An innovative Michigan study sheds light on engineering strategies to curtail the number and severity of wrong-way crashes on freeways.

Across the country, crashes caused by wrong-way drivers are few and far between, but when they do occur, they often provide fodder for terrifying and heartbreaking headlines. These crashes kill or severely injure drivers and passengers at a much greater rate (per crash) than other types of freeway incidents. It stands to reason that in a wrong-way head-on crash, the highest potential for injury and death occurs on the roads where drivers travel at the highest speeds: the Nation's freeways. For instance, a study published by the California Department of Transportation and Federal Highway Administration (FHWA), Prevention of Wrong-Way Accidents on Freeways (FHWA/CA-TE-89-2), found that the fatality rate was 12 times greater for wrong-way crashes compared to all other crashes on California freeways in 1987.

In 2010 and 2011, safety staff with FHWA and the Michigan Department of Transportation (MDOT) analyzed 110 wrong-way crashes that occurred on the Michigan freeway system during the 5-year period from 2005 to 2009. What they found regarding the characteristics of wrong-way drivers corroborated earlier studies, but what they found regarding the road system shed new light on the roadway engineering aspect of wrong-way freeway entries.

The safety researchers restricted their study to vehicles that were known or presumed to have entered the freeway system by traveling the wrong direction on an exit ramp. The team made every effort to exclude cross-median and other crashes in which a vehicle was traveling the wrong way by virtue of the driver losing control of the vehicle. Thus, an accurate description of the incidents researched in this study would be "wrong-way freeway entry" crashes.

Wrong-Way Crashes Make Horrific Headlines

The Wrong-Way Driver

The most noticeable characteristics of the wrong-way drivers in this study were their degree of impairment and a tendency toward late-night driving. The 110 crashes in the study included 9 for which the extent of impairment was unknown, either because they involved drive-aways in which the wrong-way driver was not identified, or fatal crashes in which impairment, if any, of the wrong-way driver was not documented.

Looking at the remaining 101 wrong-way drivers for which the extent of impairment was known, nearly 60 percent were under the influence of either alcohol or drugs in their systems when tested: 48 tested positive for alcohol, 7 for drugs, and 5 for both drugs and alcohol. The vast majority of these impaired-driver crashes (54 of 60) occurred at night.

For the full set of wrong-way crashes, the tendency toward night crashes held as well. Of the full 110 wrong-way crashes, a heavy concentration occurred late at night and early in the morning: 57 percent between 11 p.m. and 6 a.m. For comparison, only 16 percent of the total Michigan freeway crashes -- wrong-way entry and all others -- in 2005-2009 occurred during a similar late-night time period.

The late-night trend was even more pronounced when the researchers examined the 35 serious wrong-way crashes: 71 percent of the fatal and incapacitating wrong-way crashes occurred during the timeframe from 11 p.m. to 6 a.m., compared with only 23 percent for other serious freeway crashes in Michigan between 2005 and 2009.

In total, 78 percent (86 crashes) of the 110 study crashes occurred under conditions of darkness. This statistic makes sense because wrong-way freeway entry is the mistake of a confused driver, and darkness masks many of the roadside cues that are more visible in the daytime.

Driver age was captured for 104 of the wrong-way drivers and showed mostly equal distribution across the range of ages, with a slight concentration toward younger drivers: 18 (17 percent) of the wrong-way drivers were age 65 or older, but 24 (23 percent) of the wrong-way drivers were under the age of 25.

Another way to look at the issue of driver age is to remove the impaired drivers from the population. For the 41 wrong-way drivers who were not impaired at the time of the crash, the distribution of driver age is quite different from the total distribution. This group of 41 unimpaired drivers included 14 drivers age 65 or older (34 percent) and only 2 drivers under the age of 25 (5 percent). Although this is just a small number of crashes, the percentage of older drivers in the unimpaired group does stand out: In general, drivers 65 or older are involved in less than 12 percent of the total crashes in Michigan and approximately 14 percent of crashes that result in serious injury or death.

It appears from this analysis that some general potential for driver confusion leading to wrong-way entry exists across the entire population. This confusion is amplified in younger drivers who are impaired by alcohol or drugs, and amplified as well in older drivers (even without impairment).

The crash statistics also captured driver gender for 108 of the wrong-way drivers: 75 male, 33 female. This distribution is similar to historical percentages for gender in serious crashes in Michigan.

The Wrong-Way Crash

Not all of the studied crashes occurred on a freeway mainline. Of the 110 crashes, 31 took place on the exit ramp that provided the entry point for the wrong-way vehicle entering from the crossroad. In the 79 remaining crashes, the wrong-way driver reached the freeway mainline and either crashed on the mainline (71), or having traveled for some distance on the mainline, crashed on a freeway-to-freeway ramp (6) or on a freeway entrance ramp (2). In these cases, the researchers assumed that the wrong-way driver entered the freeway system through wrong-way movement on an earlier (but unidentified) exit ramp.

In most of the crashes (96 out of 110), the wrong-way vehicle hit another vehicle that was traveling in the appropriate direction. The remaining crashes were single-vehicle crashes in which the wrong-way vehicle hit a highway barrier (10 crashes), ran into a ditch, or simply overturned.

As a group, these wrong-way crashes were highly severe: 35 (32 percent) resulted in at least one fatality or incapacitating injury. For comparison, only 2 percent of all Michigan freeway crashes in 2005-2009 led to deaths or incapacitating injuries. Together, those 35 severe wrong-way crashes resulted in 30 fatalities and 36 incapacitating injuries -- a staggering price to pay for just 35 confused or impaired drivers.

Severity Depends on the Crash Location

The severity of an individual wrong-way crash was clearly linked to how far the wrong-way vehicle progressed onto the system, most likely due to the speed of the involved vehicles. Only 6 percent of the crashes that occurred on an exit ramp resulted in a death or incapacitating injury, while 42 percent of the mainline crashes resulted in a fatality or incapacitating injury.

This fact has implications for the engineering community. "Although engineering solutions to alert a wrong-way driver are most commonly applied at or near the crossroad -- and rightfully so -- it seems clear that any successful intervention along the entire length of the exit ramp is likely to be rewarded by a dramatic decrease in crash severity, should a collision occur," says Mark Bott, MDOT traffic and safety engineer.

Could Interchange Design Be an Issue?

Geographically, the locations of the 110 Michigan wrong-way crashes match up fairly well with the pattern and extent of freeway travel throughout the State. Detroit and the surrounding areas, which lay claim to 40 percent of the freeway vehicle miles traveled (VMT) in Michigan, were home to 47 percent (52) of the wrong-way crashes. Similarly, Grand Rapids and the surrounding areas experienced 18 of the crashes, and other areas of the State had smaller numbers, roughly correlating with the extent of their freeway VMT.

Of more interest is the type of interchange through which the wrong-way vehicle gained entry to the system. The purpose of an interchange is to route traffic onto and off the freeway from the crossroad. Freeways have a variety of interchange designs, and apparently not all designs (such as cloverleaf and diamond) are equal in terms of clarity of navigation to the motorist, and particularly to impaired, disoriented, or confused drivers.

Earlier research in California, North Carolina, and Washington State all suggested that the partial cloverleaf design, which puts an exit ramp adjacent to an entrance ramp on the crossroad, might be more conducive to wrong-way entry than other designs. The North Carolina and Washington State research also implicates full cloverleaves, and the California study points to potential navigation problems with full and partial diamond interchanges. Because the exact freeway entry points for the wrong-way drivers were unknown in most crashes investigated in these States, the researchers based their conclusions mainly on systemwide or corridorwide design features, not specific knowledge of individual entry ramps.

For the Michigan study as well, most of the crashes occurred on mainline freeways with no indication of how the wrong-way drivers entered the system. However, the Michigan dataset does contain 31 crashes that occurred on exit ramps, and 4 mainline crashes in which the wrong-way entry point was identified by the reporting police officer. Thus, for those 35 crashes, the wrong-way access ramp was known. Comparing those wrong-way access ramps to the total inventory of Michigan interchanges enabled the researchers to offer educated commentary on the culpability of various types of interchange ramps to the driver confusion that resulted in wrong-way entry.

Michigan has 791 freeway interchanges classified by MDOT as follows: directional (206), partial cloverleaf (163); tight diamond or modified diamond (154); diamond (136); urban diamond (50); trumpet (23); full cloverleaf (20); and others (39).

In the 35 Michigan crashes for which the wrong-way entry point was known, the wrong-way driver entered the freeway system at the following type of interchange: partial cloverleaf (21); trumpet (4); tight diamond (3); urban diamond (3); directional (2); full cloverleaf (1); other (1).

Based on these data, the researchers confirmed the concerns raised in earlier studies about the potential for confusion at partial cloverleaf designs. A partial cloverleaf interchange provided the wrong-way ramp entry for 60 percent of the known wrong-way Michigan drivers, even though that interchange type accounts for only 21 percent of the interchanges in the State. The trumpet interchange design is implicated as well, hosting 11 percent of the known wrong-way crashes even though trumpets comprise only 3 percent of the State's interchanges.

Another telling statistic is in the geographic spread of the study crashes. The 35 crashes with known wrong-way access points all occurred at different interchanges, with only 2 exceptions: 10 of the crashes occurred at one partial cloverleaf interchange, and 2 crashes occurred at a single trumpet interchange.

The Problem with Parallel Ramps

The primary interchange design implicated in this study -- partial cloverleaf -- has one predominant feature that appears to be the source of confusion leading to wrong-way freeway entry: a pair of freeway exit/entrance ramps that are adjacent and parallel to each other, and typically meet the crossroad at or near a 90-degree angle.

The wrong-way entry mode, then, for a disoriented, distracted, or otherwise confused driver is to turn onto the freeway exit ramp, thinking that he or she is entering onto the freeway entrance ramp. Although any entering traffic from the crossroad could be subject to confusion, it is the left turner who is presumed to be most susceptible to wrong-way entry in these situations. The left turner has to drive past the wrong ramp to reach the correct one, whereas the right turner encounters the correct ramp immediately. Not surprisingly, most of the engineering solutions that can be brought to bear on this problem involve positive cues to showcase the entrance ramp so that it looks like an entrance ramp, and negative cues that make the exit ramp appear uninviting to a potential wrong-way driver.

The Michigan Strategy

The study results were an eye-opener for MDOT staff. "The first thing I noticed in the data is that the severity of these crashes, as a group, is off the charts," says Bott.

MDOT officials wanted to move quickly but strategically on the issue. Given the design-specific findings from the study, they were able to do so. Of the 791 existing Michigan interchanges, MDOT staff identified 161 interchanges that exhibit the suspect feature of partial cloverleaf -- that is, adjacent and parallel ramps extending to the crossroad. These interchanges are being targeted for systematic installation of various low-cost countermeasures at an estimated cost of $2 million over the next 5 years.

Prevention of wrong-way freeway entry is not a new issue for MDOT. The department currently implements the nationally recognized standard approaches to wrong-way prevention: wrong-way marking arrows, DO NOT ENTER/WRONG WAY signing, and other recommended signs and markings to the degree required by the Manual on Uniform Traffic Control Devices (MUTCD). However, armed with the knowledge of the frequency and severity of actual wrong-way crashes in the State, MDOT now regards some treatments as mandatory that the MUTCD lists as optional. Also, MDOT is installing other treatments that are considered beyond the current state-of-the-art in wrong-way prevention.

Targeting the Bad Actors

When feasible, MDOT plans to apply the following seven countermeasures to the target ramps at the 161 identified interchanges over the next 5 years, using either State maintenance funds or Federal-aid safety funding as necessary:

• Lower the bottom height of DO NOT ENTER and WRONG WAY signs to 4 feet (1.2 meters). These signs typically are set at a bottom height of 7 feet (2.1 meters) to meet standard uniformity concerns regarding crash worthiness and sight distance. Neither of those concerns are paramount on ramps, but visibility of the DO NOT ENTER and WRONG WAY signs is. A lower sign is better positioned to catch a motorist's eye, both at night and during the day. California, Idaho, and Virginia are among the States that have been leading the way with this countermeasure, some of them since the early 1970s.
• Install reflective sheeting on the sign supports of DO NOT ENTER and WRONG WAY signs. Currently, reflective white strips are used in Michigan for additional marking on railroad crossbuck supports (both front and back). Reflective yellow strips have been growing in popularity to emphasize chevron signs and other warning signs, and red strips have been used on STOP signs at selected locations. This countermeasure extends the strategy to include the two sign messages that relate most directly to the prevention of wrong-way entry.
• Place stop bars at exit ramps. Current MDOT practice is to place stop bars on exit ramps where the intersection of the ramp and crossroad is signalized; the stop bar is optional at unsignalized intersections under stop control. MDOT will now consider placing stop bars on paired exit ramps in the target group of interchanges, regardless of the type of intersection control, as a further cue that the exit ramp is intended only for traffic leaving the freeway.
• Install wrong-way pavement marking arrows at exit ramps. This treatment, like some of the others, is listed as an optional measure in the MUTCD and is regarded as secondary to the more important lane-use pavement markings that are recommended for placement in each lane of an exit ramp, near the crossroad. MDOT had been following MUTCD guidance allowing for optional placement of the wrong-way arrow farther back along the ramp. Moving forward, the department plans to require wrong-way arrows at all target exit ramps, providing a second level of pavement marking warning to wrong-way drivers.
• Install pavement marking extensions that will guide crossroad left-turning traffic past the exit ramp and safely onto the entrance ramp. Pavement marking extensions are an established strategy that MDOT and other agencies use to guide traffic into the correct ramps at single-point urban interchanges, or to keep vehicles in their correct lanes during double-left-turn movements at standard intersections. This countermeasure extends the strategy to address what could be the primary failure mode for wrong-way entry at paired exit/entrance ramps -- the left turn into the first ramp (the exit ramp).
• Paint the island between the exit and entrance ramp for a sufficient distance up the ramp. This positive delineation of the island between the pair of ramps could prove helpful to a confused driver. Delineating an island indicates that there are two ramps, and that might be just enough to prompt driver realization that he or she does not want to be in the ramp to the left of the island.
• Place red delineators along the exit ramp to discourage wrong-way vehicles that are headed up the exit ramp. This strategy counts on some basic amount of driver recognition that red is used as a discouragement in several forms in the highway environment. Where guardrail is in place on a ramp, the red delineators can be attached to the guardrail either individually or in a continuous delineation system. MDOT already has experimented with several delineation strategies and prefers the continuous design. The bad news is that only a few ramps in the Michigan system have stretches of guardrail near the crossroad where this strategy could be immediately effective in stopping a driver who had just turned onto the ramp. The good news, according to the study's results, is that MDOT could apply this strategy as far up the ramp as the engineers feel is warranted as a means to discourage wrong-way drivers -- and ideally capture their attention in time to minimize the severity of wrong-way crashes. Also, in the absence of guardrail, MDOT is considering installing red delineators on delineator posts.

Eventual Improvement To All Interchanges

As MDOT directs resources to the 161 target interchanges it deems the most likely to host wrong-way entry, the agency nonetheless recognizes that the remaining 29 percent of the known freeway wrong-way entries in the study did not result from confusion on paired ramps. Wrong-way entry is a potential danger at all ramps, for a variety of driver behavior failure modes.

With that in mind, MDOT identified two of the seven low-cost countermeasures that could be cost effective at all ramps. Specifically, MDOT revised its signing standards to require the lower bottom height and to apply retroreflective sheeting strips to WRONG WAY and DO NOT ENTER signs at all exit ramps. This action will result in field changes at the remaining interchanges as they come up for routine work. Two countermeasures that are being applied to the target ramps -- painted medians and pavement marking extensions for left turns -- are not likely to be applicable to the nontarget ramps. The option to install the remaining treatments at nontarget ramps will be left as a field decision.

A Close Look at the Big Standout

Of all the interchanges in the State, one stood out in the Michigan study. The interchange of I-94 at Gratiot Avenue in Detroit was the site of 10 of the 35 known wrong-way vehicle entries. And unlike the general set of Michigan's wrong-way crash data, the 10 crashes at this interchange were not primarily at night and did not principally involve impaired drivers. Clearly, other contributing factors must have been involved here, and MDOT conducted a small-scale road safety audit in an attempt to identify what could be causing driver confusion at the two exit ramps of this interchange.

At this interchange, the exit and entrance ramps are in a partial cloverleaf configuration, side by side, and meet the crossroad at a single point, which is under the control of a traffic signal. This design occurs in the northwest and southeast quadrants, so there are two sets of partial cloverleaf ramps, and both quadrants were host to wrong-way vehicle crashes. Each of the two exit ramps carries two lanes to meet the crossroad. The crossroad itself (Gratiot) is five lanes, with a dedicated (and signalized) left-turn lane for traffic turning onto the entrance ramps.

Upon visiting the interchange ramps, the audit group quickly suggested a number of signing and marking treatments that, if implemented, would improve guidance to drivers searching for the entrance ramp and improve deterrence for drivers about to turn onto the exit ramp. These measures were along the lines of the low-cost countermeasures mentioned earlier: painting the median island and installing pavement marking extension lines to guide vehicles turning left from the crossroad, lowering the height of the DO NOT ENTER and WRONG WAY signs, and adding reflective red strips to those sign posts.

Because these exit ramps carry two lanes, the audit team recommended that lane assignment arrows (rather than wrong-way arrows) be applied to the mouth of the exit ramps. In addition, the group suggested enlarging the existing I-94 directional signs on the crossroad. The group also noted that the southbound stop bar on Gratiot Avenue was pulled back significantly from the intersection which, the audit members surmised, could give drivers the false impression that the left turn should be made into the exit ramp instead of the entrance ramp.

All of these suggested improvements are likely to be helpful, but these countermeasures mostly treat the symptoms of the problem at these ramps. That is, they attempt to mitigate the confusion that is already brewing in drivers' minds. The confusion itself, according to the audit team, is an entirely different matter.

The audit group surmised that the wrong-way entries at these exit ramps are being produced in large part by the visual picture that is presented to left-turning crossroad traffic. This visual picture includes a median guardrail that extends nearly entirely to the curb lane. This shuts off a left-turn driver's view of the entrance ramp and distorts the view of the median, which now looks more like a guardrail-aided, right-side curb than a median. Without a clear view of the entrance ramp, the driver's impression is that this exit ramp is the only roadway that is available and may be the only roadway connecting to or from the freeway at that point.

Medians typically do not look like this. Normally a driver can easily spot a median as a median, and that perception has something to do with being able to take in the entire set of lanes in one view, including a pavement-level median at the mouth of the terminal. In reviewing the literature, the Michigan researchers found that Washington State had this same issue and came to the same conclusion regarding concrete barriers that extend all the way to the STOP sign at one of the partial cloverleaf interchanges in that State.

In 2012 MDOT is looking at implementing many of the suggested low-cost treatments at the Gratiot ramps. Included are signings and pavement marking improvements, such as painting the median island and installing pavement marking extensions, installing special arrow markings for lane assignments at the mouths of the exit ramps, lowering the height of the DO NOT ENTER and WRONG WAY signs, adding reflective red sheeting to the sign supports, upgrading the existing I-94 directional signs on the crossroad to a larger size, and installing a raised lane separator system that would prohibit left-turning vehicles on Gratiot Avenue from entering the incorrect ramp. An unmaintained street light already has been removed to allow for greater visibility of the entrance ramp. In the longer term, MDOT expects to remove some portion of the guardrail during planned 2014 bridge work and plans to completely reconfigure the current partial cloverleaf ramps to a diamond configuration during a future reconstruction project.

Last Word

From an engineering point of view, reducing wrong-way entries starts with making the correct ramp choices appear inviting, while making the incorrect choices seem uninviting. That is the direction MDOT is taking.

Greg Johnson, MDOT chief operations officer, is pleased with his agency's efforts but takes a realistic view of the task at hand: "We have about 1.3 billion vehicles per year entering our freeway system in Michigan. Some of these entries are made in darkness and under less-than-perfect weather conditions, and some of the drivers are tired, or impaired, or just not as mentally alert as they could be. We don't expect to stop all wrong-way entries. But this research has helped us to better understand the driver failure modes that we are trying to prevent."

David A. Morena has been the highway safety specialist at FHWA's Michigan Division Office in Lansing since 1983. Past and current safety initiatives to which he has contributed, both in Michigan and nationwide, include rumble strips, elderly mobility countermeasures, traffic signal placement, road diets, wrong-way driving countermeasures, and engineering/emergency medical system collaboration. Morena has a B.S. in industrial engineering and an M.S. in traffic engineering from Ohio State University.

Tracie J. Leix, P.E. is the manager of the safety programs unit at MDOT, where she has been since 2005. She holds a degree in civil engineering from Michigan Technological University.

For more information, contact David Morena at 517-702-1836 or dmorena@dot.gov, or Tracie Leix at 517-373-8950 or LeixT@michigan.gov.