The City of Somerville, MA, USA, undertook a test of strobes in red signal lenses to evaluate their effectiveness in improving the violation rate at pedestrian-actuated intersections. The City's Department of Traffic and Parking had documented a pattern of high violation rates for red indications at pedestrian signals. The City decided to test the effectiveness of strobes to determine if a heightened state of awareness would result in a reduced violation rate.
To document the effectiveness of the strobes for the City's purposes, the Department of Traffic and Parking conducted a before and after study of violation rates. The department was permitted by City officials to conduct tests at only two locations.
The intent of the study-and this feature-was to provide the City's elected officials with a greater understanding of the potential of this new technology in addressing safety concerns. In addition, the authors wanted to ensure that their experiences would be shared with other practitioners for the benefit of the industry.
The City selected this technology for testing because multiple local elected officials had observed the devices in other jurisdictions. These officials were interested in local applications to improve safety. Although the strobes are not currently included in the Manual on Uniform Traffic Control Devices (MUTCD), the City felt that there could be tremendous benefit from this product.
The City did not elect to seek Federal Highway Administration (FHWA) approval for the experimental use of the technology. Because the strobes themselves were not a device requiring mandatory actions by motorists, the City determined that a test of the technology would be appropriate.
However, consistent with the requirements in MUTCD and articulated by FHWA, the authors felt that a published study of the test results would be critical to better understanding this experimental technology.
PREVIOUS RELATED STUDIES
In November 1994, the Virginia Transportation Research Council (VTRC) published a Technical Assistance Report entitled "Evaluation of the Use of Strobe Lights in the Red Lens of Traffic Signals." This paper, which subsequently was published in the Transportation Research Record, evaluated the effectiveness of strobes in red lenses at 22 intersections under Virginia Department of Transportation (VDOT) jurisdiction.1 The study of VDOTs practices focused on its use of strobes for:
* Areas with a high truck volume and high speed;
* Areas with a high accident rate;
* Areas with road geometries, especially grades (downgrade), horizontal curves and other features, that result in limited sight distance; and
* Isolated intersections where a signal is unexpected.
The VTRC study did not evaluate the effectives of strobes on improving violation rates at pedestrian signals. However, several of the study's findings were considered relevant in the context of Somerville's application of the technology.
VTRC found that there was no statistical evidence to suggest that strobes were effective in reducing accidents. It also found that that willful, defiant behavior is not likely to be affected by the presence of a strobe light to improve driver complaints with signal indications.
The authors set out to determine if the failure to obey red indications at pedestrian traffic signals was a result of this sort of intentional disobedience or an absence of awareness.
The City of Somerville is a dense urban municipality to the west-northwest of Boston, MA. It is bordered by Boston (Charlestown) to the east, the Mystic River to the northeast, the city of Medford to the north, the Town of Arlington to the west and the city of Cambridge to the south. The city is very small-only 4.1 square miles-but boasts a population of just under 80,000 residents.
More than 72 percent of the developed land in the city is dedicated to residential use. With a density of almost 20,000 residents per square mile, a large percentage of the population walk to recreational and business destinations in the immediate area of their homes. Most of the business activity is concentrated in neighborhood squares, which allow for local businesses to thrive on pedestrian activity. Many residents walk to public transit stations and stops, as shown in Figure 1.
The city is not without its share of vehicular activity. More than 55,000 vehicles are registered to owners residing in the city (a density of more than 13,000 vehicles per square mile). In addition, all of the major arterials in the City carry more than 20,000 vehicles per day. Much of this is daily commuter traffic traversing the city from the western suburbs to Boston.
In this setting, the City has experimented with a variety of traffic control devices to improve pedestrian safety. Previous experimental traffic control technologies included in-pavement lighting systems; flexible cone systems equipped with signage; and extensive use of new fluorescent yellow green signage in very high intensity sheeting. These systems have been successful in the past. However, they were not applicable to the problem of violation rates at pedestrian crossings that already were signalized.
The City also considered other measures to address the problems at these locations, including traffic calming measures such as raised crosswalks. However, raised crosswalks were eliminated from consideration because they would violate the City's Neighborhood Traffic Management and Traffic Calming Regulation. Moreover, it did not seem likely that raised crosswalks would improve violation rates (as opposed to reducing vehicular speeds).
The City's policy is not to install traffic calming devices on arterial roadways. This policy is consistent with the policies of the Massachusetts Highway Department. In addition, the policy reflects a philosophy of keeping regional traffic on arterial roadways at a speed appropriate to the type of roadway.
THE STROBE SYSTEM
The City identified the strobe system as a potential means of addressing the violation rate. The strobe system is a device that is mounted inside the red lens of a signal housing. When the signalized approach to be controlled enters the red interval, the system emits a series of strobe pulses of white light. The strobes are actuated by the change from the dwell phase of an arterial to the clearance intervals (yellow and red) while the signal is cycling to a pedestrian phase. The intent of the device is to alert motorists that there has been a change in right of way.
The City determined that one cause of the high violation rate could be the infrequency of the change in assigned right-of-way during off-peak times. At pedestrian-actuated signals, the traffic signal does not cycle on a routine basis. If the violation rate problem were caused by inattentiveness or awareness, the strobe could address this cause.
However, if the violations were intentional (such as motorists who were aware that the indications had changed but were proceeding through the crossing anyway), enforcement would be the way to address the problem.
This is heightened by the fact that the pedestrian signals have a high demand at certain times of the day, such as commuter hours and lunch hours. During all other times of the day, the pedestrian traffic signals have a lower demand rate. As a result, motorists become conditioned to ignore the traffic signal indications because they are almost always green at the time they approach the crossing.
The strobe system had the potential to increase motorists' awareness of the changing right of way by providing them with a completely new indication. The City hoped this new indication (the strobe to accompany the red indication) would cause motorists to become more alert and attentive to the pedestrian phase. If the strobe system were successful, the City would be able to reduce the violation rate and improve pedestrian safety. The strobe is shown in Figure 2.
The City selected two locations for testing the strobe system. The first, Holland Street at Hodgkins Park, is a mid-block pedestrian crossing. The second, Somerville Avenue at Beech Street, is a signalized intersection where an arterial intersects a local roadway. Both locations are at the entry to city-owned parks used by schoolchildren during daylight hours. Both locations met MUTCD traffic control signal warrants, specifically, warrant 4, "Pedestrian Volume."
Both tested locations were equipped with traffic signals in which all of the signal heads were post mounted. All of the signal heads had been equipped with light-emitting diode (LED) lenses, in part, to improve visibility. No sight distance constraints or other factors influenced compliance with red indications. Signal visibility was good at both locations and met the cone of vision requirements specified in MUTCD. Other than the fact that one location was at an intersection and another was at a mid-block crossing, the locations were similar in pattern of use, traffic volumes and urban environment.
It is important to note that the signal cycles used at these two locations were different. At the Holland Street location, the arterial dwelled on a green indication. When the pedestrian button was pushed and the pedestrian phase was actuated, the arterial indications proceeded to the standard clearance intervals (yellow and then red). The arterial indications remained red while the pedestrian indications (WALK and DONT WALK) were displayed. At the conclusion of the pedestrian phase, the arterial once again displayed a green indication.
In contrast, at the Somerville Avenue location, the arterial roadway displayed a flashing yellow indication during the dwell period. The local street displayed a flashing red indication. When the pedestrian phase was actuated, the arterial indications proceeded to solid yellow and then solid red displays. Pedestrians then received standard WALK and DON'T WALK indications. At the conclusion of the pedestrian phase, the arterial and the local street returned to dwell indications (flashing yellow and red, respectively).
These two locations were selected due to their documented violation rate. The violation rate had been documented over time as a result of complaints from residents. The proximity of parks and the propensity for school-age children to visit the parks was a leading cause of residents' complaints. Figures 3 and 4 show Somerville Avenue at Beech Street and Holland Street at Hodgkins Park, respectively.
The violation rate was measured for a one-hour period before and after the strobe installations. The morning peak hour was selected for measuring violations due to the high concentration of pedestrian activity at both locations during this time. During morning peak hours, commuter pedestrians use these crosswalks to travel to transit stops (both subway and bus stops).
The violation rate was determined by observing the number of vehicles that went through the stop line and crossed the intersection after the red indication had been displayed to the arterial roadway. Each vehicle that did so was labeled a violation. The violation rate then was calculated as the number of times at least one vehicle per pedestrian phase was in violation.
For the purpose of this study, bicycles were counted as vehicles because, under Massachusetts General Law, bicyclists are obliged to stop at traffic signals if so indicated. Moreover, the arterials at both locations have heavy bicycle traffic because they serve as commuter bicycle routes.
This heavy bicycle traffic has presented serious safety problems to pedestrians in the past, including severe bicycle-pedestrian collisions and injuries. Bicycle-pedestrian conflicts have been the source of significant local discussion and the focus of ongoing local legislative efforts to address these concerns.
PRE-INSTALLATION VIOLATION RATES
For a statistically valid comparison sample, the City measured the violation rate at each location just prior to the installation of the strobes. The pre-installation violation rate was documented in the month of July. There were no exceptional weather events on the days the violation rates were documented. The field data are provided in Table 1.
At Holland Street at Hodgkins Park, the documented pre-installation violation rate was 26.2 percent. The pedestrian phase was actuated a total of 42 times in the morning peak hour. During this same time period, at least one vehicle violated the red display during 11 of the pedestrian phases.
At Somerville Avenue at Beech Street, the pre-installation violation rate was calculated to be 34.8 percent. Of 46 pedestrian phases in the morning peak hour, there were 16 violations. This represented a significantly higher violation rate than the first location.
POST-INSTALLATION VIOLATION RATES
The post-installation violation rate was documented in the month of April. Nine months had lapsed since the installation of the strobes. Again, there were no exceptional weather events on the days the violation rates were documented.
The City determined that adequate time had passed to allow for the normalization of motorists' responses since the installation of the strobe systems. The City made a conscious decision to conduct the post-installation violation rate study after a significant amount of time to ascertain the long-term value of having the devices installed, as opposed to the short-term shock value of the devices as might be perceived on the day after installation. In addition, data were required to inform decisions regarding future capital purchases. The field data are presented in Table 2.
At the Holland Street at Hodgkins Park location, the documented post-installation violation rate was still 26.2 percent. Once again, the pedestrian phase was actuated a total of 42 times and at least one vehicle violated the red display during 11 of the pedestrian phases. This represented no change in the violation rate as a result of the strobe system installation.
The results of the installation at the other location, Somerville Avenue at Beech Street, were even less encouraging. The violation rate at this location increased from 34.8 percent to 36.9 percent. Of 46 pedestrian actuations, there were 17 violations in the morning peak hour after the installation of the strobe system. Interestingly, the total number of violations was reduced from 26 to 20 when the strobe was installed. However, the City felt that the primary objective of overall compliance-to improve pedestrian safety-had not been met.
A result of 0.05 would indicate that the results did have a statistical significance. Although, as indicated above, there was no documented improvement at all for either location, there also was no statistical significance in the change in violation rates.
For the intersection of Holland Street at Hodgkins Park, the pre- and post-violation rates were exactly the same. The numerator in the Z Test equation would be zero. Therefore, no statistical significance is indicated.
For the intersection of Somerville Avenue at Beech Street, there was an increase in violation rates from 34.8 percent to 36.9 percent. This would result in a Z of 0.029, which is less than 0.05. Accordingly, no statistical significance can be assigned to the increase in violation rates at this location.
The City measured the violation rate of motorists during a one-hour period before and after the installation of the strobes. Each car that was observed going through the red indication was documented as a violator. Under the test conditions, the strobes did not produce any improvement in the violation rate.
However, the City did receive extensive complaints regarding the strobes from adjacent residents. Residents felt that the strobes were too bright. Residents in close proximity to the tested locations complained that the strobes would keep them awake during hours of darkness.
The City concluded that the strobes did not have the desired effect of raising motorists' awareness of the change in right-of-way at pedestrian traffic signals. The inability of the strobe system to improve the compliance rate at the studied intersections indicates that driver inattentiveness may not be the cause of the red signal violations at pedestrian crossings. It is possible that the violation problem is a direct result of motorists making a conscious decision to violate the red indication in spite of the pedestrian phase.
The results indicate that more intensive enforcement may be a better solution to obtaining compliance. Depending on the availability of funding, the City may consider a next phase to determine the effectiveness of enforcement to reduce violations. As a result of this study and complaints from residents, the strobe system was removed from the Hodgkins Park location shortly after the study was conducted.
Based on the study, the City was not able to conclude whether or not the strobes would be effective in other applications. It is conceivable that the strobes might be effective at improving compliance at intersections where there is a problem with vehicle-vehicle crashes-perhaps in a rural or suburban setting. More research is needed to determine when strobes are effective and in what applications.
© 2005 Institute of Transportation Engineers Provided by ProQuest LLC. All Rights Reserved.