Spotlighting Speed Feedback Signs
An FHWA study links dynamic messages to a reduction in roadway departures on two-lane rural curves that have high crash histories
Roadway departures are a significant safety concern on U.S. roads. According to the latest data from the National Highway Traffic Safety Administration’s Fatality Analysis Reporting System, roadway departures continue to account for more than half of U.S. highway fatalities annually and nearly 40 percent of serious injuries.
Most departure crashes occur on rural two-lane roadways, with a disproportionate number taking place on horizontal curves. The average crash rate at horizontal curves is about three times that of other types of highway segments. These curves, which change the alignment or direction of the road, are associated with more than 25 percent of fatal crashes, and the majority of those fatalities are associated with roadway departures. In addition, about 75 percent of curve-related fatal crashes involve single vehicles leaving the roadway.
“The reduction of roadway departures must be a major emphasis if we want to significantly reduce fatalities and serious injuries in the United States,” says Monique Evans, director of the Federal Highway Administration’s Office of Safety Research and Development.
Not surprisingly, speed is a factor in whether drivers negotiate curves successfully. Dynamic speed feedback signs are one type of traffic control device that State departments of transportation use to reduce vehicle speeds, and therefore crashes, by giving drivers who are traveling over the posted or advisory speed a targeted message such as “YOUR SPEED XX” or “SLOW DOWN.”
These sign systems include a speed-measuring device, which consists of loop detectors or radar, and a message sign that displays feedback to those drivers who exceed a predetermined speed threshold. The feedback can include displaying the driver’s actual speed, showing a message such as SLOW DOWN, or activating some warning device, such as beacons or a curve warning sign.
To better understand the effectiveness of speed feedback signs in reducing speeds on curves, the Center for Transportation Research and Education at Iowa State University conducted a national field evaluation of the signs at horizontal curves on rural two-lane roadways. The study is described in a January 2015 report, Evaluation of Dynamic Speed Feedback Signs on Curves: A National Demonstration Project (FHWA-HRT-14-020).
Sponsors of the project included FHWA, the Midwest Transportation Center at Iowa State University, the Iowa Department of Transportation, the Iowa Highway Research Board, and the Texas Department of Transportation. In addition, the Texas A&M Transportation Institute and Portland State University were partners in the research. Here’s how the researchers did the study and what they found.
Selection of Sites
Seven States participated in the field evaluation: Arizona, Florida, Iowa, Ohio, Oregon, Texas, and Washington. The researchers asked each State DOT or corresponding local agency to identify at least 20 high-crash curve sites on rural two-lane roadways. The research team defined “rural” as 1mile (1.6 kilometers) or more outside an incorporated area.
The study started in 2007 and concluded in 2013. The team required that, during the 2-year evaluation period foreach project site, the State DOTs or corresponding local agencies would schedule no rehabilitation or reconstruction activities that would change the geometry of the roadways under consideration. Nor were the DOTs to have conducted any geometric or cross-section changes for 3 years prior to the beginning of the study. In addition to these requirements, the posted speed limit on the preceding tangent section of road had to be 50 miles per hour (mi/h) (80 kilometers per hour, km/h) or greater.
The research team also asked each DOT to provide data on crash frequency, traffic volume (annual average daily traffic and percent of trucks), geometry (including lane and shoulder width), and the posted and advisory speed limits. The researchers ranked the sites in each State by the number of crashes. They also counted sites above a predetermined threshold as high-crash locations and included them on a list for site visits.
The team conducted a preliminary speed study using a radar gun at each site to determine whether a speeding problem existed, and those findings led to picking a final list of sites. Overall, the researchers selected 22 treatment sites and 46 control sites. They used the control sites only for crash analysis. For each treatment site, they randomly assigned one of the various types of speed feedback signs for the States to implement.
Selection of Signs
The selection of systems focused on what a sign can display. The most common sign simply shows a vehicle’s speed when it exceeds a set threshold. This kind of sign also can activate a flashing beacon. Another type of sign can show a static message, such as SLOW DOWN or TOO FAST. More complex signs display unique messages, limited only by the number of alphanumeric characters the sign can show.
The research team developed the following set of minimum criteria to guide the final selection of the type of speed feedback sign:
- Can be mounted permanently on a standard wooden or metal pole.
- Can display a warning or a simple message (for example, TOO FAST or XX mi/h).
- Is durable enough to survive the 2-year study period and perform in various climates.
- Has self-contained power (for example, alternating current or solar).
- Costs less than $10,000 per sign (including installation, support, and maintenance).
- Meets all applicable Manual on Uniform Traffic Control Devices requirements or is capable of being approved under MUTCD.
- Provides repeatable and accurate speed measurements.
- Projects a clear, bright, nonglare message that motorists can read easily.
Location Of The Study’s Treatment Sites
|State||ID||Location||Posted Speed (mi/h)||Advisory Speed (mi/h)||ADT*||Crashes/ Year||Number of Control Sites|
|AZ||2||SR 95||45 NB**
|FL||6||3 SR 267||55||None||4,300||2.6||5|
|8||3 SR 20||55||None||5,400||2.2|
|32||2 SR 20||55||45||8,100||1.0|
Truck 60 day
Truck 55 night
Truck 70 day
Truck 65 night
Truck 60 day
Truck 55 night
|18||SR 7||50||40 NB/35 SB||1,976||3.3|
|Average for Treatment Sites||3,565||2.2|
|Average for Crash Control Sites||3,362||1.8|
*ADT = Average daily traffic.
**NB = Northbound, SB = Southbound.
***Crashes were over several curves.
For the first message type, the team selected the dynamic display of YOUR SPEED XX or SPEED LIMIT XX, with the message determined by the speed threshold.
For the second message type, the researchers chose a sign that displays an advance curve warning symbol. When activated, the sign displays a standard curve warning symbol as specified by the MUTCD and the words SLOW DOWN.The sign also has two lights on the top and bottom that blink in an alternating pattern while the curve warning symbol is displayed. Popular in Europe, this message type has had limited application in the United States.
A commonly accepted view is that speed displays should have an upper speed threshold above which they no longer display speed, so that drivers do not “test” the sign at unsafe speeds. The researchers settled upon 20 mi/h (32 km/h) over the posted speed limit as the upper threshold. For each site, they also selected a unique lower threshold--the lowest speed at which the speed display would be activated.
Based on the upper and lower speed thresholds, the sign face for the speed display showed the following for each situation (driver speed was measured at the point of curvature):
- Blank sign: When a curve advisory sign was present, no message was given for drivers who were traveling at or below the advisory speed limit plus 5 mi/h (8 km/h). When no advisory sign was present, the sign was blank for drivers traveling at or below the posted speed plus 5 mi/h (8 km/h).
- YOUR SPEED followed by the vehicle’s speed XX in miles per hour: When drivers were traveling 5 mi/h (8 km/h) or more over the advisory speed if present or posted speed limit with no advisory speed, up to 20 mi/h (32 km/h) over the posted speed limit.
- SPEED LIMIT XX with the actual speed limit displayed: When drivers were traveling 20 mi/h (32 km/h) or more over the posted speed limit.
Based on the upper and lower speed thresholds, the sign face for the curve warning display showed the following for each situation:
- Blank sign: When a curve advisory sign was present, no message was given for drivers who were traveling at or below the advisory speed plus 5 mi/h (8 km/h). When no advisory sign was present, the sign was blank for drivers traveling at or below the posted speed plus 5 mi/h (8 km/h).
- Curve warning sign plus alternating lights and the words SLOW DOWN: When drivers were traveling 5 mi/h (8 km/h) or more over the advisory speed if present or posted speed limit with no advisory speed.
Methodology of the Study
The researchers conducted a full-scale, before-and-after speed study. They collected speed and volume data at the 22 test sites for 2 days about 1 month before the State DOTs installed the signs, and again about 1 month, 1 year, and 2 years after installation. Altogether, the research team collected data for 2 years to determine whether the effectiveness of the speed feedback signs decreases over time as drivers habituate to the signs.
The researchers used pneumatic road tubes and counters for data collection. The advantage of the road tubes is that they are fairly accurate, can collect individual vehicle speeds (enabling spot-checking of the data), are relatively low cost, and can be placed without cutting the pavement. The team also decided they are practical because other technologies, such as video, are more cumbersome, less accurate, or more expensive.
For each data collection period, the counters recorded time, vehicle speed, and vehicle class for individual vehicles. The team calculated other metrics, such as volume, headway, and average speed, from the data collected by the counters.
At each site, the team placed the speed-activated feedback sign near the point of curvature for one direction of travel. For each data collection period, the team collected data from the road tubes approximately 0.5 mile (0.8 kilometer) upstream of the point of curvature, at the point of curvature, and at the center of the curve.
Each collection period consisted of 48 hours and took place from Mondays through Fridays. The researchers chose the 48-hour period to ensure that a large sample size would result and that the data would not be biased toward a specific time of day.
The team calculated several speed metrics for the direction of travel toward the sign. The metrics included average speed, standard deviation, 50th percentile speed, 85th percentile speed, and the number of vehicles traveling 5, 10, 15, or 20 mi/h (8, 16, 24, or 32 km/h) over the posted or advisory speed limit. The team expected the signs to affect driver behavior shortly upstream of the curve and throughout it. As a result, the researchers evaluated the effectiveness of the signs by the change in speed at the point of curvature and at the curve’s center.
Average Change In Speeds At The Point Of Curvature
|1 Month||12 Months||24 Months|
|All Sites||Curve Sign Sites||Speed Sign Sites||All Sites||Curve Sign Sites||Speed Sign Sites||All Sites||Curve Sign Sites||Speed Sign Sites|
|Average Mean Speed (mi/h)||-1.82||-1.68||-1.95||-2.57||-2.47||-2.66||-1.97||-1.99||-1.96|
|Average 85th Percentile Speed (mi/h)||-2.19||-1.90||-2.45||-2.86||-2.40||-2.70||-2.17||-2.00||-2.30|
|Average change in fraction of vehicles exceeding posted or advisory speed by||5 mi/h||-11.8%||-9.8%||-13.7%||-18.6%||-22.1%||-15.0%||-19.8%||-27.1%||-13.3%|
Point of Curvature. The team examined the change in speed metrics averaged over all treatment sites at the point of curvature. The speed data facilitated determining the difference between the before-period speeds (1 month before sign installation) and the after-period speeds (1, 12, and 24 months after sign installation).
Comparing the difference, the average mean speed for all sites was reduced 1.82 mi/h (2.93 km/h) after 1 month, 2.57 mi/h (4.14 km/h) after 12 months, and 1.97 mi/h (3.17 km/h) after 24 months. Changes in the 85th percentile speeds were similar, with reductions of 2.19, 2.86, and 2.17 mi/h (3.52, 4.60, and 3.49 km/h) after 1, 12, and 24 months, respectively.
Strong decreases also occurred in the average percent of change in the fraction of vehicles exceeding posted or advisory speeds. After 1 month, the number of vehicles traveling 5, 10, 15, and 20 mi/h (8, 16, 24, and 32 km/h) or more over the posted or advisory speed limits decreased 11.8, 29.9, 36.3, and 28.5 percent, respectively. Similar decreases were recorded after 12 and 24 months. The highest decrease (49.8 percent) came from vehicles going 20 mi/h (32 km/h) or more over the posted or advisory limit at the 12-month mark.
The team also tabulated and compared data by sign type. In general, the researchers noted larger decreases for the speed signs than for the curve signs, although the differences were not statistically significant.
Center of Curve. Similar to the data from the point of curvature, the average change in mean speed for all sites at the center of the curve also decreased. Reductions of 2.08, 1.65, and 1.76 mi/h (3.35, 2.66, and 2.83 km/h) were recorded after 1, 12, and 24 months, respectively. For the 85th percentile speed at the center of the curve, the average decreases were 2.52, 1.55, and 1.89 mi/h (4.06, 2.49, and 3.04 km/h) after 1 month, 1 year, and 2 years, respectively.
The average percent change in the fraction of vehicles exceeding the posted or advisory speed tended to have greater decreases at the center of the curve when compared to the point of the curve. After the first month, the fraction of vehicles exceeding posted or advisory speeds at 5, 10, 15, and 20 mi/h (8, 16, 24, and 32 km/h) decreased 28, 42, 57, and 31 percent, respectively. After 12 months, the decreases ranged from 14 to 37 percent, and after 2 years ranged from 26 to 44 percent.
These data anecdotally suggest that the signs remained effective over time. However, the researchers used a statistical test to determine whether the differences were due to the treatment for the 1-, 12-, and 24-month-after periods. The analysis indicated no statistically significant differences among changes in mean speeds at the point of curvature and the center of the curve for any of the time periods. This finding suggests that the signs might have a long-term impact on reducing speeds.
The researchers modeled the crashes by quarter rather than by year. By using quarters, they could exclude from the analysis the quarter in which installation occurred without having to exclude the entire installation year. In addition, the signs stopped functioning at several sites for various periods, so the quarter in which the signs were nonfunctional also could be excluded from the analysis without discarding the entire year’s data.
Total crashes in both directions decreased by 0.08 crashes per quarter for the control sites, while crashes per quarter at the treatment sites decreased by 0.22 (17-percent reduction compared to 40-percent reduction). Single vehicle crashes for both directions decreased by 0.07 crashes per quarter at the control sites and by 0.21 at the treatment sites (19-percent decrease compared to 47-percent decrease). Reductions at treatment sites were 2.75 and 3.0 times greater than at control sites. Fluctuations in speed at the control sites could be due to a number of factors that were not known and could not be controlled. For instance, short-term maintenance in the vicinity of one of the curves could have impacted speeds. Every attempt was made to collect data under similar circumstances, but it was impossible to be aware of every situation that might have impacted speed.
Total crashes in the direction of the outside of the curve increased by 0.02 crash per quarter for control sites and decreased by 0.12 crash per quarter in the direction of the sign for the treatment sites (9-percent increase compared to 35-percent decrease). Similarly, single vehicle crashes decreased by 0.01 crash per quarter at the control sites compared with a decrease of 0.14 at treatment sites (4-percent decrease compared to 49-percent decrease). Reductions at treatment sites were 6 to 14 times greater than at control sites.
The results show that a much greater decrease in crashes per quarter occurred for treatment sites compared to control sites. However, caution should be used in applying the results for the simple analysis because the data are not adjusted to account for the seasons, and more quarters of a particular season might have been present in the before period than the after period.
The team also conducted a before-and-after analysis using a full Bayes model to develop crash modification factors. The model accounts for trends in the data that cannot be accounted for using other models. For instance, crashes might increase or decrease at a treatment site due to random fluctuations in the data not related to the treatment. Full Bayes is able to account for this phenomenon.
Average Change In Speeds At The Center Of The Curve
|1 Month||12 Months||24 Months|
|All Sites||Curve Sign Sites||Speed Sign Sites||All Sites||Curve Sign Sites||Speed Sign Sites||All Sites||Curve Sign Sites||Speed Sign Sites|
|Avg. Mean Speed (mi/h)||-2.08||-2.01||-2.15||-1.65||-1.47||-1.84||-1.76||-1.46||-2.00|
|Avg. 85th Percentile Speed (mi/h)||-2.52||-2.50||-2.55||-1.55||-0.82||-2.27||-1.89||-1.25||-2.40|
|Average change in fraction of vehicles exceeding posted by||5 mi/h||28%||28%||27%||20%||21%||18%||26%||30%||23%|
The researchers developed predictive models using data from control sites for all periods and before data for treatment sites. The models accounted for season, differences in the length of sites, and multiple measures at the same site. The team then used the models to calculate the number of crashes for the after period for treatment sites that would have been expected had no treatment been applied. They also calculated crash modification factors by dividing the observed crashes by the predicted values.
The model indicated that expected total crashes for both directions would decrease by 5 percent (0.95 crash modification factor) with installation of the speed feedback signs. The team expected total crashes in the direction of the signs to decrease by 7 percent (0.93 crash modification factor). Both figures are statistically significant.
The model indicated that expected single vehicle crashes in both directions would decrease by 5 percent, and single vehicle crashes in the direction of the sign to decrease by 5 percent as well. Both changes are statistically significant.
The goal of this national demonstration project was to evaluate the effectiveness of two types of speed feedback signs in reducing speed and crashes on rural horizontal curves. If the signs were effective, that would provide traffic safety engineers with additional tools to improve roadway safety.
The results indicate that the systems are reasonably effective in reducing both vehicle speeds and crashes. And, it is noteworthy, the reductions were maintained for more than 2 years, indicating drivers did not habituate to the dynamic signs, although the study did not specifically look at this.
On average, most sites had decreases in mean speeds, with decreases up to 10.9 mi/h (17.5 km/h) noted for both the point of curvature and center of curve. Most sites experienced changes in the 85th percentile speed of 3 mi/h (4.8 km/h) or more at the point of curvature, with the majority of sites having a decrease of 2 mi/h (3.2 km/h) at the center of the curve.
Large reductions in the number of vehicles traveling over the posted or advisory speeds occurred for all of the after periods at the beginning and center of the curves, indicating that the signs were effective in reducing high-end speeds, as well as average and 85th percentile speeds.
“In the right place and for the right situation, dynamic speed feedback signs are a good option to consider to reduce vehicle speeds,” says Sandra Larson, systems operations bureau director, highway division, Iowa DOT. “We have used these signs effectively for interstate and noninterstate work zones, rural expressway intersections where there is a speed limit reduction, school zones, and with pavement painting operations.”
Abdul Zineddin, Ph.D., is a transportation specialist with FHWA’s Office of Safety Research and Development at the Turner-Fairbank Highway Research Center. He oversees the speed management research program. Zineddin holds bachelor of science, master of engineering, and doctorate degrees in civil engineering with two graduate minors in human factors and statistics from Pennsylvania State University.
Shauna Hallmark, Ph.D., is a professor of civil engineering at Iowa State University and is director of Iowa State’s Institute for Transportation. She holds a Ph.D. from Georgia Institute of Technology, an M.S. from Utah State University, and a B.S. from Brigham Young University, all in civil engineering.
Omar Smadi, Ph.D., is an associate professor of civil engineering at Iowa State University. He also is director of the Roadway Infrastructure Management & Operations Systems program and is a research scientist at the Center for Transportation Research and Education. He holds a Ph.D. and an M.S. in transportation engineering from Iowa State.
Neal Hawkins is the director of the Center for Transportation Research and Education and also the Center for Weather Impacts on Mobility and Safety at Iowa State University. He has an M.S. from Iowa State and a B.S. from the University of Oklahoma in civil engineering.
For more information, contact Abdul Zineddin at 202–493–3288 or firstname.lastname@example.org or Shauna Hallmark at 515–294–5249 or email@example.com, or see Evaluation of Dynamic Speed Feedback Signs on Curves: A National Demonstration Project at www.fhwa.dot.gov/publications/research/safety/14020/index.cfm.