Traffic Control Devices and Measures for Deterring Wrong-Way Movements (2018)

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.

A-1 In order to develop a comprehensive assessment of what is known from existing research with respect to the wrong-way driving (WWD) problem, its causes, and the development, implementation, and evaluation of wrong-way movement countermeasures, the research team conducted a literature review of domestic and international research. A summary of the findings from this effort are documented in this appendix. Wrong-Way Crash Characteristics Since the mid-1960s, many states in the United States have analyzed crash data to quantify the WWD problem and examine wrong-way crash characteristics. However, only a few past studies have focused on wrong-way crashes at the national level and in other countries. In addition, most of the previous research on WWD has focused on freeways. Below is a summary of the major findings related to the occurrence and severity of wrong-way crashes, the age and gender of the wrong-way driver, the role of driver impairment, the time of day in which wrong-way crashes occur, and the origination of the wrong-way movement. Occurrence and Severity of Wrong-Way Crashes Some of the earliest WWD research in the United States was conducted by the California Department of Transpor- tation (Caltrans) in the mid-1960s. Tamburri and Theobald (1965) found that during two separate 9-month periods, there were approximately 1200 wrong-way events in California: 763 on freeways, 354 on expressways (i.e., divided highways), and 97 on conventional roads. Additionally, Tamburri and Theobald found that from 1961 to 1964, there was an aver- age of 20 fatal wrong-way crashes on freeways. These fatal crashes resulted in an average of 31 deaths per year. When compared to all freeway and expressway crashes, Tamburri and Theobald found that wrong-way crashes were about six times more likely to produce a fatality. In the 1980s, Copelan (1989) published an updated analy- sis of wrong-way crashes in California. In 1987, wrong-way crashes accounted for approximately 0.24 percent of all free- way crashes. In addition, about 3 percent of the fatal crashes on California freeways were attributed to wrong-way drivers. Over a 20-year period, the number of fatal wrong-way crashes had averaged 35 per year. This number remained relatively constant even though the number of miles of freeway and travel substantially increased during this time. Therefore, the fatal wrong-way crash rate had decreased. Copelan attributed the reduction to the many actions taken by Caltrans to combat WWD over the 20-year period, some of which will be dis- cussed later in this appendix. Copelan also found that wrong- way crashes occurred more in urban areas than rural areas. Vaswani (1973) found that in Virginia, wrong-way crashes only accounted for 0.1 percent of the total crashes, but the fatality rate per wrong-way crash was 30 times that for other types of crashes. Vaswani (1977a) also reviewed trends in wrong-way events and crashes from 1970 to 1976. During this time, a total of 114 wrong-way crashes, in which 54 people died and 120 were injured, occurred on interstate highways in Virginia. An additional 167 wrong-way crashes occurred on other four-lane, divided highways during the same period, killing 33 people and injuring 173 people. Although the number of wrong-way crashes was still relatively small com- pared to total crashes, they remained more severe. Similar to Vaswani’s earlier study, the fatality rate per wrong-way crash was 31 times greater than for other types of crashes on inter- state highways and 10 times greater than for other types of crashes on four-lane divided highways. Two Texas studies (Cooner et al. 2004a; Finley et al. 2014) also confirmed that although wrong-way crashes represent less than 1 percent of all traffic crashes, they tend to be more severe and have a greater proportion that result in death or serious injury when compared with other types of crashes. In addition, these Texas studies and a study in Illinois (Zhou et al. 2012) found that urban areas have more wrong-way crashes than rural areas. Cooner et al. (2004a) also found A P P E N D I X A State of the Practice

A-2 that most wrong-way collisions occurred in the leftmost (i.e., inside) lane of the correct direction. Similarly, the Illinois study reported statistics regarding which lane the wrong-way crash was in: 51 percent in the lane closest to the median (i.e., inside or left), 16 percent in the middle lane, 20 percent in the outside or right lane, and 8 percent on the shoulder. Another 7 percent occurred on a ramp. A study of 110 wrong-way crashes on Michigan freeways during the 5-year period from 2005 to 2009 also found that wrong-way crashes were highly severe (Morena and Leix 2012). In 87 percent of the crashes, the wrong-way vehicle hit another vehicle traveling in the correct direction. The remaining crashes were single-vehicle crashes that involved the wrong-way vehicle only. The severity was also linked to the location of the crash. Only 6 percent of the crashes that occurred on an exit ramp resulted in a fatality or incapacitat- ing injury. In contrast, 42 percent of the crashes on the main lanes resulted in a fatality or incapacitating injury. A study in North Carolina for the 6-year period from 2000 to 2006 found that wrong-way crashes accounted for less than 0.2 percent of all freeway crashes (North Carolina DOT 2006). A similar study conducted with data from 2006 to 2012 yielded similar results (North Carolina DOT 2006). The latter study also found that 72 percent of all wrong-way crashes occurred on interstates and 52 percent happened in rural areas (North Carolina DOT 2006). More recent studies in Florida (Kittelson & Associates, Inc. 2015) and Arizona (Simpson and Bruggeman 2015) also found that wrong-way crashes on freeways were more severe than other crashes on urban and rural freeways. In Florida, more than half of the wrong-way crashes resulted in injury and over 25 percent resulted in a fatality. In addition, approximately 76 percent of the wrong-way crashes on free- ways occurred in urban areas compared to 24 percent in rural areas. In Arizona, 25 percent of all wrong-way crashes were fatal compared to 1 percent overall. Similarly, Zhang et al. (2016) found that on divided highways in Alabama, 12.5 per- cent of wrong-way crashes were fatal compared to 0.6 percent of all other crashes. An analysis of wrong-way crashes from the Fatality Acci- dent Reporting System (FARS) database from 1996 to 2000 (Moler 2002) revealed that 1753 people died in wrong-way crashes on U.S. freeways (ramps and main lanes). During the 5-year period, on average about 350 people were killed each year nationwide in wrong-way freeway crashes. Thousands more were injured. In December of 2012, NTSB released a special investigation report on WWD on freeways and access ramps. This report documented the results of an analysis of fatal crash data in the United States from 2004 to 2009 and summarized nine NTSB wrong-way collision investigations. On average, 360 fatalities occurred each year from about 260 wrong-way crashes. Most of the wrong-way crashes occurred in the lane closest to the median (i.e., inside lane). NTSB concluded that although wrong-way crashes occur relatively infrequently (account- ing for only about 3 percent of all crashes on freeways), they are more likely to result in fatal and serious injuries than other types of highway crashes because the vast majority are head-on collisions. A more recent query of 8 years of crash data (2004 to 2011) from the FARS database (American Traffic Safety Services Association 2014; Baratian-Ghorghi et al. 2014) revealed that on average, 269 fatal wrong-way crashes continued to occur each year, resulting in 359 fatalities. These findings showed that these numbers remained steady even though the overall number of fatal crashes and fatalities decreased in the United States by more than 20 percent during the study period. On average, wrong-way fatalities account for about 2.3 percent of total fatalities on freeways. Further analysis on these data found that a higher proportion of wrong-way fatal crashes occur on urban roadways. Wrong-way crash studies have also been conducted in other countries. A study performed by the Japanese Institute for Traffic Accident Research and Data Analysis (ITARDA 2002) found that less than 1 percent of crashes were caused by a wrong-way driver. However, 12 percent of the wrong- way crashes resulted in a death compared to 2 percent of total crashes. Similarly, studies in the Netherlands (SWOV Institute for Road Safety Research 2007) found that wrong- way crashes were less than 1 percent of all registered crashes, and that wrong-way crashes were rather severe. In another Japanese study (Xing 2013), researchers found that 62 per- cent of wrong-way crashes occurred in the median (i.e., inside) lane. The other lanes and shoulders accounted for 3 to 5 percent of the crashes. An additional 19 percent happened on ramps at interchanges, junctions, and rest areas. In France (Kemel 2015), wrong-way crashes represented 1 percent of all injury crashes and 6 percent of fatalities. Wrong-way crashes were also six times more likely to be fatal than other crashes. Overall, wrong-way crashes represent a small portion of the total crashes on U.S. freeways and divided highways. However, these crashes result in more fatalities than other types of crashes on these facilities because most are head-on collisions. Wrong-way crashes tend to occur more in urban areas than rural areas, and most tend to take place in the lane closest to the median (i.e., inside lane). Wrong-Way Driver Characteristics Gender and Age Tamburri and Theobald (1965) were some of the first researchers to investigate the characteristics of the wrong- way driver. They found that 80 percent of the wrong-way drivers in California were male, and that the rate of WWD per vehicle miles traveled increased with age. Lew (1971) also reported on the age of the California wrong-way driver. He found that drivers 16 to 19 years old were underrepresented

A-3 in wrong-way driving crashes and those 70 to 79 years old were over twice the number expected based on their propor- tion of the driving population. More recently, two Texas studies (Cooner et al. 2004a; Finley et al. 2014), a New Mexico study (Lathrop et al. 2010), a Michigan study (Morena and Leix 2012), an Illinois study (Zhou et al. 2012), and an Alabama study (Pour-Rouholamin et al. 2016) confirmed that most of the wrong-way crashes still involved a male driver (59 to 77 percent). As with previ- ous studies, the percentage of Texas, Michigan, Illinois, and Alabama drivers over 65 years old involved in wrong-way crashes was higher compared to their involvement in other types of crashes. In addition, almost half of the wrong-way drivers in Texas were 16 to 34 years old. Furthermore, in Michigan and North Carolina (North Carolina DOT 2006), nearly one-quarter of the wrong-way drivers were under the age of 25 or 20 to 29 years old, respectively. In Illinois, Zhou et al. (2012) found that younger drivers (under the age of 25) were also proportionally overrepresented. In Florida (Kittelson & Associates, Inc. 2015), although drivers less than 30 years old accounted for 42 percent of the wrong-way crashes on freeways, this proportion was similar to all freeway crashes for this age group. In contrast, older drivers (≥ 75 years old) accounted for about 5 percent of the wrong-way crashes on freeways. This proportion was more than three times the expected proportion from statewide trends on freeways. On divided highways in Alabama, Zhang et al. (2016) found that drivers less than 44 years old accounted for approximately 47 percent of the wrong-way crashes. How- ever, this proportion was similar to all other divided highway crashes for this age group. In contrast, drivers over 65 years old accounted for approximately 27 percent of the wrong- way crashes on divided highways. However, this proportion was more than two times the proportion for all other crashes on divided highways. A query of 8 years of crash data (2004 to 2011) from the FARS database (Baratian-Ghorghi et al. 2014) found that male wrong-way drivers outnumbered female wrong-way drivers by more than 2 to 1. About 15 percent of the wrong- way drivers were age 65 or older. Comparing this to the 10 percent of older drivers involved in other freeway crashes showed that older drivers are overrepresented in wrong-way fatal crashes. The NTSB (2012) report also noted findings that drivers over the age of 70 are overrepresented in wrong- way crashes. A study performed by ITARDA (2002) in Japan determined that the following age groups were overrepresented in wrong- way crashes: 25 to 29, 45 to 54, and 75 to 79. Furthermore, 29 percent of the wrong-way crashes on highways involv- ing injury or death were caused by senior citizens. A more recent Japanese study (Xing 2013) found that approximately half of the wrong-way drivers were more than 65 years old, and 40 percent were over 70 years old. Studies in the Netherlands (SWOV Institute for Road Safety Research 2007) also found that drivers at least 70 years old were 23 times more likely than other age groups to be involved in a wrong- way crash. In France (Kemel 2015), drivers over the age of 65 were also overrepresented in wrong-way crashes, and male drivers accounted for approximately 75 percent of the wrong- way crashes. Overall, male drivers tend to be involved in wrong-way crashes more than female drivers. Many studies also found that younger drivers and senior citizens tend to be propor- tionally overrepresented in wrong-way crashes. Driver Impairment Tamburri and Theobald’s (1965) study also revealed that approximately one-third of all the wrong-way drivers in California had been drinking. Of those who had been drink- ing, most (58.6 percent) were 30 to 49 years old. In the 1980s, Copelan (1989) found that almost 60 percent of all wrong- way crashes and 77 percent of fatal wrong-way crashes in California were caused by a driver who was impaired by drugs or alcohol. An analysis of wrong-way crashes in Indiana from 1970 to 1972 showed that approximately 55 percent of the drivers were impaired (Scifres and Loutzenheiser 1975). Similarly, a 1977 Virginia study (Vaswani 1977a) found that 53 percent of the wrong-way drivers on interstates and 33 per- cent of the wrong-way drivers on non-interstate, four-lane, divided highways were driving under the influence. Further- more, a Washington State DOT study (Moler 2002) indicated that half of the 30 wrong-way crashes along an interstate cor- ridor were alcohol or drug related. From 1983 to 1990 in the Netherlands (SWOV Institute for Road Safety Research 2007), alcohol was involved in 45 percent of the wrong-way crashes. However, this percentage decreased to 20 percent from 1991 to 1998. Over the whole period (1983 to 1998), 56 percent of the wrong-way drivers 25 to 54 years old were under the influ- ence of alcohol. A 2004 Texas study (Cooner et al. 2004a) found that almost 61 percent of all the wrong-way crashes on freeways had some influence of alcohol and/or drugs cited. In 2013, Finley et al. (2014) investigated the blood alcohol concentration (BAC) level of wrong-way drivers in Texas that were tested for alco- hol. Key findings regarding impairment were: • Almost 90 percent had a BAC level equal to or greater than the legal limit (0.08 g/dL). • Approximately 50 percent had a BAC level equal to or greater than twice the legal limit (0.16 g/dL). • Approximately 10 percent had a BAC level equal to or greater than three times the legal limit (0.24 g/dL). • The BAC ranges with the highest percentage of drivers (30 percent) were 0.16 to 0.199 g/dL and 0.20 to 0.239 g/dL. • Most drivers (60 percent) had a BAC level of 0.16 to 0.239 g/dL.

A-4 • The average BAC level was 0.18 g/dL (over twice the legal limit). Many other states have also found alcohol to be a major contributing factor for wrong-way crashes on freeways. In Arizona, impaired drivers were the cause of 65 percent of all wrong-way crashes (Simpson and Bruggeman 2015). In Michigan, nearly 60 percent of the wrong-way drivers for which the impairment was known were under the influence of either alcohol or drugs (Morena and Leix 2012). Similarly, in New Mexico, 63 percent of the wrong-way drivers tested had a BAC level above the legal limit (0.08 g/dL) (Lathrop et al. 2010). In North Carolina (North Carolina DOT 2006), 48 percent of all wrong-way crashes were alcohol related. An Illinois study (Zhou et al. 2012) found that 50 percent of the wrong-way drivers were under the influence of alcohol and 5 percent were under the influence of drugs. Further- more, 80 percent of the drivers completing a test for alcohol and drugs had a BAC level greater than 0.1 g/dL. In Florida, alcohol and/or drug use was involved in 45 percent of the wrong-way crashes on freeways (Kittelson & Associates, Inc. 2015). In Alabama, almost half of the wrong-way drivers were intoxicated (Pour-Rouholamin et al. 2016). For divided highways in Alabama, Zhang et al. (2016) found that driving under the influence was a contributing factor in only 23 percent of wrong-way crashes. However, this propor- tion was more than seven times the proportion for all other crashes on divided highways. The NTSB (2012) report revealed that approximately 60 percent of fatal wrong-way crashes involved alcohol. Further more, in seven of the nine wrong-way crashes investigated by NTSB, the wrong-way driver had BAC over 0.15 g/dL. A more recent query of 8 years of crash data (2004 to 2011) from the FARS database (Baratian-Ghorghi et al. 2014) found that on average, 58 percent of wrong-way crashes in the United States were related to driving under the influence of alcohol and/or drugs. This was nearly twice the rate of alcohol or drug involvement for all fatal crashes. Furthermore, impaired male wrong-way drivers outnum- bered impaired female wrong-way drivers by nearly 3 to 1. Nearly two-thirds of the wrong-way drivers under the age of 65 were impaired by alcohol or drugs. Overall, driving under the influence of alcohol and/or drugs is the primary contributing factor in most of the wrong-way crashes. Impaired wrong-way drivers involved in crashes tend to be younger and have BAC levels that exceed the legal limit by two to three times (i.e., 0.16 to 0.239 g/dL). Time-of-Day Characteristics In the 1960s, Tamburri and Theobald (1965) found that in California, WWD events peaked at 11:00 a.m. and again at 2:00 a.m. The late morning peak was attributed to drivers over 60 years old, while the early morning peak coincided with the typical time for establishments that serve alcohol to close in California. Likewise, both Japanese studies (ITARDA 2002; Xing 2013) found that most crashes involving senior citizens occurred during the day and most crashes involving the other age groups occurred at night. In the 1980s, Copelan (1989) found that the number of WWD crashes was higher at night than during the day and confirmed that wrong-way crashes in California still peaked at 2:00 a.m. Two studies in Texas (Cooner et al. 2004a; Finley et al. 2014) also identified a peak in wrong-way crashes on free- ways around the typical closing time for establishments that serve alcohol. These two Texas studies also found that more than half of the wrong-way crashes occurred between mid- night and 6:00 a.m. This finding led Cooner et al. (2004a) to conclude that wrong-way crashes on freeways were five times more likely to occur during the early morning hours. Studies in North Carolina (North Carolina DOT 2006), New Mexico (Lathrop et al. 2010), Michigan (Morena and Leix 2012), Florida (Kittelson & Associates, Inc. 2015), Arizona (Simpson and Bruggeman 2015), and Alabama (Pour-Rouholamin et al. 2016) identified similar trends with respect to time of day. Many state studies have also found that wrong-way crashes on freeways occur most frequently at night on the weekend (e.g., Scifres and Loutzenheiser 1975; Zhou et al. 2012; Finley et al. 2014; Kittelson & Associates, Inc. 2015). For the United States, the NTSB (2012) report noted that fatal wrong-way crashes on freeways occurred more frequently at night (78 percent) and on the weekend (57 percent). Zhang et al. (2016) found that most of the wrong-way crashes on divided highways in Alabama occurred in the evening (6:00 p.m. to midnight, 46 percent) and at night (midnight to 6:00 a.m., 20 percent). Overall, many daytime wrong-way crashes are caused by senior citizens, while those occurring at night are attributed to other age groups and involve driving under the influence of alcohol and/or drugs. All studies seem to suggest that wrong- way crashes are more prevalent at night during the early morning hours. Origin of Wrong-Way Movement In California, Tamburri and Theobald (1965) found that on freeways, 53 percent of the events began when a driver entered the freeway via an exit ramp. Other drivers made a U-turn on the main lanes (19 percent), made a U-turn from the on-ramp (9 percent), or drove across the median (4 per- cent). On freeways, the origin of a wrong-way maneuver most frequently occurred at diamond interchanges (38 percent), although the researchers noted that this was the most com- mon type of ramp in California at that time. Copelan’s (1989) review of the California data and counter- measure implementation from the 1960s to the 1980s revealed that cul-de-sac, button hook, trumpet, and two-quadrant

A-5 cloverleaf interchanges had a greater number of wrong-way crashes than other types of interchanges. California studies also found that left exits appeared to be on-ramps to some wrong-way drivers. In Virginia, Vaswani (1977a) found that about 50 percent of the wrong-way entries onto an interstate highway originated at interchanges. Another 15 percent originated at crossovers and rest areas or were associated with U-turns and median crossings. Vaswani (1973) also found that most wrong-way entries were at partial interchanges of the diamond type. The NTSB (2012) report also noted that entering an exit ramp in the wrong direction was the primary origin of wrong- way movements onto freeways. Other actions resulting in wrong-way movements included making U-turns on the main lanes of the freeway and using an emergency turn- around through the median. Parsonson and Marks (1979) determined that the half- diamond, partial cloverleaf loop ramp, and partial cloverleaf AB loop ramp had the highest wrong-way entry rates. These researchers found that since half-diamonds are incomplete, drivers often made intentional wrong-way maneuvers at these types of interchanges. The main issue identified with the partial cloverleaf designs was that the entrance and exit ramps are in close proximity to each other. Cooner et al. (2004a) gathered data on the locations where 323 freeway-related wrong-way crashes originated from crash reports. These researchers found that they could only deduce the entry point for about one out of every three crashes. Even so, the researchers were able to note several important char- acteristics about wrong-way crashes. • Several locations with left-side exit ramps experienced multiple wrong-way crashes. • Several locations where a one-way street transitions directly into a freeway section resulted in multiple wrong- way crashes. • Almost 4 percent of the wrong-way crashes were the result of a driver making a U-turn on the main lanes. In New Mexico, Lathrop et al. (2010) could only deter- mine the origination of the wrong-way maneuver for 12 out of 49 wrong-way collisions. Five of the drivers entered the interstate the wrong way by driving up an exit ramp, four of the drivers made a U-turn on the main lanes, and three of the drivers entered from external highway/nonstandard entrance points. In Michigan, Morena and Leix (2012) were able to con- firm that 31 out of 110 crashes entered at a specific exit ramp since the collision occurred at the same ramp. In addition, the reporting officer was able to identify the point of entry for four main lane crashes. For all the other crashes, the point of origin was unknown. Using the known entry points, Morena and Leix reviewed the type of interchange and offered edu- cated commentary on the potential driver confusion issues that could have led to the wrong-way entry. Their review con- firmed previous study findings about the potential confusion at partial cloverleaf designs. This type of interchange design was involved in 60 percent of the known wrong-way entry points, even though they only account for 21 percent of the interchanges in Michigan. The main culprit is the pair of free- way exit and entrance ramps that are adjacent and parallel to each other and intersect with the cross street at a 90-degree angle. Disoriented, distracted, or impaired drivers can mis- takenly turn onto the exit ramp instead of the entrance ramp. Morena and Leix also found that trumpet interchanges expe- rienced 11 percent of the wrong-way entries but only com- prised 3 percent of the interchanges in Michigan. In Illinois, Zhou et al. (2012) confirmed how difficult it is to identify the wrong-way entry point from crash report narratives. Only 20 percent of the wrong-way crashes had wrong-way entry points recorded in the reports. For all other crashes, the researchers had to estimate the wrong-way entry point. For the 47 known origins, 14 began with a driver making a U-turn on a freeway. Compressed diamond and diamond interchanges resulted in 26 percent and 34 per- cent of the known wrong-way entries, respectively. Partial cloverleaf interchanges were connected with 11 percent of the known entries. With respect to divided highways, Tamburri and Theobald (1965) found that approximately half of the California crashes originated at intersections with median openings. However, some wrong-way maneuvers (6 percent) occurred at inter- sections without median openings (i.e., only right turns permitted). Furthermore, 19 percent of the drivers traveled through the median (not at an intersection) and 19 percent made a U-turn on the main lanes. Vaswani (1977a) found that on non-interstate, four-lane, divided highways in Virginia, about 40 percent of the wrong- way entries occurred at intersections with crossroads or exit ramps connecting with interstate routes. Another 25 percent happened when the driver exited from a business establish- ment. About 20 percent occurred at crossovers, the beginning of divided sections, and construction sites or were associated with U-turns and median crossings. Beginning in 2013, the Iowa department of transportation piloted the use of high definition radars and video analyt- ics to detect wrong-way vehicles. Findings from this effort showed 68 percent of the 51 confirmed wrong-way events entered via at-grade intersections on divided highways. The most common wrong-way entry path was a vehicle turning left from an intersecting roadway into the near main lanes (Athey Creek Consultants 2016). Overall, identifying the location of a wrong-way entry is difficult. Nevertheless, many studies have found that most

A-6 of the wrong-way movements on freeways originate at exit ramps. Wrong-way maneuvers on freeways also occur from drivers making U-turns on the main lanes, traveling across the median, and entering from a nonstandard entrance point. Most wrong-way entries on divided highways occur at inter- sections with median openings. However, they also happen at intersections without median openings and from drivers making U-turns on the main lanes or driving through the median. Several types of interchange designs have been found to be more susceptible to wrong-way entries. The side-by-side entrance and exit ramp configuration of partial cloverleaf interchanges may cause drivers to unintentionally turn into the exit ramp instead of the entrance ramp. Other incomplete or partial interchanges may lead to intentional wrong-way maneuvers. In addition, in some cases, drivers mistake an exit ramp at a diamond interchange as a two-way frontage road that runs parallel to the main lanes. Left-side exit ramps are also more susceptible to wrong-way entries. Other Contributing Factors for Wrong-Way Crashes As discussed previously, many studies have identified driv- ing under the influence of alcohol and/or drugs, older drivers, and interchange design as factors that contribute to wrong- way crashes. Other contributing factors include intentional behavior (e.g., suicide, avoiding congestion, or missed exit), driver inattention, driver confusion, physical or mental ill- ness, insufficient lighting, insufficient sight distance, and inclement weather (Zhou et al. 2012). Wrong-Way Crash Countermeasures Common countermeasures for WWD include traffic control devices (signs and pavement markings), intelligent transportation systems (ITSs), and geometric modifications (e.g., access management and interchange reconstruction). Enforcement and education are also tools that can be used to mitigate WWD. In the future, connected vehicle applications will also be used to detect and alert wrong-way drivers, right- way vehicles in the path of the wrong-way driver, enforce- ment, traffic management centers, and other entities through the use of vehicle-to-infrastructure and/or vehicle-to-vehicle technologies. With respect to the objectives of this project, the research team focused the literature review on traffic control devices and associated ITS/technology systems. In the mid-1960s, Caltrans research also involved identify- ing wrong-way countermeasures. Caltrans examined the use of preventive measures including the spike strip on exit ramps and a detection and warning system that featured a WRONG WAY sign that was automatically illuminated when a wrong- way vehicle was detected on an exit ramp in conjunction with an electric horn warning to alert the WWD driver (Tamburri and Theobald 1965; Doty and Ledbetter 1965; Tamburri 1965). Caltrans staff concluded that spike strips: • Were not a safe and viable countermeasure because they disabled but did not stop a wrong-way vehicle, • Could create a hazard when spikes were broken, • Were an ongoing maintenance concern to ensure proper operation, and • Could be misinterpreted by right-way drivers as a hazard. In the 1970s, Caltrans examined a modified form of the wrong-way vehicle detection system it originally developed in the late 1960s (Rinde 1978). Caltrans placed this system in over 4000 freeway exit ramp locations throughout California to assess which ramp designs and other factors were associ- ated with wrong-way driving. This research showed that the following changes to the standard exit ramp signing were effective in reducing wrong-way entries onto freeways to less than or equal to two per month at 90 percent of the ramps identified as having a significant wrong-way entry problem: • The bottom of the lower portion of the wrong-way move- ment sign package (WRONG WAY sign mounted below a DO NOT ENTER sign on the same post) was placed 2 ft above the edge of the pavement. • At least one wrong-way movement sign package was placed in the area covered by a vehicle’s headlights and visible to drivers from all approaches. • ONE WAY signs were mounted 1.5 ft above the pavement. • FREEWAY ENTRANCE signs were placed as close as pos- sible to the intersection of the ramp and cross street. While this research has been used to encourage the use of a lower mounting height for DO NOT ENTER and WRONG WAY signs, the positive findings were a result of the imple- mentation of a combination of traffic control devices, not just lowered signs. Similarly, the Georgia Department of Trans- portation sponsored research in the late 1970s that used the wrong-way camera system from Caltrans in a study to moni- tor exit ramps in order to correlate various ramp designs with WWD activity (Parsonson and Marks 1979). The first research regarding WWD in Texas occurred from the late 1960s to the early 1970s. Texas A&M Transportation Institute (TTI) researchers conducted a survey of state and local highway engineers and law enforcement personnel in an attempt to qualitatively determine the nature of WWD in Texas (Messer et al. 1971). Researchers also summarized the state of the knowledge on WWD on freeways and express- ways, including a review of countermeasures and the devel- opment of a detection and communication system to warn drivers of WWD (Friebele et al. 1971). As discussed previously, in the mid-1970s, the Virginia Transportation Research Council (VTRC) conducted research

A-7 to identify the causes of wrong-way movements (Vaswani 1973, 1977a). VTRC also developed countermeasures to address the causes identified. The countermeasures were mainly directed at exit ramp configurations and included improved pavement markings that used reflectorized wrong-way pave- ment arrows on all exit ramps, implementation of sensors on exit ramps for detecting WWD in future construction projects, and consideration of lowered DO NOT ENTER and WRONG WAY signing to address alcohol and nighttime problem locations. Vaswani (1977b) also experimented with divided highway crossing signs on the minor approaches to an intersection with a divided highway. According to Vaswani, the Delaware Department of Highways and Transportation had success- fully used this sign for 20 years to mitigate wrong-way maneu- vers. The field observations in Virginia found nine wrong-way entries during the 3 years of before data and no wrong-way entries in the 7-month after period. Although the after evalu- ation period was too short to allow for definite conclusions, the results were encouraging. In the 1980s, the Illinois DOT experimented with sensors embedded in the roadway to detect wrong-way traffic move- ment, which, if activated, would lower a signal arm across the road and initiate a dynamic message sign (DMS) to alert exit- ing traffic about the WWD hazard ahead (Knight 1983). In New Mexico in the 1990s, a directional traffic sensor system was implemented on an exit ramp near Albuquerque. This system used loop sensors that detected wrong-way vehicles and activated red flashers on a WRONG WAY sign to warn the wrong-way driver. Additionally, yellow flashers on a STOP AHEAD sign for right-way ramp vehicles were used to warn traffic of an exit ramp obstacle (Cooner et al. 2004a). In the 2000s, the Washington State DOT used video moni- toring systems on select exit ramps to detect wrong-way drivers. When a wrong-way vehicle was detected, a blank- out sign with the message WRONG WAY and flashers were activated. Concurrently, the system videotaped the vehicle’s movements and the driver’s behavior to further assess the problem (Moler 2002). In 2003, the Texas DOT sponsored WWD research fol- lowing several severe wrong-way crashes around the state (Cooner et al. 2004a, 2004b). The major findings from the research called for the use of reflectorized wrong-way arrows on exit ramps, lowered DO NOT ENTER and WRONG WAY signs mounted together on the same sign support, and devel- opment of a field checklist for wrong-way entry problem locations. In 2005, Schrock et al. conducted a before-after study to determine if lane direction arrows on a two-way frontage road reduced the number of wrong-way maneuvers. In the before period, these researchers found that one of every 13 drivers exiting the freeway onto the frontage road incor- rectly used the left lane for travel. In the after period, only one of 150 vehicles used the incorrect lane. Overall, these researchers found a 90 percent reduction in the number of wrong-way maneuvers. In 2006, a wrong-way detection system was implemented on the Pensacola Bay Bridge in Florida (Williams 2007). This system used a low-power microwave radar detector that was not affected by adverse weather conditions. The detector was mounted approximately 20 ft above the roadway and could detect a wrong-way movement at approximately 1000 ft prior to the bridge. When a wrong-way movement was detected, flashing beacons visibly enhanced the DO NOT ENTER and WRONG WAY signs above the travel way. In October 2008, the Harris County Toll Road Authority (HCTRA) began to operate a wrong-way driver detection system on a 13.2-mi portion of the West Park Tollway, a con- trolled-access roadway in Houston. The system used Doppler radar detection sensors supplemented with in-pavement loop sensors at 14 points along the tollway. Incident management center (IMC) personnel received all wrong-way movement detections and monitored the system 24 hours a day and 7 days a week. Once a vehicle was detected, operators at the IMC could immediately dispatch law enforcement officers, monitor the vehicle’s whereabouts via closed-circuit television (CCTV) and a geographic information system wrong-way detection map integrated into the software platform, and warn other motorists of the detected wrong-way vehicle using DMSs. This deployment was the first of its type in the United States and incorporated a number of innovative aspects including site-specific design, configuration, and communications dis- patch and response protocols (ITS International 2010). The original cost in 2007 was $337,000 (about $25,530 per mile). In 2011, HCTRA spent an additional $175,000 to enhance the system, which increased the cost per mile to approxi- mately $38,788. Additional features included: • Once the alarm is activated, the nearest CCTV camera automatically pans toward the detection site so that IMC dispatchers can track a wrong-way vehicle and relay infor- mation to first responders. • Warning messages conveyed to other drivers on DMSs can be displayed in automated incident response plans based on the direction of travel and location of the detection. • Light-emitting diode (LED) in-ground lighting was installed to warn motorists at South Post Oak and Richmond Avenue. • WRONG WAY signs with flashing LEDs around the border were installed at locations that have a higher rate of incidents. • Through attrition, in-ground puck loop systems are replac- ing radar sensors. Finley et al. (2014) reviewed the WWD alerts received by the HCTRA system from 2009 to 2013. Most of the verified

A-8 WWD alerts showed that the driver self-corrected (74 per- cent). Law enforcement caught only about 13 percent of the wrong-way drivers that resulted in an alert. Of these, two-thirds of the drivers were arrested for driving while intoxicated. The HCTRA detection system alert review also revealed that about 7 percent of the alerts were the result of a reversing down the roadway. In most instances, this appeared to occur because the driver missed the exit. Finley et al. (2014) also examined the HCTRA WWD alert data to determine the effectiveness of the flashing LED in-pavement lighting and WRONG WAY signs with flashing red LEDs around the border at South Post Oak. A before-after evaluation of these two devices could not be completed since the addition of the puck sensors (also in 2011) increased the number of overall alerts received. Even so, the HCTRA data did show that out of the 62 WWD alerts received for this loca- tion between January 2012 and December 2013, 86 percent of the drivers self-corrected before reaching the main lanes. In 2009, in response to wrong-way crashes on the Dallas North Tollway, the North Texas Tollway Authority (NTTA) formed a WWD task force and deployed a number of signing and marking countermeasures, including wrong-way pave- ment markings created with retroreflective raised pavement markers (RRPMs) at every exit ramp and red retroreflective sheeting on exit ramp sign supports (November 2009). Further countermeasure implementation included: • WRONG WAY signing with flashing red LEDs around the border at three exit ramp locations in December 2010. These signs flash continuously (i.e., day and night). • Pavement marking and signing modifications at cross street approaches at problem locations (January 2011 at Wycliff Avenue and June 2012 at the south end of the Dallas North Tollway). With respect to diamond inter- changes, NTTA replaced left-turn arrows with through- lane-use arrows when the left-turn lane for the exit ramp extended back beyond the entrance ramp. Based on previous research recommendations (Cooner et al. 2004a) and success in other states (notably California), NTTA also considered the use of lowered DO NOT ENTER and WRONG WAY signing. Although NTTA was aware that a 3-ft mounting height was an option, it was unable to locate any crash tests to verify that signs at this height would not be hazardous to an errant vehicle that was traveling the right way on the system. TTI researchers also had concerns regarding how a sign mounted at 3 ft would perform using the latest crash test criteria in the AASHTO Manual for Assessing Safety Hardware (MASH) (2009). In addition, modeling revealed that a sign mounted at 2 ft was almost twice as bright as a sign mounted at 3 ft. Thus, the 2-ft mounting height (measured vertically from the bottom of the sign to the elevation of the near edge of the pavement) was proposed as a height capable of catching an impaired driver’s attention while still being able to alert unimpaired drivers of restricted movements and meet current crashworthiness criteria. Using standard 36-inch by 36-inch DO NOT ENTER signs and 24-inch by 36-inch WRONG WAY signs, the sign assemblies had a total height of 5 ft and 4 ft, respectively (measured vertically from the top of the sign to the elevation of the near edge of the pavement). NTTA contracted with TTI to determine if the 2-ft sign assemblies described above would meet the provisions of the AASHTO MASH. The testing was conducted, and the findings were submitted to the FHWA Office of Safety for review. On December 7, 2010, NTTA received a letter from FHWA stating that the 24-inch by 36-inch WRONG WAY sign mounted 2 ft above the ground was acceptable to use on the National High- way System under the provisions of the AASHTO MASH. In January 2011, TTI completed a crash test on a 36-inch by 36-inch DO NOT ENTER sign mounted at a 2-ft height. The results showed that the test assembly and sign passed. In spring 2011, NTTA, in cooperation with Texas DOT, requested experimentation to mount 36-inch by 36-inch DO NOT ENTER and 24-inch by 36-inch WRONG WAY signs at 2 ft instead of the standard 7-ft mounting height. FHWA approved this request on July 14, 2011. At that time, NTTA had 142 exit ramps on its system; of these, 51 were tolled (meaning they had in-ground loops send- ing wrong-way driver alerts to the command center). NTTA decided to install lowered signs at 28 exit ramps (11 tolled and 17 non-tolled locations) based on the frequency of WWD events, geometry of the ramp, presence of pedestrians, and desire to have system-wide coverage. NTTA implemented the following three configurations of the DO NOT ENTER and WRONG WAY signage to accommodate pedestrian and cross- traffic visibility concerns: • Configuration 1: DO NOT ENTER sign at the 2-ft mount- ing height and WRONG WAY sign at the standard 7-ft mounting height. This configuration was installed at 12 locations. • Configuration 2: DO NOT ENTER sign at the standard 7-ft mounting height and WRONG WAY sign at the 2-ft mounting height. This configuration was installed at two locations. • Configuration 3: DO NOT ENTER and WRONG WAY signs both at the 2-ft mounting height. This configuration was installed at 14 locations. At the remaining 114 exit ramps in the NTTA system (40 tolled and 74 non-tolled), the DO NOT ENTER and WRONG WAY signs remained at the standard height (7 ft). The standard configuration served as a control group during the evaluation period. Finley et al. (2014) used WWD event data collected by NTTA from August 2010 to July 2013 to conduct an initial

A-9 analysis regarding the effectiveness of the lowered signs. However, the limited sample size produced statistically insignificant results. An updated analysis in 2015 (Finley et al. 2016b) found a 56 percent reduction in WWD events after the installation of lowered signing. In other words, the WWD events at the exit ramps with lowered signing were cut in half. This percent change was statistically significant at a 5 percent significance level, and the 95 percent confidence interval ranged from an 85 percent reduction to a 27 percent reduction. The analysis method used accounted for changes in events due to factors other than the lowered signing (e.g., traffic volumes), but it did not account for regression to the mean, which may have occurred since NTTA installed lowered signing at some ramps with high frequencies of WWD events. Thus, the analysis may overestimate the effectiveness of low- ered signing at sites with an average WWD event history. In May 2011, public transportation and law enforcement agencies in the San Antonio area created a WWD task force to share information and identify means to address and reduce WWD activity. The task force used various methods to doc- ument WWD activity in San Antonio, with the purpose of identifying where WWD countermeasure deployment would be most meaningful and effective. After analyzing the vari- ous WWD event data sources and the information details available from each source, analysts determined that insuf- ficient information existed to link WWD events with specific freeway ramps where wrong-way drivers entered the freeway network. Accordingly, there was no logical means that could be devised for prioritizing the treatment of one freeway ramp over another. Thus, the task force concluded that treatment of an entire freeway corridor was necessary in order to deter- mine the effectiveness of WWD countermeasures. The task force selected the 15-mi US 281 corridor from I-35 (near downtown) to just north of Loop 1604 (the far north central side of San Antonio) as the Wrong-Way Driver Countermeasure Operational Test Corridor. Between March 2012 and June 2012, Texas DOT staff and contractors installed WRONG WAY signs with flashing red LEDs around the border at each exit ramp in the US 281 test corridor. The purpose of the flashing red LEDs was to increase the conspicuity of WRONG WAY signing at night. The signs were set to flash under low ambient light conditions (i.e., at night and during some inclement weather events), whether or not a wrong- way vehicle was detected. Texas DOT felt that this operation would catch the attention of a wrong-way driver approach- ing on the frontage farther away instead of waiting until he or she was driving up the ramp. Also, sustained false alarm issues with the detection equipment led to deactivation of the detection component. Where the length and design of the exit ramp allowed, WRONG WAY signs with flashing red LEDs around the bor- der supplemented the existing, static WRONG WAY signs. On shorter ramps, the WRONG WAY signs with flashing red LEDs around the border replaced the existing static WRONG WAY signing. The battery for the signs was encased in the sign pole and charged by a small solar array attached to the top of the sign support. Even before the task force was created, the San Antonio Police Department (SAPD) and Texas DOT implemented sev- eral procedures with regard to responding to WWD events. In August 2010, SAPD began to use an emergency call signal (i.e., E-tone) for its radio network when a wrong-way driver was reported to 911. In January 2011, SAPD implemented a code in its computer-aided dispatch (CAD) system that spe- cifically identified all wrong-way driver events. Similarly, in March 2011, Texas DOT TransGuide traffic management cen- ter (TMC) operators began logging all WWD events, not just those that resulted in a crash. In May 2011, Texas DOT Trans- Guide operators began displaying wrong-way driver warning messages on DMSs when an E-tone was issued (previously they waited to display warning message until the wrong-way driver was visually verified). Two of these procedures (code in the SAPD CAD system and Texas DOT logging all WWD events) created databases that could be used to determine the WWD trends in San Antonio. Institutional actions also included site reviews of select freeway exit ramps around San Antonio, an ongoing effort that involves staff from Texas DOT, the City of San Antonio Public Works Department, and TTI, and employs a site review checklist developed during previous research (Cooner et al. 2004a). Finley et al. (2014) used the WWD subcomponent of SAPD 911 call log data to determine the effectiveness of the LED border-illuminated WRONG WAY signs. Using 14 months of before data and 22 months of after data, TTI researchers found a 38 percent reduction in WWD events on the US 281 cor- ridor after the installation of the LED border-illuminated WRONG WAY signs. This percent change was statistically significant at a 5 percent significance level (the 95 percent confidence interval ranged from a 63 percent reduction to a 13 percent reduction). Finley et al. (2014) also reviewed the Texas DOT Trans- Guide operator log data and found that for 87 percent of the WWD events documented since March 2011, there was no crash or the driver of the vehicle was not apprehended. Of those WWD events where the driver was apprehended or a crash occurred (92), 67 percent were attributed to alcohol impairment and approximately 1 percent were disoriented elderly drivers. For over three-quarters of the WWD events, Texas DOT was able to post a WWD warning message on at least one DMS in the area. This message warned drivers of the potential for a wrong-way driver in the area. While extensive human factors and traffic operations research on DMS message design has been previously con- ducted, these efforts have not looked at the design of WWD warning messages. Therefore, Finley et al. (2014) used the focus group discussion method to obtain motorists’ opinions regarding the design of WWD warning messages for DMSs. Researchers also reviewed previous DMS message design

A-10 literature and manuals to gain insight into the design of WWD warning messages. Based on the findings, researchers suggested the single-phase message shown in Figure A-1. During a WWD event, the location of a wrong-way driver and the lanes affected can be difficult to verify and can change rather quickly. In addition, the focus group results showed that motorists understand the dynamic nature of the situa- tion and difficulty with providing this type of information in a timely manner. Texas DOT personnel also noted that even if the operators have the capability to monitor a wrong-way driver via camera, these operators do not always have time to continuously update a DMS message. Instead, the prior- ity is to convey information to law enforcement so they can apprehend the wrong-way driver. Therefore, researchers did not include location information in the suggested message. The focus group results also showed that WRONG WAY adequately conveyed the effect on travel (i.e., motorists might encounter a wrong-way driver) and the proper driving action (i.e., motorists should slow down and proceed with caution). In addition, most of the focus group participants did not think drivers should be told to do a specific driving action (e.g., pull over to the shoulder or exit the freeway). Again, due to the dynamic nature of the situation, providing a spe- cific action would be difficult. Thus, researchers also did not include an action statement in the suggested message. Finley et al. (2014) also suggested that anytime one of these messages is displayed on a DMS, the beacons located on the DMS should be activated. If the DMS does not have beacons, the entire message may be flashed. One line of these messages should never be flashed. The suggested message should be posted on DMSs whenever a wrong-way driver is reported, even if not confirmed, in order to alert motorists to the pos- sibility of a wrong-way driver as soon as possible. When a wrong-way driver is confirmed, there is no need to change the third line of the messages. The message should be displayed along the entire length of the roadway in both directions of travel, and should be displayed until the wrong-way driver is apprehended or the report is canceled. The Michigan DOT also implemented an initiative to address serious crashes that included low-cost counter- measures to deter wrong-way movements onto freeways (Michigan DOT 2012; Morena and Leix 2012). In 2012, the Michigan DOT began implementing several low-cost safety improvements over a 5-year period at locations where wrong- way maneuvers were more frequently observed (i.e., a partial cloverleaf configuration). These improvements included: • Lowered height of DO NOT ENTER and WRONG WAY signs, • Reflective sheeting on the supports of lowered signs, • Stop lines at exit ramps, • Wrong-way pavement marking arrows, • Left-turn pavement marking guides, • Painted islands between exit and entrance ramps, and • Increased two-sided delineation along the exit ramp. In October 2012, the Illinois Center for Transportation fin- ished a study for the Illinois DOT related to WWD on freeways (Zhou et al. 2012). The research included analysis of wrong- way crashes in Illinois over a 6-year period to determine the contributing factors to wrong-way crashes on freeways and the development of promising, cost-conscious counter measures to reduce the WWD errors and their associated crashes. Researchers developed a method to rank the high-frequency crash locations based on the number of recorded or estimated wrong-way freeway entries. Interchanges were identified for field reviews, with site-specific and general countermeasures identified for future implementation. Some of the wrong- way countermeasures identified for implementation included: • Larger DO NOT ENTER and WRONG WAY signs, • Red reflective sheeting on sign supports, • WRONG WAY signs with flashing LEDs around the border at high-frequency crash locations, and • Pavement marking and geometric design enhancements at on-off ramp configurations. Also in 2012, the Ohio DOT finalized systematic upgrades of DO NOT ENTER, WRONG WAY, and ONE WAY signs to the 2012 Ohio Manual on Uniform Traffic Control Devices standards (Ohio DOT 2012). The work was part of an ongo- ing sign replacement program, which provided for traf- fic control signs to be replaced regularly to assure adequate nighttime visibility. As part of this effort, Ohio DOT: • Upgraded signage along freeway and expressway inter- changes to enhance the visibility of signage for wrong-way drivers, • Installed supplemental WRONG WAY signs at 3-ft mount- ing heights on the non-cloverleaf exit ramps, and • Installed dual directional route marker assemblies at the ramp ends and pavement marking arrows for positive guid- ance on the entrance ramps for interchanges where the entrance and exit ramps are side by side. In 2013, the New York State Thruway Authority installed an ITS-based warning system at one exit along I-190 in WARNING WRONG WAY DRIVER REPORTED Source: Finley et al. 2014. Figure A-1. Suggested DMS WWD message.

A-11 Buffalo (American Traffic Safety Services Association 2014). The system used Doppler radar detection and a small DMS that activated after a wrong-way vehicle was detected. The initial installation displayed the following sequence of mes- sages to a wrong-way driver via the DMS: WRONG WAY, STOP, and PULL OVER. The system was also designed to send alerts to the TMC, police, and other DMSs along the main lanes. The cost of the system was $10,000 per sign, inclusive of development and testing. Also in 2013, the first National Wrong-Way Driving Summit was held in Illinois. Approximately 130 people from 23 states attended the summit. These individuals represented NTSB, FHWA, American Traffic Safety Services Association, state departments of transportation, enforcement, tollway authorities, universities, and consulting firms. Based on the results of a survey distributed to the attendees, group discussion, and presentations, researchers summarized the various WWD countermeasures implemented by different agencies (see Zhou and Rouholamin 2014b; Zhou and Pour- Rouholamin 2015; Pour-Rouholamin et al. 2014). Sign- ing countermeasures included lowering sign height, using oversized signs, mounting multiple signs on the same pole, applying red retroreflective strips to sign posts, and using FREEWAY ENTRANCE signs. Pavement marking counter- measures included stop lines, wrong-way arrows, lane-use arrows, red retroreflective raised pavement markers (RRPMs), and short-dashed lane delineation through turns. ITS tech- nologies implemented included LED border-illuminated signs, use of DMS to warn right-way drivers, and use of global positioning system (GPS) navigation technologies to provide wrong-way movement alerts. Other countermeasures included entrance/exit ramp separation, raised curb medians, longitudi- nal channelizers, and changing ramp geometrics. With respect to the survey, 16 states responded, including states that had already implemented and tested various WWD countermeasures and those planning to address WWD in the future. Half of the survey participants had recently con- ducted WWD studies in their jurisdictions, and many more were planning to begin their evaluations soon. In addition, one-third of the agencies were using ITS technologies to detect wrong-way drivers and alert the wrong-way driver and/or right-way drivers. The survey findings showed that all of the entities use WRONG WAY signs on exit ramps and all but two entities install DO NOT ENTER signs on exit ramps. The use of these signs on frontage roads is not as consistent (about 70 percent for DO NOT ENTER signs and 56 percent for WRONG WAY signs). On divided highways, 81 percent of the participants use DO NOT ENTER signs and 75 percent use WRONG WAY signs. Most of the participants (85 percent) use identical signs on both sides of the roadway and increase the size of the signs (77 percent). In addition, most of the respondents (81 per- cent) install DO NOT ENTER and WRONG WAY signs at the standard height. Nearly half of the entities surveyed have lowered the height of these signs in special conditions. About 40 percent of the entities surveyed install various combina- tions of WRONG WAY and DO NOT ENTER signs on the same sign post assembly. Sixty-two percent of the partici- pants add a strip of retroreflective material to the sign post. The survey also found that 63 percent of the respondents face these signs perpendicular to the roadway instead of orienting them toward the target user. Roughly 70 percent of the participants use the wrong-way arrow, and most of these entities place the arrow on the exit ramp near the intersection with the cross street (71 percent) and at the middle of the exit ramp (64 percent). In some cases, the wrong-way arrow is also located on the exit ramp near the gore area off the main lanes (21 percent) and on the main lane (7 percent). In addition, more than half of the states use red RRPMs to supplement or replace the wrong-way arrow. The NTSB (2012) special investigation report aimed to identify relevant safety recommendations to prevent wrong- way collisions on highways and access ramps on a national level. As discussed previously, the report characterized WWD in the United States and summarized nine NTSB wrong-way collision investigations. In addition, the report made the fol- lowing suggestions to address wrong-way collisions: • Installation of alcohol ignition interlocks on the vehicles of all driving-while-intoxicated offenders, • Widespread implementation of new in-vehicle alcohol detection technologies in U.S. vehicles, • Use of traffic control devices to make exit ramps more dis- tinguishable from entrance ramps, • Use of wrong-way monitoring programs to identify wrong- way drivers, • Use of navigation system alerts in the vehicle to inform drivers that they have performed a wrong-way movement, • Development of an assessment tool that states can use to select appropriate countermeasures, and • Development of a best practices guide for law enforcement on how to respond to a wrong-way driver. Since alcohol has been found to be the primary contrib- uting factor in many wrong-way crash studies, Finley et al. (2014) investigated the behaviors of alcohol-impaired drivers through two nighttime closed-course studies. While data in response to a simulated environment cannot be directly com- pared to data collected on an actual road, closed-course study data can be used to compare the relative differences in perfor- mance between the various treatments evaluated. The studies were designed to: • Determine where alcohol-impaired drivers look in the for- ward driving scene,

A-12 • Provide insight into how alcohol-impaired drivers recog- nize and read signs, and • Assess the conspicuity of select WWD countermeasures from the perspective of alcohol-impaired drivers. Finley et al. (2014) found that alcohol-impaired drivers tend to look less to the left and right and more toward the pavement area in front of the vehicle. In addition, research- ers confirmed that alcohol-impaired drivers do not actively search the forward driving scene as much as non-impaired drivers. Instead, alcohol-impaired drivers concentrate their glances in a smaller area within the forward driving scene. Finley et al. (2014) also confirmed that drivers at higher BAC levels took longer to locate signs and must be closer to a sign before they can identify the background color and read the legend. In addition, alcohol-impaired drivers have to be closer to signs with flashing red LEDs around the border before they can read the legend compared to signs without flashing LEDs. Researchers also found that as the BAC level increased, more drivers misidentified the red sign background color, with most thinking that a red sign was an orange sign. Lowering the height of the white-on-red signs studied did not improve the ability of alcohol-impaired drivers to locate signs, identify the background color, or read the legend com- pared to the standard sign height (7 ft). Making the sign larger (i.e., oversized), adding red retroreflective sheeting to the sign support, or adding flashing red LEDs around the border of the sign also did not improve the ability of the alcohol- impaired drivers to locate WRONG WAY signs. However, the participants felt that these three countermeasures made it easier to find the WRONG WAY sign. The participants also thought that these three countermeasures caught their atten- tion more than the lowered WRONG WAY sign and the nor- mal size WRONG WAY signs without a conspicuity element. In an effort to reduce the occurrence of missing RRPMs, Finley et al. (2014) modified the design of the current RRPM wrong-way arrow that Texas DOT uses. Researchers did not find a significant difference in the recognition time between the two wrong-way arrow marking designs. In addition, the participants similarly assessed the ease with which they could find the arrow among the other markings. Thus, it appeared that the modified design performed as well as the current design. Researchers also found that at higher BAC levels, the participants took longer to locate the wrong-way arrow pave- ment markings, independent of the design, among the other markings. Overall, Finley et al. (2014) concluded that a wide variety of countermeasures and mitigation methods are needed to combat WWD on freeways. However, based on the findings of this research and anecdotal evidence, Finley et al. suspect that highly intoxicated drivers will not be attracted to or understand most traditional countermeasures and possibly even some innovative dynamic countermeasures. Therefore, Finley et al. also concluded that WWD detection systems are needed. Detection systems can be used to detect wrong-way drivers as they enter the freeway and/or on the main lanes. In addition, detection systems provide data regarding actual wrong-way driver entry points, a critical piece of informa- tion that is needed to help practitioners further combat WWD. Detection systems, in conjunction with cameras, can also provide data about wrong-way drivers that self-correct before reaching the main lanes. These data would help prac- titioners further assess the effectiveness of implemented countermeasures. Finley et al. compiled a catalog of known WWD countermeasures and mitigation methods being used or under research in the United States. Furthermore, Finley et al. developed guidelines for Texas DOT districts to follow to assess the WWD issue and implement countermeasures. These guidelines are shown in Figure A-2. In 2014, Zhou and Rouholamin (2014a) also developed guidelines for reducing wrong-way crashes on freeways. Their document also contains guidance regarding geometric design elements in general and for some specific interchange designs, as well as information regarding enforcement and education. The Handbook for Designing Roadways for the Aging Popu- lation (Brewer et al. 2014) recommends: • The use of larger-than-minimum-size DO NOT ENTER and WRONG WAY signs to increase the letter size, • The use of retroreflective fluorescent red sheeting material to increase sign conspicuity and legibility for aging drivers, • The placement of DO NOT ENTER and WRONG WAY signs on both sides of the road, • The lowering of sign height to 36 inches above the pave- ment to maximize brightness under low-beam headlights, where all other engineering options have been tried or considered, • The application of wrong-way arrow pavement markings near the terminus of all exit ramps, and • The use of red/white bidirectional RRPMs to supplement wrong-way arrow pavement markings where engineering judgment indicates a need for increased conspicuity. The American Traffic Safety Services Association (2014) has also published a document that contains various case studies regarding emerging WWD countermeasures and a wrong-way driving road safety audit prompt list developed by FHWA. Most of the case studies included in this document have already been discussed, so a simple list of the counter- measures is provided below. • Low-mounted DO NOT ENTER and WRONG WAY signs, • LED border-illuminated WRONG WAY signs,

A-13 • Red reflective strips on DO NOT ENTER and WRONG WAY sign posts, • Red RRPMs, • Access management, • ITS detection systems, • Oversized DO NOT ENTER and WRONG WAY signs, and • Enhanced pavement markings. In 2016, the ENTERPRISE Pooled Fund Study published a report that summarized the current practice of wrong-way countermeasures on freeways, including those that use ITSs (Athey Creek Consultants 2016). This report includes details on wrong-way countermeasures used by 13 agencies, as well as evaluation plans, lessons learned, and standards/ policies. Many of the countermeasures included in this report have already been described; thus, details are not provided herein. Some additional countermeasures include using dual directional route marker sign assemblies at exit ramps (Ohio) and replacing traditional green ball signal indications with green straight arrow indications (Rhode Island). Researchers concluded that determining the effectiveness of wrong-way countermeasures is challenging due to the random nature of wrong-way crashes, lack of data before the counter measures were implemented, inconsistency in countermeasure deploy- ment, simultaneous installation of multiple counter measures, and lack of agency resources. More recently, Ponnaluri (2016) presented the Florida DOT’s policy-oriented framework toward addressing Source: Finley et al. 2014. • Include local agencies • Share information about WWD activity • Use data to determine locations/corridors where WWD occurs Form Task Force • Conduct field reviews of exit ramps at identified locations/corridors • Inspect condition of signing and pavement markings during the day and at night • Identify other items in the surrounding area that may impact WWD activity • Correct any identified traffic control device deficiencies Perform Field Reviews • Install or repair wrong-way arrows on exit ramps • Install red retroreflective sheeting on WRONG WAY and DO NOT ENTER sign supports • Install oversized WRONG WAY and DO NOT ENTER signs • Install WRONG WAY signs with red flashing lights around the border • Display wrong-way driver warning messages on DMSs Implement Low-Cost Countermeasures • Install at identified locations or along identified corridors • Use data to identify and respond to wrong-way entries • Use data to assess effectiveness of countermeasures • Use data to warrant implementation of additional countermeasures and mitigation methods Consider Detection Systems • Use access management near exit ramps • Make geometric modifications to downstream intersection • Make geometric modifications to interchange Consider Higher-Cost Countermeasures Figure A-2. Guideline recommendations for Texas DOT districts.

A-14 WWD in a systematic manner. Florida DOT’s method included: • Implementing pilot projects, • Conducting a statewide study with crash evaluation and field reviews, • Evaluating and deploying experimental devices, • Conceptualizing a human factors study, • Transforming recommendations to design guidance, • Discussing with planners the interchange types susceptible to WWD, • Retrofitting exit ramps with the recommended counter- measures, and • Leveraging the media to promote awareness and to educate the public. As part of this process, Florida’s Turnpike Enterprise (FTE) began a pilot effort on an 18-mi section of the Homestead Extension. In Phase 1, the DO NOT ENTER, WRONG WAY, ONE WAY, NO LEFT/U-TURNS, and KEEP RIGHT signs were replaced with their respective oversized signs. Wrong- way arrows were also added along exit ramps. In Phase 2, FTE implemented vehicle-alerting technology on the main lanes. In Phase 3, FTE installed LED border-illuminated WRONG WAY signs with vehicle detection and enhanced Florida DOT SunGuide® software. Florida DOT District 3 also installed additional signs and pavement markings, as well as vehicle-activated blank-out WRONG WAY signs at four interchanges along Interstate 10. Large WRONG WAY signs were also installed on existing overhead guide sign trusses over the main lanes, and interstate pavement shields with straight arrows were added on the arterial turn lanes. Simi- lar to NTTA, Florida DOT also decided to replace left-turn pavement markings, when they precede the turn lanes, with a combination of interstate pavement shield, cardinal direc- tion, and straight arrows. While these efforts did not include a post-deployment evaluation, Florida DOT and FTE offices are closely monitoring citizen response, media interest, and law enforcement input. Ponnaluri concluded that no one sign, pavement marking, or technology by itself can alert a driver to a potential wrong-way maneuver, but their com- bined effect seems to be effective. The Florida DOT Traffic Engineering and Research Lab- oratory also received permission from FHWA to experiment with red rapid rectangular flashing beacons (RRFBs) and a three-row pattern of in-pavement red internally illuminated RPMs placed along exit ramps. Studies to assess the effective- ness of these emerging countermeasures are currently under- way. Florida DOT is also currently sponsoring a statewide WWD study to understand the reasons why drivers enter freeways going the wrong direction. The Central Florida Expressway Authority (CFX) has also assessed the extent of wrong-way crashes on CFX tollways. Based on the findings, the University of Central Florida inves- tigated and tested potential WWD countermeasures and developed a field inspection checklist for inspectors to use to ensure that the standard signs and pavement markings were installed and properly maintained (Al-Deek et al. 2015). After receiving permission to experiment from FHWA in October 2014, CFX installed red rectangular flashing beacons (RFBs) above and below WRONG WAY signs at five ramps (one on SR 520 and four on SR 408). In addition, CFX installed two cameras (one forward-facing and one side facing) and two radar sensors (one forward-facing and one rear facing) at each of these ramps. When a wrong-way vehicle is detected, the RFBs will flash in a wig-wag pattern to warn the driver, and an alert will be sent to the nearby TMC. In 2015, TTI and the Southwest Research Institute began working with Texas DOT to develop CV applications that would detect wrong-way vehicles, notify traffic management agencies and law enforcement, and alert affected travelers. In Phase I, the research team reviewed the state of the practice regarding ITS and CV technologies being applied as wrong- way driving countermeasures. The research team then identi- fied user needs associated with the implementation of a CV wrong-way driving system, assessed motorist understanding of wrong-way driver warning messages posted on dynamic message signs, and ascertained preliminary ways to connect with law enforcement. Phase I culminated in the develop- ment of a concept of operations, functional requirements, and high-level system design for a CV test bed for wrong-way driving applications (Finley et al. 2016a). In Phase II, the research team developed a proof-of-con- cept CV wrong-way driving detection and management sys- tem at a closed-course facility. The purpose of the test bed was to provide an off-roadway location to test and fine-tune the system components and operations prior to installing them on an actual roadway. As part of the prototype development, the research team generated a detailed system design based on requirements that were established in Phase I. The research team then procured the hardware components needed to build the prototype system. Furthermore, the research team developed detailed system architecture, integrated hardware and software components, performed validation testing, and conducted a demonstration of the system. In Phase II, the research team also conducted human factors studies to inves- tigate the in-vehicle information needs of right-way drivers when a wrong-way driving event occurs. Researchers con- ducted a formal task analysis to identify critical stages where right-way drivers could make a better decision if information was provided to them through connected vehicle technology. Researchers then used structured interviews and surveys to identify the information needs of right-way drivers and eval- uate comprehension and preference of message wording and timing (Finley et al. 2017).

A-15 Summary The research team reviewed previous domestic and inter- national research to identify the characteristics of wrong-way crashes and summarize the use of traffic control devices and ITS technologies to deter wrong-way movements on freeways and divided highways. Overall, to date, most of the WWD research has focused on freeways; thus, very little is known about the characteristics of wrong-way crashes and effective- ness of countermeasures on divided highways. The key find- ings regarding wrong-way crashes were: • Wrong-way crashes on freeways and divided highways tend to be more severe (i.e., a greater proportion resulting in a fatality or serious injury than other types of crashes), • Wrong-way crashes on freeways tend to occur more fre- quently in urban areas, at night, and on the weekend, • The primary origin of wrong-way movements on freeways is entering an exit ramp, • Most wrong-way crashes on freeways occur in the lane closest to the median, • The primary origin of wrong-way movements on divided highways is at intersections with median openings, • Wrong-way drivers tend to be young males, • Driving under the influence of alcohol and/or drugs is the primary contributing factor in most of the wrong-way crashes, and • Elderly drivers are overrepresented in wrong-way collisions. Other contributing factors include interchange design, inclement weather, driver inattention, driver confusion, physical or mental illness, insufficient lighting, and insuf- ficient sight distance. Wrong-way maneuvers also occur because of intentional driver behaviors, such as driving backward, making U-turns, and attempting suicide. Many traffic control devices and ITS technologies have been developed and tested since the 1960s. However, simi- lar to the crash data analyses, most implementations and evaluations have occurred on freeways. Table A-1 contains a summary of the traffic control devices and associated ITS/ technology systems documented in previous research. Preventive ITS/Technology Traditional signs • DO NOT ENTER • WRONG WAY • ONE WAY • Divided highway crossing • Movement prohibition • Freeway entrance • KEEP RIGHT • Directional route marker assembly Enhanced static signs • Additional signs (mounted separately or on the same sign post) • Oversized signs • Lower mounting height • Red retroreflective tape on sign post Active signs • Blank-out • Beacons • LEDs around sign border • RRFBs/RFBs (experimental) Pavement markings • Wrong-way arrows • Lane-use arrows • Stop lines • Lane line extensions for turning movements • Route designation shields Active pavement markings • In-pavement lighting (experimental) Other traffic control devices • Red delineators along exit ramp • Green straight arrow signal indications Detects wrong-way driver • Sensors • Cameras Warns wrong-way driver • Active signs • Pavement markings • In-vehicle (emerging approach) Notifies TMC and law enforcement Verifies and/or monitors event • Sensors • Cameras • TMC operator • GPS Warns right-way drivers • DMS • In-vehicle (emerging approach) Table A-1. Summary of traffic control device and ITS/technology wrong-way driving countermeasures.

A-16 References Al-Deek, H., J. Rogers, A. Sandt, and A. Alomari. 2015. Wrong-Way Driving Incidents on CFX Toll Road Network, Phase-2 Study: Devel- oping Countermeasures. Final Report Version 3.0, University of Central Florida, Orlando, Florida. AASHTO. 2009. Manual for Assessing Safety Hardware, First Edition. Washington, D.C. American Traffic Safety Services Association. 2014. Emerging Safety Countermeasures for Wrong-Way Driving. Fredericksburg, Virginia. Athey Creek Consultants. 2016. Countermeasures for Wrong-Way Driving on Freeways. Project Summary Report. ENTERPRISE Transporta- tion Pooled Fund Study TPF-5 (231). Baratian-Ghorghi, F., H. Zhou, and J. Shaw. 2014. “Overview of Wrong- Way Driving Fatal Crashes in the United States.” ITE Journal, Vol. 84, No. 8, pp. 41–47. Brewer, M., D. Murillo, and A. Pate. 2014. Handbook for Designing Road- ways for the Aging Population. FHWA, U.S. DOT, Washington, D.C. Cooner, S. A., A. S. Cothron, and S. E. Ranft. 2004a. Countermeasures for Wrong Way Movement on Freeways: Overview of Project Activities and Findings. Report No. FHWA/TX-04/0-4128-1, Texas Transpor- tation Institute, College Station, Texas. Cooner, S. A., A. S. Cothron, and S. E. Ranft. 2004b. Countermeasures for Wrong Way Movement on Freeways: Guidelines and Recommended Practices. Report No. FHWA/TX-04/4128-2, Texas Transportation Institute, College Station, Texas. Copelan, J. 1989. Prevention of Wrong Way Accidents on Freeways. Report No. FHWA/CA-TE-89-2, California Department of Transporta- tion, Traffic Operations Division, Sacramento, California. Doty, R. N., and C. R. Ledbetter. 1965. Full Scale Dynamic Tests on One-Way Spike Barriers. Traffic Department, Division of High- ways, Department of Public Works, State of California. Finley, M. D., S. P. Venglar, V. Iragavarapu, J. D. Miles, E. S. Park, S. A. Cooner, and S. E. Ranft. 2014. Assessment of the Effectiveness of Wrong Way Driving Countermeasures and Mitigation Methods. Research Report 0-6769-1, Texas A&M Transportation Institute, College Station, Texas. Finley, M. D., K. N. Balke, R. Rajbhandari, S. T. Chrysler, C. S. Dobrovolny, N. D. Trout, P. Avery, D. Vickers, and C. Mott. 2016a. Conceptual Design of a Connected Vehicle Wrong-Way Driving Detection and Man- agement System. Research Report 0-6867-1, Texas A&M Transporta- tion Institute, College Station, Texas. Finley, M. D., S. P. Venglar, and Y. Ouyang. 2016b. “Operational Field Studies of Two Wrong-Way Signing Countermeasures.” Paper No. 16-1827. In Conference Proceedings: 95th Annual Transporta- tion Research Board Meeting, Washington, D.C. Finley, M. D., S. T. Chrysler, K. N. Balke, H. Charara, D. H. Florence, R. Rajbhandari, C. Mott, P. K. Sturgeon II, and D. M. Parish. 2017. Development and Testing of a Prototype Connected Vehicle Wrong- Way Driving Detection and Management System. Research Report 0-6867-01-1, Texas A&M Transportation Institute, Col- lege Station, Texas. Friebele, J. D., C. J. Messer, and C. L. Dudek. 1971. State-of-the-Art of Wrong Way Driving on Freeways and Expressways. Report No. 139-7, Texas Transportation Institute, College Station, Texas. Institute for Traffic Accident Research and Data Analysis. 2002. High- way Accidents Involving Dangerous Wrong-Way Traveling. Informa- tion Bulletin No. 36, Japan. ITS International. 2010. “Wrong Way Detection System Prevents Acci- dents, Improves Safety.” http://www.itsinternational.com/sections/ cost-benefit-analysis/features/wrong-way-detection-system-prevents- accidents-improves-safety/ (as of October 30, 2017). Kemel, E. 2015. “Wrong-Way Driving Crashes on French Divided Roads.” Accident Analysis and Prevention, Vol. 75, pp. 69–76. Kittelson & Associates, Inc. 2015. Statewide Wrong Way Crash Study Final Report. Florida DOT Central Office, Tallahassee, Florida. Knight, D. 1983. “Danger! Wrong Way Driver Ahead!” Traffic Safety (Chicago), Vol. 83, No. 5. Lathrop, S., T. B., Dick, and K. B. Nolte. March 2010. “Fatal Wrong Way Collisions on New Mexico’s Interstate Highways, 1990–2004.” Jour- nal of Forensic Sciences, Vol. 55, No. 2, pp. 432–437. Lew, A. 1971. Final Report on Wrong-Way Driving. Phase III Driver Characteristics, Effectiveness of Remedial Measures, and Effect of Ramp Type. State of California, Division of Highways, Traffic Department. Messer, C. J., J. D. Friebele, and C. L. Dudek. 1971. A Qualitative Analysis of Wrong Way Driving in Texas. Report No. 139-6, Texas Transporta- tion Institute, College Station, Texas. Michigan DOT. 2012. Highway Safety White Paper: 2035 MI Trans- portation Plan: State Long-Range Transportation Plan. Lansing, Michigan. Moler, S. September/October 2002. “Stop. You’re Going the Wrong Way!” Public Roads, Vol. 66, No. 2. Morena, D., and T. Leix. May/June 2012. “Where These Drivers Went Wrong.” Public Roads, Vol. 75, No. 6. NTSB. December 11, 2012. Highway Special Investigation Report: Wrong Way Driving. Report NTSB/SIR-12/01, Washington, D.C. North Carolina DOT. 2006. Wrong Way Crashes in North Carolina Update. Traffic Safety Unit, Transportation Mobility and Safety Division, Raleigh, North Carolina. Ohio DOT. 2012. ODOT to Enhance Wrong Way Signage. http://www. dot.state.oh.us/districts/D02/newsreleases/Pages/ODOT.aspx (as of October 30, 2017). Parsonson, P., and J. Marks. 1979. Wrong-Way Traffic Movements on Freeway Ramps. Georgia Institute of Technology, School of Engi- neering, Atlanta, Georgia. Ponnaluri, R. V. 2016. “Addressing Wrong-Way Driving as a Matter of Policy: The Florida Experience.” Transport Policy, Vol. 46, pp. 92–100. Pour-Rouholamin, M., H. Zhou, J. Shaw, and P. Tobias. December 2014. “Overview of Safety Countermeasures for Wrong-Way Driving Crashes.” ITE Journal, Vol. 84, No. 12, pp. 31–38. Pour-Rouholamin, M., H. Zhou, B. Zhang, and R. E. Turochy. 2016. “Comprehensive Analysis of Wrong-Way Driving Crashes on Ala- bama Interstates.” Transportation Research Record: Journal of the Transportation Research Board, No. 2601, Transportation Research Board of the National Academies, Washington, D.C., pp. 50–58. Rinde, E. 1978. Off-Ramp Surveillance: Wrong Way Driving. Report No. FHWA-CA-TE-78-1, California Department of Transporta- tion, Office of Traffic, Sacramento, California. Schrock, S. D., H. G. Hawkins, Jr., and S. T. Chrysler. 2005. “Effective- ness of Lane Direction Arrows as Pavement Markings in Reducing Wrong-Way Movements on Two-Way Frontage Roads.” Trans- portation Research Record: Journal of the Transportation Research Board, No. 1918, Transportation Research Board of the National Academies, Washington, D.C., pp. 63–67. Scifres, P., and R. Loutzenheiser. 1975. Wrong Way Movements on Divided Highways. Purdue University Joint Highway Research Project No. JHRP-13-75, West Lafayette, Indiana. Simpson, S., and D. Bruggeman. 2015. Detection and Warning Systems for Wrong-Way Driving. Arizona Department of Transportation, Phoenix, Arizona. SWOV Institute for Road Safety Research. 2007. SWOV Fact Sheet: Wrong-Way Driving. Leidschendam, The Netherlands.

A-17 Tamburri, T. 1965. Report on Wrong Way Automatic Sign, Light and Horn Device. Traffic Department, Division of Highways, Depart- ment of Public Works, State of California. Tamburri, T. N., and D. J. Theobald. 1965. Wrong Way Driving (Phase II). Traffic Department, Division of Highways, Department of Public Works, State of California. Vaswani, N. 1973. Measures for Preventing Wrong Way Entries on Highways. Virginia Highway Research Council, Charlottesville, Virginia. Vaswani, N. K. 1977a. Further Reduction in Incidences of Wrong Way Driving. Virginia Highway and Transportation Research Council, Charlottesville, Virginia. Vaswani, N. K. 1977b. Experiments with a Divided Highway Crossing Sign to Reduce Wrong-Way Driving. Virginia Highway Research Council, Charlottesville, Virginia. Williams, C. 2007. District 3 ITS Project Update Presentation. Florida DOT. http://www.ce.siue.edu/faculty/hzhou/ww/paper/District3 %20ITS%20Engineer.pdf (as of October 30, 2017). Xing. J. 2013. “Characteristics and Countermeasures Against Wrong- Way Driving on Motorways in Japan.” In Conference Proceedings for the 20th ITS World Congress, Tokyo, Japan. Zhang, B., M. Pour-Rouholamin, and H. Zhou. 2016. “Investigation of Confounding Factors Contributing to Wrong-Way Driving Crashes on Partially/Uncontrolled-Access Divided Highways.” Paper No. 17-05146. In Conference Proceedings: 96th Annual Transportation Research Board Meeting, Washington, D.C. Zhou, H., and M. P. Rouholamin. 2014a. Guidelines for Reducing Wrong-Way Crashes on Freeways. Report FHWA-ICT-14-010, Illi- nois Center for Transportation, Department of Civil and Environ- mental Engineering, Urbana, Illinois. Zhou, H., and M. P. Rouholamin. 2014b. Proceedings of the 2013 National Wrong-Way Driving Summit. Report FHWA-ICT-14-009, Illinois Center for Transportation, Department of Civil and Envi- ronmental Engineering, Urbana, Illinois. Zhou, H., and M. Pour-Rouholamin. 2015. Investigation of Contributing Factors Regarding Wrong-Way Driving on Freeways, Phase II. Report FHWA-ICT-15-016, Illinois Center for Transportation, Depart- ment of Civil and Environmental Engineering, Urbana, Illinois. Zhou, H., J. Zhao, and R. Fries. 2012. Investigation of Contribut- ing Factors Regarding Wrong Way Driving on Freeways. Report FHWA-ICT-12-010, Illinois Center for Transportation, Depart- ment of Civil and Environmental Engineering, Urbana, Illinois.