transmitters and receivers. Each LED flasher is 6 inches wide and 2.5 inches high and is placed 9 inches apart. Each unit is dual indicated, with LEDs on the front and back. Each side of the LED flasher illuminates in a wig-wag sequence (left and then right). The left LED flashes two times in a slow volley each time it is energized (124 milliseconds (ms) on and 76 ms off per flash). This is followed by the right LED, which flashes four times in a rapid volley when energized (25 ms on and 25 ms off per flash) and then has a longer flash for 200 ms. The effect has been described as a "stutter-flash effect."(11) This pattern was selected because it is similar to one of the patterns used by emergency vehicles. Advance yield markings were installed prior to collecting baseline data to reduce the risk of multiple-threat crashes.

Figure 3. Photo. RRFB with two forward-facing LED flashers and a side-mounted LED flasher. Click here for more information.

Figure 3. Photo. RRFB with two forward-facing LED flashers and a side-mounted LED flasher.

METHODOLOGY

The general methodology for all of the experiments was measuring driver yielding behavior and vehicle/pedestrian conflicts. Details on methodology are contained in the final report.(9) For driver yielding, observers scored the percentage of drivers who did and did not yield to pedestrians. Drivers were scored as yielding if they stopped or slowed and allowed the pedestrian to cross. Conversely, drivers were scored as not yielding if they passed in front of the pedestrian but would have been able to stop when the pedestrian arrived at the crosswalk.

The Institute of Transportation Engineers signal formula applied to calculate the duration of the yellow signal phase was used to determine whether a driver could stop safely.(12) A landmark associated with this distance was identified for each approach to the crosswalk. Drivers who passed this landmark before the pedestrian started to cross could be scored as yielding to pedestrians but not as failing to yield because they may not have had sufficient distance to stop safely. Drivers beyond the landmark when the pedestrian entered the crosswalk could be scored as yielding or not yielding because they had sufficient distance to stop safely. When the pedestrian first started to cross, only drivers in the first half of the roadway were scored for yielding. Once the pedestrian approached the painted median, the yielding behaviors of drivers in the remaining two lanes were scored. This procedure was used because it conformed to the obligation of drivers specified in the statutes of each of the three cities that were studied.

RESULTS

Geographic Sustainability

Yielding during the baseline period before the introduction of the RRFB ranged between zero and 26 percent. The introduction of the RRFB was associated with yielding that ranged between 72 and 96 percent at the 2-year follow-up. Table 1 shows the percentage yielding at each of the 22 sites.

Table 1. Baseline and follow-up yielding data at sites in Florida, Illinois, and Washington, DC.

Site

Day (Percent)

Baseline (Percent)

7

30

60

90

180

270

365

730

Florida

31st Street south of 54th Avenue S

0

54

76

N/A

59

N/A

91

75

83

4th Street at 18th Avenue S

0

63

72

N/A

69

N/A

69

80

80

22d Avenue N and 7th Street

0

97

96

91

93

92

91

98

96

9th Avenue N and 26th Street

0

80

82

85

95

81

88

77

78

22d Avenue N and 5th Street

8

87

89

92

92

87

96

92

95

Martin Luther King Street and
15th Avenue S

1

86

84

85

82

N/A

89

88

88

Martin Luther King Street and
17th Avenue N

0

96

94

80

82

83

88

82

83

1st Avenue N and 13th Street

2

85

87

75

78

N/A

91

88

N/A

9th Avenue N and 25th Street

0

86

90

83

90

N/A

88

81

79

1st Street and 37th Avenue N

0

79

87

85

87

N/A

90

97

95

58th Street and 3d Avenue N

0

85

84

85

85

79

92

82

88

Central Avenue and 61st Street

0

94

95

77

73

72

79

67

72

1st Avenue S and 61st Street

5

68

72

73

75

72

90

72

78

1st Avenue N and 61st Street

0

75

75

68

82

42

76

79

83

83d Avenue N and Macoma Drive

0

86

93

91

73

88

84

80

90

9th Avenue N and 45th Street

0

54

91

89

90

80

83

77

78

22d Avenue S west of 23d Street

0

89

86

78

77

60

75

81

82

62d Avenue S and 21st Street

0

77

76

77

53

78

81

84

80

9th Avenue N and 31st Street

16

93

95

89

88

82

82

89

N/A

Florida Average

2

81

86

82

80

76

86

83

84

Illinois

Midlothian Road and Kilarny
Pass Road

7

62

62

N/A

N/A

N/A

N/A

N/A

N/A

Hawley Street and Atwater Drive

19

71

68

N/A

N/A

N/A

N/A

N/A

N/A

Illinois Average

13

67

65

N/A

N/A

N/A

N/A

N/A

N/A

Washington, DC

Brentwood Road and 13th Street

26

62

74

N/A

N/A

80

N/A

N/A

N/A

Average Yield (All Sites)

4

78

82

83

80

77

85

83

84

N/A indicates that the measure was missed or has not yet been scheduled.

The general statistical methodology used in this study was based on the general time-series intervention regression modeling approach described in Huitema and McKean and McKnight et al. (See references 13–16.)

The five main parameter estimates are shown in table 2. There is an immediate and large increase in yielding from the baseline to day 7, a small but statistically significant additional increase from day 7 to day 30, a minor and not statistically significant decrease at day 60, and a general trend after day 60 that has little slope across the remaining observation days. Therefore, the evidence for change is overwhelming, and the change is maintained for the 2-year duration of the study.

Table 2. Florida data estimates of treatment effect parameters and associated t-ratios and p-values.

Treatment Effect Parameter

Parameter
Estimate

t-Ratio

p-Value

Baseline level

  1.79

 

 

Level change day 7

  77.25

29.22

0.001

Level change day 30

  6.03

2.38

0.02

Level change day 60

 -4.26

-1.75

0.08

Follow-up slope

0.0059

1.62

0.11

Note: Certain cells were left blank because only t-ratios and p-values that show change from the baseline were included. In the table, there are 166 degrees of freedom for all tests.

Two-Beacon Versus Four-Beacon Systems

This experiment evaluated the effectiveness of the installation of only two RRFBs (one for each direction of an approach mounted at the right-hand side of the approach) as opposed to the installation of four RRFBs (two per approach with one on the roadway median and one on the right-hand side). The average yielding during baseline conditions across four sites was 18.2 percent. Installation and activation of the two RRFB systems increased the average yielding to 81.2 percent. The addition of the median beacons produced a further increase in yielding to 87.8 percent. Yielding for the four-beacon system continued to improve over time during follow-up data collection.

Aimed Versus Unaimed Beacons

This experiment evaluated the effectiveness of RRFBs with LEDs aimed parallel to the approach roadway compared to RRFBs with LEDs specifically aimed toward the eyes of approaching drivers at a given distance in advance of the crossing. The percentage of drivers yielding to pedestrians during the baseline condition was zero percent. The average yielding compliance 7 days after RRFB installation increased to 33.4 percent. There was an additional increase to 72 percent 30 days after installation. The change from parallel LEDs to LEDs that could be aimed produced an increased average of 89 percent.

Night Versus Day Operations

Night data were collected at one site where data had also been collected during the daytime. During daytime collection, the site had a baseline average yielding rate of 18.3 percent. The initiation of the two-beacon and four-beacon RRFB systems increased yielding to 86.7 and 89.6 percent, respectively. When the site was evaluated during nighttime hours, baseline yielding was only 4.8 percent. Introduction of the two-beacon and four-beacon RRFB systems showed increases in yielding to 84.6 and 99.5 percent, respectively.

Standard Beacons Versus RRFBs

Two sites were selected for this experiment. The first site had an above roadway standard yellow flashing beacon, while the second site was equipped with a side-mounted standard yellow flashing beacon attached to a pedestrian warning sign. The average baseline yielding when the standard beacons were present but not activated was 12 percent for the above roadway beacon and zero percent for the side-mounted beacon. Activating the overhead standard beacon produced an average yielding compliance of 15.5 percent. The introduction of a two-beacon RRFB system at this site produced an increase in yielding to 78.3 percent. The introduction of the four-beacon RRFB system was associated with 88 percent yielding compliance. At the second site, activation of the side-mounted standard beacon produced 12 percent yielding compliance. The two-beacon RRFB produced 72 percent yielding compliance. A four-beacon RRFB system was not available for this second site.

CONCLUSION

The results show that the rectangular LED yellow RRFBs appear to be an effective tool for producing large numbers of drivers who yield right-of-way to pedestrians in crosswalks at sites where drivers rarely yielded to pedestrians in the past. The results seem to be maintained over time. Because 19 systems were introduced in St. Petersburg, FL, it is evident that the effects do not diminish when a modest number of systems are installed. However, it is not clear whether the effects will diminish if the device is installed at hundreds or thousands of sites. The findings of this study suggest that the RRFB used in conjunction with advance yield markings can increase yielding and may increase safety at uncontrolled crosswalks at high ADT multilane sites. The data also indicate that mounting additional beacons on a pedestrian refuge island or median increases yielding behavior over using just side-mounted beacons alone. Additionally, aiming the beacon to vehicles at the dilemma zone may also increase yielding behavior. Future research should examine crash data using time-series or empirical Bayes (EB) methodology to determine the safety benefits of the RRFB.

ResearchFHWA
FHWA
United States Department of Transportation - Federal Highway Administration