Slower drivers and drivers with emergencies sometimes travel on or near the road shoulder. This practice may cause a problem when there is inadequate notice about shoulder gaps--such as a "shoulder drop," a discontinuation of the paved shoulder across a bridge span.

Recent controlled field tests studied driver behavior when a bridge shoulder drop was marked with an edgeline taper/transverse marking delineation pattern. Drivers encountering this pattern left the shoulder farther from the upcoming bridge than when they only saw object markers or post-mounted delineators at the bridge.

The shoulder-drop delineation study, sponsored by the Texas Department of Transportation (TxDOT) and conducted by the Texas Transportation Institute, evaluated alternative shoulder-drop bridge delineation methods. This study resulted in a report, "Delineation at Bridges with Paved Shoulder Drops," by Gerald L. Ullman and Val J. Pezoldt. The report appeared as Transportation Research Board Paper No. 980721, prepared for the January 1998 meeting of the Transportation Research Board in Washington, D.C.

BACKGROUND

Reviewing previous dropoff studies, the authors found several warning devices suggested for reducing the number of bridge crashes. They did not find specific data about which improvement or combination of improvements was most effective, or under which conditions. In field testing, many site-specific factors affect driving behavior. Moreover, narrow bridges and culverts are often on rural highways, where low vehicle volumes make it difficult for meaningful statistical comparisons.

The researchers used a model to divide the approach to a road hazard into five information-handling zones:

• advance zone,
• approach zone,
• non-recovery zone,
• hazard zone, and
• downstream zone.

Recognizing these zones helped researchers assess the visibility of a hazard and identify where to locate hazard warning information. The diagram below shows these zones.

FIGURE 1
Information Handling Zones Upstream of a Bridge

The middle three zones--approach, non-recovery, and hazard are the most critical for hazard delineation system design. Drivers must change their speed, path, and (or) direction by the time they reach the non-recovery zone--170 to 260 meters from the bridge for speeds of 88 to 113 km/h.

The researchers considered alternative bridge delineation systems at a simulated narrow bridge site. They chose to use a controlled field study approach for several reasons:

• to allow each subject exposure to each treatment;
• to reduce the number of site-specific factors; and
• to efficiently collect more detailed comparative information.

STUDY METHODOLOGY

Researchers created a closed driving course. Test subjects drove at night on dry pavement in a vehicle that logged continuous speed, distance, and lateral position data. Each subject tested each of the delineation treatments.

Because roadway delineations deteriorate over time, the test simulated a worn condition with a mist of paint to produce a "near minimum" acceptable level of visibility. Sections of W-beam guardrail, bolted to the pavement, created a bridge mock-up to represent rural highway conditions.

Study Procedures

Study subjects received a description of their driving task but no information about what the researchers were testing. The test vehicle contained this test equipment:

• a Datron non-contact fifth wheel;
• accelerometers;
• a small automatic-iris video camera;
• a VHS recorder and monitor;
• a portable laptop computer;
• a light at the left front corner of the car; and
• colored lights in a small box mounted in the rear window.

The last two items simulated the appearance of other cars on the road and other distractions. Test drivers practiced driving the vehicle before the test began. As much as possible, they drove the test course at 90 km/h.

Delineation Treatments

The research studied five treatments chosen to test a range of delineation materials, their placement within the visual field, and their location on the approach to the bridge.

• Treatment 1 used Type OM-3 object markers on each side of the road and OM-2 at the end of the W-beam transition section.

• Treatment 2 added three post-mounted delineators (PMDs) behind the W-beam transition section.

• Treatment 3 added a symbolic narrow bridge sign 150 meters ahead of the bridge.

• Treatment 4 was identical to Treatment 2, except for the addition of a "Shoulder Ends-500 ft" sign 150 meters ahead of the bridge.

• Treatment 5 used a 100-mm-wide crosshatching or transverse marking pattern in the shoulder, beginning 150 meters from the bridge.

The diagrams below show these delineation treatments.

FIGURE 2
Schematic Layout and Delineation Treatments Tested in Studies

Subject Characteristics

The study examined two age groups: drivers younger than 25 and drivers 55 or older. Drivers tested each delineation treatment twice. Researchers recognized that having the drivers pass the same bridge site more than once may have had some effect on the driver response data.

Measures-of-Performance

The test vehicle provided the following information, with recording beginning 300 meters from the bridge:

• speed and cumulative distance traveled (at 1/10th second intervals),
• lateral position within the lane or shoulder (measured approximately every 6 meters),
• times when the secondary task occurred (i.e., lights in the rearview mirror were illuminated), and
• lateral and longitudinal accelerations (x- and y-coordinate directions), measured with accelerometers.

Performance measures included (the numbers in brackets refer to Figure 3 below):

• average speed,
• maximum acceleration or deceleration in the longitudinal direction, and the distance from the bridge at which this acceleration occurred,
• maximum acceleration to the left (away from the bridge) in the lateral direction, and the distance from the bridge at which this occurred,
• acceleration noise (standard deviation of accelerations about the mean acceleration) for both the longitudinal and lateral directions,
• the lateral position at the beginning of the bridge [1],
• the variability in the lateral position at the beginning of the bridge,
• the change in lateral position at the bridge relative to the average lateral position [2],
• the distance from the bridge at which the subject began to exit the shoulder [3], and
• the distance traveled while completing the shoulder exit maneuver [4].

FIGURE 3
Measures Taken During Shoulder Drop Delineation Studies

Analysis of these measures used two-factor Analysis-of-Variance (ANOVA) statistical techniques, except for the variability in lateral positions at the beginning of the bridge, which used the Hartley Test of equal variances.

Results

Test results showed that older drivers tended to drive slightly more to the left than younger drivers, regardless of delineation. Older drivers also applied more deceleration force, regardless of the delineation treatment.

Delineation treatment only affected the distance from the bridge at which drivers of all ages began to leave the shoulder. Data clearly showed that drivers left the shoulder farther away from the bridge at Treatment 5 than at Treatments 1 and 2. Treatments 3 and 4 were less definitive. Researchers expected these results.

Test drivers ranked all the treatments as making their driving task "very easy" or "easy" as far as knowing where they were supposed to drive and as making "very clear" or "clear" what they were supposed to do.

The researchers concluded: