Manual on Uniform Traffic Control Devices (MUTCD) arrows logo

Efficacy of Rectangular-shaped Rapid Flash LED Beacons


Participants and Setting

Participants consisted of drivers traveling on 58th St. N s/of 3rd Avenue and the crossing pedestrians. The location at 58th St. N s/of 3rd Avenue traversed four lanes of traffic, a posted speed limit of 35 MPH and an ADT of 19,192. It also had a median island and provided a crossing for the residents of a near-by retirement center. The second location was at 4th St. S & 18th Avenue. This location was equipped with a side-mounted system. This roadway traversed four lanes, has an ADT of 9,600, and a posted speed of 35 MPH.


The treatment in this experiment was the standard over-head yellow flashing beacon and a standard side-mounted yellow beacon (see Figure 7). These systems are activated with a pedestrian call button. The system employed two 12-inch diameter yellow beacons facing each direction. The beacons were flashed at a rate of 55 times per minute and the illumination period of the beacon was 50 percent of the time.

Experimental Design

An ABCD design was employed at the 58th St. N s/of 3rd Avenue site to measure the efficacy of standard over-head beacons. This design was comprised of collecting baseline data in the absence of activation of the standard system. The system was activated during treatment. Collecting seven data sheets of data comprised of 20 crossings each. Following the standard beacon treatment, a rapid-flash two-beacon system was implemented followed by the four-beacon system. The rapid-flash treatments were repeated. Each rapid-flash treatment was observed for five data sheets each. This gave a total of 680 crossings.

FIGURE 7 Photographs showing a traditional over-head circular incandescent flashing beacon (left photograph) and a round side-mounted beacon (right photograph).
Figure 7 - Standard Over-head and Side-mount Round Flashing Beacons

An ABC design was used at the 4th St. S @ 18th Avenue location. Baseline consisted of 46 crossings. After baseline, a side-mounted standard beacon system was evaluated for 70 crosses at seven and 30-day intervals. Following the B phase of treatment, a two-beacon rapid-flash system was installed and evaluated at the seven and 30-day intervals after rapid-flash installation. The standard and rapid-flash evaluations each consisted of 70 crossings.

Statistical Analysis

Driver Yielding Behavior

The average yielding compliance at the 58th St. N s/of 3rd Avenue site during baseline recording was 10.9%. The activation of the over-head standard beacon produced an average yielding compliance of 15.5%. This is an average increase of only 4.6% above baseline. The introduction of a two-beacon, rapid-flash, system produced an increase in yielding to 78.3%. A four-beacon system followed giving 88% yielding compliance. Reversal back to two beacons yielded 84.6% compliance followed by 89.3% yielding for the second four-beacon system treatment. The average yielding percentage for a two-beacon system was 81.5%. The average yielding compliance for the four-beacon system was 88.7%. With the introduction of a two- and four-beacon system came increases of 70.6% and 77.8% increases over baseline, respectively, and increases of 66% and 73.2% over the standard-beacon efficacy. (See figure 8).

FIGURE 8 Set of two bar graphs illustrating the effectiveness of the rectangular-shaped rapid flash LED beacon when compared to either a traditional over-head mounted or side-mounted circular flashing beacon.

Figure 8 - Yielding Comparison between Standard and Rectangular-shaped

Baseline data at the 4th St. & 18th Avenue site showed a 0.0% yielding compliance. Activating the side-mounted standard beacon produced a 12.2% yielding compliance after seven days. A 30-day analysis of the standard side-mounted system yielded 17% compliance. The rapid-flash produced 63.4% yielding compliance after 7 days and the 30-day analysis showed 72% yielding. The rapid-flash percentages are representative a two-beacon system only. The average yielding percentage for each of the two-beacon analysis is 67.7%. This number is 55.5% more yielding over the 12.2% observed during the standard beacon treatment.

Driver Yielding Distance Behavior (58th St. N s/of 3rd Avenue only)

The absence of the standard beacon actually produced a better effect on yielding distance than during standard overhead beacon. During treatment, light on, a higher percentage (one percent more) of the vehicles yielded at less than 30ft. However, there are more cars yielding during treatment and this produces a larger number of cars that yielding at closer distance than in the absence of the light. There were 48 cars yielding at less than 30 feet during treatment with only 27 during baseline. There were also a smaller percentage of cars yielding at >100ft. during treatment, 5.6%, as opposed to 8.4% of vehicles yielding at >100ft. during baseline. The majority of yielding during both conditions occurred at the same distance, 30ft.-50ft. During baseline, 41% of motorists yielded at this distance and 42.7% during the standard beacon treatment. The majority of yielding during the two-beacon system occurred at the 30ft.-50ft. interval (43.5%). During the four-beacon system, the majority was at the same interval with 41.7%. The percentage of motorists yielding greater than 100ft. more than doubled from the two-beacon system to the four-beacon system with an increase from 5.6% to 12%.

There were no significant results reported for Evasive action: pedestrian/vehicle, Pedestrian trapped in median or Car behind yielding or drivers jamming on brakes.

Inter-observer agreement

Inter-observer agreement on the occurrence of a yielding behavior averaged 92% with a range of 80% to 98%. Inter-observer agreement on evasive conflicts was 100%. Inter-observer agreement on whether the pedestrian was trapped in the center of the road averaged 100%, inter-observer agreement on vehicle passes or pass attempt averaged 100%, inter-observer agreement on vehicles that jammed on brakes averaged 100%, and inter-observer agreement on stopping distance averaged 99%.

previous | next