Work Zone Speed Management (2015)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

27 chapter three ENGINEERING TECHNOLOGIES INTRODUCTION This chapter discusses technologies—predominantly electronic technologies—intended to enhance driver awareness and/or compliance with work zone speed limits. CHANGEABLE SPEED LIMIT SIGNS FOR WORK ZONES Changeable speed limit (CSL) signs, such as those shown in Figure 10, appear to be gaining popularity in jurisdictions where work zone speed limits are based on criteria that change frequently, such as the presence or absence of workers. CSL sys- tems typically combine an ordinary sign with a digital display panel that can be toggled between two or more values, such as a workers-present and a workers-not-present speed limit. Some systems allow the change to be made remotely. (Similar equip- ment is sometimes used at school zones to change the speed limit when children are present.) Some examples of the use of this type of equipment are as follows: • VDOT reports the use of permanently mounted CSL signage at tunnel approaches near Hampton Roads (D. Rush, VDOT, personal communication, 2014). The ordi- nary speed limit in this section is 45 mph; however, a CSL system is activated (along with other pretunnel signage) to reduce the speed limit to 25 mph during recurring tunnel cleaning and maintenance. • In 2008, Minnesota DOT (MnDOT) issued a Guideline for Intelligent Work Zone System Selection, which pro- vides operational notes for the use of CSL signs (MnDOT 2008). • In 1999–2000, MnDOT conducted a small-scale demon- stration project intended to make it easier to change work zone speed limits on high-volume urban freeways (FHWA 2002). Portable speed limit signs mounted on U-channel supports were used for the project. Each sign was equipped with a two-digit digital display panel. When construction workers were not present, a 65 mph the speed limit was displayed; when construction workers arrived at the site, a designated worker changed the speed limit to 45 mph. Only two studies were found comparing driver compli- ance with CSL to compliance with ordinary fixed work zone speed limit signs: 1. A study of CSL-type signage displaying mandatory work zone speed limits on I-80 near the Wyoming– Utah border concluded that at night under uncongested conditions and a 65 mph speed limit, the CSL-type signs resulted in hourly average speeds that were somewhat lower than the standard static speed limits signs previ- ously used at the site. Speed variation was also reduced. The digital signs were highly conspicuous at night. During daylight hours the travel speeds appeared to be more consistent when the digital signs were used, but the results were difficult to interpret because of work zone congestion, traffic volume fluctuations, and other factors (Rifkin et al. 2008; McMurtry et al. 2009). 2. A study of advisory CSL-type signage at three Missouri work zones found that average traffic speeds were gen- erally 1.5 to 2.0 mph lower with the digital signs than with static signs; 85th percentile speeds and speed vari- ance were also reduced (Edara et al. 2013). Additional evaluations may be necessary to understand more fully whether the differences found in these two studies are attributable to the greater conspicuity of the digital signs or simply novelty effects that will diminish with more wide- spread use of CSL. VARIABLE SPEED LIMITS In contrast to the relative simplicity of CSL signs, large-scale variable speed limit (VSL) systems modify the speed limit in response to real-time changes in traffic conditions. VSL systems typically involve corridor-level deployment of a series of speed sensors, digital signs, and automated control algorithms. A significant level of planning, system engineer- ing, and system integration may be required. The require- ments vary with the VSL system design and the extent of existing Intelligent Transportation System (ITS) infrastruc- ture at the site. Administrative procedures for approving the system-recommended speed limit and communicating it to law enforcement are also a consideration, particularly if the limit is mandatory rather than advisory. Typically the objective of VSL is to delay the onset of congestion by slowing traffic gradually as it approaches a bottleneck. Consequently, the goal is usually to improve the temporal and spatial uniformity of traffic speeds, rather than simply reducing speeds. This characteristic makes assess- ment of VSL systems more difficult than evaluation of most other work zone speed management techniques. Some of

28 found that the system was effective in reducing the lon- gitudinal speed differences along the work zone area dur- ing the weekday morning peak period. Traffic throughput also increased (Kwon et al. 2007). • As discussed in Case Example 2, effects of the 2008–2010 VSL deployment on the Capitol Beltway in Virginia were inconclusive, in part because of driver noncompliance with the variable speed limits. • A 2013 Missouri study of a VASL system deployed at an uncongested site on I-270 on the outskirts of St. Louis found that average freeway speeds were reduced by 2.2 mph, but the standard deviation of speeds increased by 4.4 mph, most likely because some drivers complied with the advisory limits and others did not. Compliance rates at the uncongested site were much higher with the VASL system than without it. Results at a nearby con- gested site were mixed (Edara et al. 2013). Case Example 2: Variable Speed Limit Pilot Project on the Woodrow Wilson Bridge** The Capital Beltway (I-95/495) encircles the Washington, D.C., area. The Woodrow Wilson Bridge (WWB) carries the Beltway across the Potomac River at the Virginia/Maryland border near Alexandria. In late July 2008, VDOT introduced a Variable Speed Limits (VSL) test program as part of a mega-project to reconstruct the bridge. The initial deployment was intended to assist drivers approaching the reconstruction of the Telegraph Road Interchange near the western approach to the bridge. The Telegraph Road project was a $240 mil- lion element of the larger WWB mega-project. The cost of deploy- ing the temporary VSL system was $3.2 million. Three to six travel lanes were available in each direction through the work area, and the work zone was approximately five miles long in each direction. To minimize impact on the already heavily congested roadway the lane closures were done during overnight hours, typically with one or two lanes closed. Given the high traffic volumes and the complexity of the work zone, VSL was deployed as a congestion management tool. The intent was to enable more vehicles to pass safely through the constricted area by gradually regulating the speed of vehicles approaching a reduction in lane capacity. Other ITS applications were also incorporated into the transportation management plan, including closed-circuit television cameras to monitor traffic conditions and Portable Changeable Message Signs (PCMS) to provide real-time travel delay informa- tion to motorists and enable drivers to modify travel routes, times, and modes. these challenges are discussed in Case Example 2, which describes a large-scale Virginia deployment. A number of relatively theoretical research papers have explored the speed harmonization and capacity benefits of work zone VSL systems; however, the primary field applica- tion for VSL in the United States has been to reduce the risk of back-of-queue crashes at work zone approaches. These queue management applications are beyond the scope of this syn- thesis report. The theoretical papers often assume low latency (short time lags between selection of a new speed limit and its display to traffic) and high levels of driver compliance, which have not always been the case in U.S. field deployments. A few field studies have examined the speed manage- ment aspects of work zone VSL systems. The small number of studies, technical differences between the systems, varia- tions in site conditions, and diversity of results make it diffi- cult to generalize about the systems’ effectiveness. Reported results include the following: • A study of VSL deployment on I-96 near Lansing, Michi- gan, found increased average speeds in the freeway work zone when the system was operational, primarily because during uncongested conditions the VSL was higher than the fixed work zone speed limits that had been in place before the system was activated. Effects on the estimated 85th percentile speed and speed variance were either undetectable or inconsistent (Lyles et al. 2004). • A study of a Variable Advisory Speed Limit (VASL) sys- tem deployed on I-494 in the Twin Cities (Minnesota) FIGURE 10 Trailer-mounted changeable speed limit (CSL) sign (MinnDOT 2008).

29 9 months of after data (May 2009 through January 2010) were used for this study. As a result of cost limitations, travel time for each segment was estimated by multiplying the average speed at each detector by the distance between detectors. Results were as follows: • Overall, there were no statistically significant changes in the various speed or capacity MOEs; however, some positive impact was indicated in the capacity of the roadway. • Travel times increased in the area upstream of the beginning of the bottleneck and significantly decreased for the remain- der of the work zone for the north/east direction of travel. An insignificant change in travel time was detected for the south/west direction of travel. • Anecdotal observations of queue lengths and associated delay time appeared slightly reduced from normal patterns. • Qualitative measures such as public outreach and commu- nity feedback were also assessed, but little feedback was received about VSL after the initial deployment. The deploy- ment and methods of use of PCMS received very positive feedback. The project study team reported several lessons learned: • The system performance and ability to evaluate the system would have been improved by using speed detectors that could record individual vehicle speeds instead of averaging all vehicle speeds over a period of time. • The speed detectors should be verified with other traffic data sources, such as speed radar guns, and system soft- ware calibrated regularly to account for changing roadway configurations. • Automated speed enforcement could have enhanced system performance by substantially reducing the number of high speed drivers. • PCMS at key locations are important to help drivers make diversion decisions. The project construction team evaluation concluded that suc- cessful VSL deployment as a work zone management tool would be viable if motorists are informed of real-time work zone condi- tions, automated speed limit enforcement is used, and if drivers have adequate time to use alternate routes. References: (Fudala and Fontaine 2010a, b; Nicholson et al. 2010; VDOT 2010). DYNAMIC SPEED FEEDBACK SIGNS Dynamic speed feedback signs (also known as dynamic speed signs, speed monitoring displays, speed indicator boards, radar speed reader boards, or radar speed display units), such as those shown in Figure 11, combine a static sign show- ing the regulatory or advisory speed limit with an adjacent digital panel that displays the observed speed of the near- est vehicle. Displaying actual speed to motorists as they approach a work zone (or proceed through the work zone) has become a useful speed management practice. The tech- nique has been successfully implemented using several differ- ent types of hardware. Although the topic has been extensively researched, a wide range of speed reduction results has been found in the field. The VSL deployment was intended to increase throughput traffic flow in a safe manner by reducing driver speed variation. Specially trained VSL officers housed at the project field office oversaw the system using roadway sensors and the other ITS hardware and systems on the project. The officers developed optimal speed limits using a vendor-supplied control algorithm that would attempt to keep traffic flowing by setting the range of speed limits from a minimum speed of 35 mph to a maximum speed of 55 mph. Drivers were informed of the varying speed limits, which changed in 5 to 10 mph increments, through the VMS along a seven-mile portion of I-95 between the Springfield Interchange and the Maryland shore. To allow sufficient time for law enforcement to be notified, the speed limit was not changed more than once every 20 minutes. The VSL system was initially deployed in July 2008, but only activated during temporary lane closures that occurred at night and weekends, and under certain “restricted” conditions that resulted in very limited use. In May 2009, the VSL was expanded to “full-time” use: the system was activated in a traffic-responsive mode during all peak congested periods, regardless of construction activity. The pilot program officially ended February 21, 2010; all VSL equipment was removed from the area and a fixed 55 mph speed limit was restored throughout the project corridor. A preliminary evaluation, done in 2009 by the Virginia Transporta- tion Research Council (VTRC), studied the initial 10-month restricted activation period. The evaluation produced inconclusive results, including no large changes in speed or queue length. Therefore, in May 2009, significant changes were made by extending the sys- tem to full-time use and the system vendor made changes in the control logic. After the VSL system operation changes were made, the VTRC researchers concluded it was not prudent to conduct a thorough empirical analysis of the system performance of the final sys- tem operation. Significant variations in the lane closure locations, demand volumes, and work activity made it difficult to perform before-and-after comparisons under similar conditions. In addi- tion, construction activities changed enough that there was never a clear one-to-one match between lane closures before and after the VSL was activated. Because of the inconclusive preliminary results of the field deployment of VSLs on the project, the researchers concentrated their efforts on using simulation modeling to gain a better understanding of the significance VSL system design has on operations. Using data collected during the second half of the deployment period, an in-depth report was prepared using a cali- brated VISSIM microsimulation model of the project area. The simulation provided an opportunity to examine a number of sys- tem configurations to assess how changes in system design and driver behavior might affect a variety of performance measures. The results indicated that VSL could provide substantial improve- ments in traffic operations, but only if demand did not exceed capacity by too large a margin, such as only reducing four lanes to two lanes versus four lanes to one lane. The simulation also dem- onstrated that algorithm design, driver compliance with the limits, and the location of the VSL signs all play important roles in opera- tional performance. A cost–benefit analysis was also conducted, which showed that VSL was most appropriate only for long-term work zone applications, because of the high cost of establishing the VSL system. The construction project team (including the VSL project manager) conducted a before-and-after study of the full-time VSL phase. Various measures of effectiveness (MOEs) were examined including speed, travel time, capacity, queues, and delays. Three months of baseline before data (February through April 2009) and

30 ment measures (such as additional active law enforcement patrols) are needed. Dynamic speed feedback can modestly reduce the vehicle speeds in the immediate vicinity of the display, with vari- ous research studies showing a range of speed reductions of approximately 1 to 8 mph in the area approaching lane reduc- tion tapers and 3 to 6 mph within the work zone, compared with conventional work zone treatments used in the before condition. The treatment was evaluated and found effective for both short-term work zones (maintenance operations) as well as long-term projects (Carlson et al. 2000; Maze et al. 2000; McCoy and Pesti 2002; Brewer et al. 2006; Hajbabaie et al. 2011). Speed reductions of 8 mph were reported when a radar- equipped portable changeable message board was used (Wang et al. 2003). There is no consensus on the long-term speed reduction from using the technique based on different studies. The research done by Wang et al. (2003) concluded that the effects remained valid three weeks after the installation. Conversely, a study by Meyer (2004) concluded there was a novelty effect. A somewhat more recent study concluded that, “After a long- term implementation, the DSD [dynamic speed display] had a slight impact on the daytime speed but significantly decreased both the average speed and percentage of speeding drivers during nighttime hours. The significant speed reduction by the DSD presented in previous studies was likely the result of a novelty effect. As observed in this study, the overall speed reduction lapsed with time” (Chen et al. 2007). Field experience indicates some potential limitations of the signs. For example, PennDOT prefers to display a slow down message instead of a numerical speed reading because some drivers attempt to see how large a number they can register on the sign (M. Briggs, PennDOT, personal communication, 2014). In 2014, the authors of this synthesis report routinely observed readings in the 59 to 63 mph range on dynamic speed feedback signs at a work zone on I-94 in Milwaukee County, Wisconsin, which had a speed limit of 50 mph and is used mainly by com- muters. In the multilane Milwaukee application it was often difficult to tell which vehicle’s speed was being displayed. Surveys and interviews with state DOT officials indicate varying levels of utilization of speed monitoring displays in the United States. In 2014, Illinois DOT began requiring con- tractors to furnish a dynamic speed feedback sign whenever workers are present and lanes are restricted by construction (IDOT 2014a). PORTABLE CHANGEABLE MESSAGE SIGNS WITH VEHICLE-ACTIVATED SPEED MESSAGES Trailer-mounted PCMS can be equipped with radar-activated changeable message displays that warn motorists they are traveling at an unsafe or undesired speed in a work zone. Work zone dynamic speed feedback systems typically uti- lize a trailer-mounted display panel for ease of installation and portability. Alternatively, a post-mounted system can be used, or the speed can be displayed on a multipurpose PCMS. The speed data are typically collected using a radar device. Depending on the design details of the system, observed speeds that exceed the speed limit are sometimes displayed in red or using flashing amber digits. A limitation of the system is that in multilane applications with moderate to heavy traffic it may be unclear which vehicle’s speed is being displayed. Although some automated speed enforcement systems incorporate a dynamic speed feedback display, standalone dynamic speed feedback signs are generally not used to enforce speed limits. Such signs serve solely as a reminder to the driver by displaying his or her actual speed in comparison with the posted speed limit. Unlike automated speed enforce- ment systems, no photographs are taken, nor are individual vehicles identified. Nevertheless, summary speed data are often recorded to assist law enforcement and the highway authority in determining whether additional speed manage- FIGURE 11 Dynamic speed feedback trailer. (Photo: Wikimedia Commons).

31 The units combined a standard PCMS, a radar speed data collection unit, and a set of blue and amber flashing lights— colors are associated with the California Highway Patrol. The flashing lights and a slow down message were acti- vated when a vehicle traveling 5 mph or 10 mph over the speed limit was detected (Figure 12). The lane closure alone without the trailer resulted in a reduction of average traffic speed by approximately 5 to 5.5 mph. The use of the light- augmented speed trailer by itself resulted in approximately 3 to 7 mph further reduction of the average traffic speed in the work zone beyond what was observed with the closure alone. Use of a highway patrol officer in a police vehicle in addition to the speed trailer resulted in approximately 5 to 9 mph further reduction of the average traffic speed in the work zone beyond what was observed with the closure alone. Some of the reductions associated with the device may have been novelty effects. AUGMENTED ENFORCEMENT SYSTEM An experimental augmented enforcement system was tested on a multilane divided rural highway in central California in 2012 (Chan et al. 2013). The system utilized Automated License Plate Recognition similar to the technology discussed in chapter seven to go beyond the capabilities of traditional speed feedback displays. As shown in Figures 13 and 14, the system combined the license plate recognition capabilities of a speed camera with a PCMS to display the speed and plate number of each vehicle that was exceeding the work zone speed limit. In addition, the system demonstrated the feasibility of using Dedicated Short Range Communications (DSRC) to relay the plate number and observed speed to a California Highway Patrol officer or workers downstream. The research prototype did not include any automated func- tions related to issuing citations. Speed data were collected using iCone drums—proprietary radar and telecommunica- tion devices concealed inside ordinary orange traffic con- trol drums. Results from an 8-week field test showed a 10% Typically, the unit is programmed to display an anti-speeding message when a vehicle exceeding a threshold speed (typi- cally 3 mph above the speed limit) approaches. Common mes- sages that are displayed when the threshold speed is exceeded include: • you are speeding, slow down • high speed, slow down • reduce speed in work zone • excessive speed, slow down. Several studies have been conducted on the effectiveness of these four sign messages and all were found to reduce the number of vehicles speeding adjacent to the PCMS, with a speed reduction of 7 mph near the sign reported in a Georgia study (Fitzsimmons et al. 2009). The reduction in speed did not extend into the work area itself, perhaps owing to the site conditions and distance to the active work site. A more advanced version of this concept was tested on a four-lane divided highway in Arizona that had a 35 mph work zone speed limit (Roberts and Smaglik 2012). On the PCMS, a your speed is xx mph message was displayed to all vehicles. For vehicles traveling 46 mph or more, the speed feedback message alternated with a second message: possi- ble fine $xxx. The dollar amount varied from $155 to $480 based on the severity of the speeding. Reductions in mean speeds were modest (about 4 mph); however, the number of severe speeders was reduced by up to 50%. PORTABLE CHANGEABLE MESSAGE SIGNS WITH GENERAL SPEED SAFETY MESSAGES In the 1990s, as PCMS technology became widely available, PCMS units were added to supplement static work zone signs and traditional traffic control devices. In general, the intent is to draw attention to the changed road condition and the need for drivers to reduce speeds. PCMS used in this appli- cation are typically programmed with general work zone safety messages and speed limit information. They became a staple speed reduction device on many large freeway and high-speed roadway projects. Starting in the 1990s, a number of studies found that this technique resulted in a speed reduc- tion for all types of vehicles; however, in more recent studies only 1 to 4 mph reductions were typically observed (Garber and Srinivasan 1998a, b; Fontaine and Carlson 2001; Brewer et al. 2006; Bai et al. 2010). PORTABLE CHANGEABLE MESSAGE SIGNS SPEED FEEDBACK TRAILER WITH “POLICE” LIGHTS In 2011, small-scale field tests of a PCMS trailer augmented with experimental flashing “police” lights were performed at moderate-volume freeway maintenance work zones near Stockton in north-central California (Ravani et al. 2012). FIGURE 12 PCMS augmented with flashing “police” lights (Ravani 2012).

32 expected to affect speeding vehicles more than general traf- fic, because it is anticipated that drivers who intend to exceed the speed limit are more likely to have radar detectors. Con- sequently, the effectiveness of decoy radar depends on the number of radar detectors in the traffic stream. In February 1995, U.S.DOT issued a directive banning the use of radar detectors in all commercial vehicles. Nevertheless, it is likely that some commercial vehicles still use radar detectors. As of 2007, radar detector usage in passenger vehicles was legal except in the District of Columbia, Virginia, and U.S. military installations (Eckenrode et al. 2007). Several studies have been undertaken on the effective- ness of decoy radar in work zones since the mid-1980s. Table 7, based on Eckenrode et al. (2007), summarizes the evaluations of decoy radar in work zones. The same authors performed the most recent evaluation of decoy radar using information gathered in South Carolina work zones. Data were collected at five sites including interstate highways, state routes, and rural routes. Results showed mean speed reduction in the percentage of vehicles travelling through the work zone at speeds above 65 mph and approximately 6% more vehicles traveling at less than 60 mph. The accuracy of the Automated License Plate Recognition system in the speed camera declined under certain lighting conditions. DECOY RADAR Decoy or “drone” radar refers to unattended systems that emit a decoy radar signal. Decoy radar units are relatively inexpensive and can be mounted on a variety of roadside objects including signs, guardrails, arrow panels, traffic con- trol drums, and vehicles. The tactic is primarily used to catch the attention of drivers with radar detectors. An August 1991 directive from U.S.DOT and NHTSA authorized the use of decoy radars (Eckenrode et al. 2007). The intent is that vehicles equipped with radar detectors will perceive the presence of decoy radar as the presence of police enforcement and reduce speeds. In general, decoy radar is FIGURE 13 Concept of operation for augmented enforcement system (Chan et al. 2013). FIGURE 14 Messages displayed to speeders by the augmented enforcement system (Chan et al. 2013).

33 Date Study Description Configuration Results 1986 Northern Kentucky (Pigman et al. 1987) Used X-band decoy radar in 6-lane Interstate work zones Speed collected at 2 locations for several days for both night and day Radar detectors (42% tractor-trailers and 11% cars) mean and 85th percentile speed decreased 1990 Texas Transportation Institute (Ullman 1991) Used low output radar transmissions in 8 Interstate work zones Station 1: 3,000 ft upstream from work zone Station 2: 1,200 ft upstream from decoy Station 3: 2,000 ft downstream of Station 2 Speeds decreased 2 mph and standard deviations increased 1992 South Dakota DOT (Fors 1999) Used 500 of Kustom Signal Pro 65 X- band decoys Decoys were placed in 500 of the 602 state vehicles 21.1% decrease in statewide crashes 1992 Champaign, Illinois (Benekohal et al. 1993) Used a radar gun that acts as decoy on rural Interstates, CB radio Experiment 1: immediate effect of decoy Experiment 2: short-term effect Experiment 3: use multiple radar guns Experiment 1 (8 to 10 mph decrease) Experiment 2 (no effect) Experiment 3 (3 to 6 mph decrease) 1993 Maryland (Teed et al. 1993) Used multiple police radar devices on Interstates Radar placed every few miles. Tested radar detector use by looking for braking and 5 mph speed reductions Both experiments showed police radar has short-term effects on speed, radar detector usage was 30% 1994 Missouri (Freedman et al. 1994) Placed decoys on pavement edge of roads of long-term work zones Station 1: 0.4 mi from work zone Station 2: 0.2 to 0.8 mi in work zone Station 3: 0.4 mi from decoy 3.4 to 1.8 mph speed decreases (passenger cars), 3.6 to 2.0 mph decrease (tractor trailers) 1995 New Mexico and Texas (Fors 1999) Placed decoys on arrow boards and barrels Monitored CB radio and collected speeds for 40 consecutive hours 3 to 4 mph decrease (tractor-trailers), 2 mph reduction in speeds (cars) 1995 University of Michigan (Streff et al. 1995) Used decoy radar and police enforcement in two major Interstates Station 1: upstream from decoy detection Station 2: detection distance from decoy Station 3: downstream of decoy 2 mph decrease in speeds, radar detector usage days (5% cars), (19% tractor- trailers day, 28% night) 1997 Virginia Polytechnic Institute and State Univ. (Turochy 1997) Used the decoy radar checkpoint model 2A by PM design lab of NC Station 1: upstream into decoy Station 2: 500 to 1,000 ft from decoy detection Mean speeds decreased 0.8 to 2.3 mph, standard deviations reduced in half with decoy on 2000 Midwest Smart Work Zone Deployment Initiative (Fors et al. 2000) Placed decoy on either end of 1-mi segment of work zone Speeds were collected 4 h before decoy deployment and 4 h after No significant changes in speeds, but may reduce 85th percentile speeds 2001 Georgia Institute of Technology (de Oliveira et al. 2002) Tested safety warning system and basic decoy radar on Interstates Station 1: collected volume using road tubes Station 2: pictures of radar detector users Station 3: decoy unit No decrease in speeds, decelerated more with SWS in place TABLE 7 SUMMARY OF DECOY RADAR STUDIES

34 conjunction with decoy radar operation to maintain its effec- tiveness over the long term. A 1991 Texas study reported that severe braking and last minute lane changing increased when decoy radar was in oper- ation (Ullman 1991); however, one could expect that a similar response would be elicited by the presence of law enforcement. State DOTs interviewed for this project reported mixed results with the use of these systems. For example, VDOT experimented with a portable decoy radar system that could be attached to a guard rail, but stopped using the system after one construction season because of “disappointing” speed reduction results and because users found the system cum- bersome to move and recharge (D. Rush, VDOT, personal communication, 2014). reductions of about 2 mph for the entire traffic stream. Speed reductions of vehicles equipped with radar detectors ranged from 5 to 8 mph. Decoy radar also caused 85th percentile speeds to decrease between 1 and 5 mph, and a 20% speed reduction was shown in vehicles exceeding the speed limit. Another important finding from this study is the prevalence of radar detectors in the traffic stream. For passenger cars it varied from 2.2% to 5% and for tractor-trailers it ranged from 7.1% to 7.5%, which helps explain the relatively modest impact decoy radar has on speeds. Collectively, these studies suggest that decoy radar can have minor impacts on speeds within the detection distance of radar transmissions. Prolonged use of decoy radar limits its effectiveness. A 2005 Maryland State Highway Administration report (MSHA 2005) recommends periodic police enforcement in