All-Weather Pavement Marking for Work Zones: Field Evaluation in North Carolina and Ohio

LATERAL LANE PLACEMENT

Lane placement describes the distance a driver deviates laterally within the travel lane. Driver behavior and vehicle placement of vehicles travelling through the work zone was recorded using two methods of data collection: a video camera mounted in a survey vehicle and distance measuring sensors. The methodology for each data collection method is described in detail in the following sections.

Data Collection

Video Calculated Distances for US-32/33/50

To collect data on the lane placement of vehicles through the work zone for the US-32/33/50 site, a digital video camera was mounted inside a survey vehicle and data were recorded through the advance warning area and through the shift area of the work zone while following a target vehicle in the traffic stream. Due to the remote location of the site and the lack of ambient lighting, video data collection from an elevated location was not feasible because the video could not delineate the pavement markings. Therefore, motorists were followed through the work zone and monitored through the use of a video camera on-site. The motorists were not aware that they were being monitored and, thus, their driving behavior was unbiased.

The video data were then analyzed in the laboratory to obtain quantifiable lateral lane placement data for each observed vehicle. When analyzing the video data, the lateral placement was determined by locating the center of the lane and comparing that location to the location of the center of the vehicle’s license plate in a 6-inch positional alignment. When a vehicle was positioned left of the centerline, or away from the work zone, that was recorded as a positive distance, whereas a vehicle traveling right of the centerline, or closer to the work zone, was recorded as a negative distance. The data were extracted from the video for every second of travel through the shift and immediately before and after the shift.

Distance Measuring Sensors for I-90

To obtain lateral placement distances of vehicles traveling through the I-90 crossovers, sensors were utilized to collect data from the shoulder of the roadway. These sensors were the same used for speed collection.

For the I-90 test site, the lane placement and speed data were collected for the double-lane crossover sections of the work zones for each pavement marking type. For the crossover section with the AWP treatment, four sensors were placed along both of the travel lanes in the middle of the crossover. The sensors were placed along the left travel lane at 100-foot (30.5-m) spacing, with the first sensor placed at approximately the point of reverse curvature in the double-lane crossover. The last sensor on the right travel lane was placed just prior to the jersey barriers dividing the eastbound and westbound lanes. The remaining sensors were then placed at 100-foot (30.5-m) spacing back from the last sensor. For the standard pavement marking treatment, data could not be feasibly collected for the left travel lane, as there was no shoulder available alongside the lane. As a result, all eight sensors were placed along the right travel lane at 100-foot (30.5-m) spacing, with the first sensor located approximately 200 feet (61 m) into the double-lane crossover. Due to the potential in comparing lane placement and speed data for left and right lanes, a statistical analysis was conducted to determine differences, if any, between the two lanes for one direction of travel. The results from the ANOVA indicated that there was no statistical difference between lane placements or speeds between the left and the right lane. Therefore, comparisons can be made between the pavement marking sites.

For each of the work zones studied, the sensors were oriented perpendicular to the roadway and placed on the graded shoulder. The distance from the sensor to the edge line of the closest travel lane was measured and recorded to determine the lane placement of the vehicle. The sensors were placed within 6 feet (1.8 m) of the edge line to remain within the operating range of the sensors. The primary focus was to be able to detect a vehicle within the travel lane adjacent to the shoulder where the sensors were located. Once all of the sensors were in place, the synchronized sampling was conducted using the MicroStrain Node Commander^® software installed on a laptop computer. The synchronized sampling was conducted for the eight sensors at a rate of 128 Hz. The data were collected for approximately 1.5 to 2 hours, depending on the life of the laptop’s battery.

Analysis Methodology

Statistical analyses were conducted for the lateral lane placement data collected at the two Ohio sites. Averages and standard deviations were calculated from the filtered lateral placement data. The data were compared using the one-way analysis of variance testing the significance at a 95 percent level of confidence.

Results

The primary difference between the two test sites was that the distances of lane placements were relative to different reference points, and therefore incomparable. At US-32/33/50, the distances were taken for lane placement relative to the center of the travel lane, whereas at I-90 distances were measured from the shoulder perpendicular to the travel way. The mean speeds, standard deviations, and mean differences were first calculated, followed by a statistical analysis using either t-tests or ANOVA for sample comparison. Table 13 displays a summary table of the lane placement data collected and analyzed for the test sites, followed by a section summarizing the findings from each site studied.

US-32/33/50

The averages and standard deviations were calculated for the samples collected for each of the lane shifts at the US-32/33/50 test site. Based on results from video data extraction, the mean lateral lane placements ranged between 0.043 to 0.771 feet (1.3 to 23.5 cm) from the center of the lane. The positive nature of each of the means indicates that motorists were traveling to the right of the center of the lane, or veering away from the work zone.

Statistical tests were used to determine if the mean lateral lane placement for the test site as compared to the control site was statistically significant. The one-way analysis of variance was utilized to compare the mean speeds. Due to heterogeneous variances, the Welch’s modification to the one-way analysis of variance was utilized, and the calculated F-value was based on an asymptotic distribution. In addition, due to the unequal variances and unequal sample sizes, the Games Howell post hoc test was utilized to examine specific differences within those samples tested in the one-way analysis of variance. Based on the one-way analysis of variance, it was determined that the mean lateral lane placements associated with the vehicles entering, within, and exiting the lane shift for the daytime rain, nighttime dry, and nighttime wet conditions were significantly different, as shown in table 14.

Table 13. Summary of statistics for lateral lane placement.

Site	Light	Weather	Paint	Location	Average Lane Placement (Feet)	Standard Deviation	Mean Difference	P(T<=t) two-tail
US-32/33/50	Day	Rain	Standard	Entry Crossover	0.61	0.46	0.15	0.214
			AWP	Entry Crossover	0.75	0.43
			Standard	Within Work Zone	0.28	0.31	0.15	0.017*
			AWP	Within Work Zone	0.43	0.4
			Standard	Exit Crossover	0.55	0.35	-0.50	<0.001*
			AWP	Exit Crossover	0.04	0.46
	Night	Rain	Standard	Entry Crossover	0.57	0.46	0.09	0.765
			AWP	Entry Crossover	0.66	0.39
			Standard	Within Work Zone	0.29	0.39	0.21	0.001*
			AWP	Within Work Zone	0.50	0.41
			Standard	Exit Crossover	0.57	0.47	-0.33	<0.001*
			AWP	Exit Crossover	0.24	0.54
	Night	Clear	Standard	Entry Crossover	0.52	0.45	0.25	0.003*
			AWP	Entry Crossover	0.77	0.31
			Standard	Within Work Zone	0.30	0.42	0.23	<0.001*
			AWP	Within Work Zone	0.53	0.4
			Standard	Exit Crossover	0.67	0.4	-0.28	0.001*
			AWP	Exit Crossover	0.39	4.42
I-90	Day	Clear	Standard	Exit Crossover	3.25	3.75	2.57	<0.001*
			AWP	Entry Crossover	5.82	2.13
	Night	Clear	Standard	Exit Crossover	9.28	3.03	-3.80	<0.001*
			AWP	Entry Crossover	5.48	3.09
	Night	Rain	Standard	Exit Crossover	6.57	3.01	-0.07	0.994
			AWP	Entry Crossover	6.50

* p-value is statistically significant (α = 0.05)
1 foot = 30.5 cm

Table 14. US-32/33/50 lane placement ANOVA results.

Comparison	Sum of Squares	Degrees of Freedom	Mean Squares	F-Calc	p-value
Daytime Rain Conditions Between Groups Within Groups Total	25.02 88.49 113.51	5 240.03	5.00 0.159	29.148	< 0.001*
Nighttime Dry Conditions Between Groups Within Groups Total	15.12 109.61 124.73	5 257.84	3.025 0.184	20.854	< 0.001*
Nighttime Rain Conditions Between Groups Within Groups Total	12.67 97.00 109.68	5 251.25	2.535 0.174	14.259	< 0.001*

* p-value is statistically significant (α = 0.05)

For US-32/33/50, the ANOVA results provided in tables 13 and 14 indicate that for all conditions:

• For daytime and nighttime rain conditions, the mean lateral lane placements within and exiting the lane shifts were statistically different between the standard pavement marking and the AWP, while means for entering the lane shift were statistically similar. While traveling through the work zone, vehicles tended to maintain a position closer to the white markings when utilizing the AWP. When exiting the work zone, vehicles maintained a position closer to the center of the lane delineated with the AWP.

• For the nighttime dry conditions, the mean lateral lane placements for all locations along the lane shifts were statistically different between the standard pavement marking and the AWP sections. When entering the lane shift and through the lane shift, those vehicles traveling through the AWP work zone maintained a position closer to the white pavement markings as compared to the standard pavement marking work zone. When exiting the work zone, vehicles traveling through the AWP work zone were able to maintain a position closer to the center of the lane than those traveling through the standard pavement marking work zone.

The US-32/33/50 results are visually represented in figure 26.

I-90

The averages and standard deviations were calculated for the samples collected for each of the two-lane crossovers at the I-90 test site. The sensor was placed 3 feet (.91 m) away from the outside edge of the lane line. Based on results from sensor data extraction, the mean lateral lane placements (em.e., the distance between the sensor and the vehicle’s closest tire) ranged between 3.25 and 9.28 feet (1 and 2.8 m). Given a lane width of 12 feet (3.7 m) and an assumed vehicle width of 7 feet (2.1 m), a vehicle positioned directly in the center of the lane would be 2.5 feet (.76 m) inside the edge lane. Therefore, a vehicle position (the edge of the vehicle) that was approximately 5.5 feet (3 + 2.5 feet), or 1.7 meters, from the sensor is assumed to be located within the center of the closest lane in the two-lane crossover. A vehicle detected 9.28 feet (2.8 m) from the sensor is assumed to be encroaching a total of 1.28 feet (0.39 m) in the second lane as follows: the near tire is 6.28 feet (1.91 m) from the near edge line (9.28 feet – 3 feet set back for the sensor), with the far tire positioned 13.28 feet (4.05 m) from the near edge line (6.28 feet + 7-foot width). Based on a 12-foot (3.7-m) lane width, the far tire is located 1.28 feet (0.39 m) inside the adjacent lane.

Table 15 shows a summary of lane placements determined for I-90. Based on the one-way analysis of variance, it was determined that the mean lateral lane placements associated with the vehicles traveling through the two-lane crossover for the daytime dry, nighttime dry, and nighttime wet conditions were significantly different, as shown in table 15.

Table 15. I-90 lane placement ANOVA results.

Comparison	Sum of Squares	Degrees of Freedom	Mean Squares	F-Calc	p-value
All Conditions Between Groups Within Groups Total	22505.34 176294.20 198799.54	6 3855.34	3750.89 12.56	341.09	< 0.001*

* p-value is statistically significant (α = 0.05)

For I-90, the ANOVA results indicate the following:

• During daytime dry conditions, the mean lateral lane placements of vehicles on the standard pavement marking and the AWP sections were statistically different. The vehicles traveling through the AWP crossover generally maintained a position in the center of the travel lane more consistently than those traveling along the standard pavement marking crossover. The edges of vehicles traveling along the standard pavement marking crossover were generally located just 1 foot (.3 m) from the white lane line.

• For the nighttime dry conditions, the post hoc analysis also indicated that the vehicles traveling through the AWP crossover were able to maintain a position in the center of the lane more consistently than those traveling along the standard pavement marking crossover. The edges of vehicles traveling along the standard pavement marking crossover were generally located closer to the lane line dividing the two lanes, thereby creating a sideswipe crash potential.

• The post hoc conditions indicated that, during the nighttime rain conditions, the lateral lane placement of vehicles were statistically similar.

The I-90 results in table 15 are visually represented in figure 27.

Note: The dashed line represents the distance from the sensor to the edge of the outer lane line. The lane placement measurements are the linear distance between the sensor and the vehicles closest tire.

Conclusions

The lateral lane placement of vehicles was quantified to assess the ability of the pavement markings in guiding motorists through the work zone. Vehicle placement within a work zone was considered a primary measure for safety due to the ability to discern a vehicle’s crash risk. The crash risk was identified by the location of a vehicle close to a lane line, indicating the potential for a sideswipe crash or intrusion into the work zone, increasing the potential for a fixed object crash or a pedestrian crash.

Along the US-32/33/50 site, motorists maintained their position either in the center of the lane or close to the white edge marking. Statistical differences were noted for the daytime and nighttime rain conditions within and exiting the lane shift and for all the locations for the nighttime dry condition. Generally, when exiting the lane shift, those vehicles traveling through the AWP site were able to maintain a position closer to the center of the lane than those traveling through the standard pavement marking site.

For the I-90 site, motorists also maintained their position in the center of the lane when traveling through the AWP site for the daytime dry and nighttime dry conditions. During the nighttime rain conditions, motorists maintained their position nearly in the center of the lane for both the AWP and the standard pavement marking sites. While traveling through the standard pavement marking site during the nighttime dry conditions, the mean lateral lane placement position of the vehicles indicated a higher potential for sideswipe crashes.

It was preferable for motorists to be able to locate their vehicle in the center of the lane to minimize the crash potential. However, based on the higher levels of retroreflectivity associated with the white pavement markings, it was not surprising that drivers would utilize the white edge line to appropriately position their vehicles. At the I-90 site, the examined lane for lateral placement utilized white pavement markings along both sides of the lane, which enabled drivers to maintain their lane appropriately more frequently.

Overall Conclusions

This research effort sought to quantify the effects of a new pavement marking product developed by 3M for work zone deployment under nighttime rainy conditions. The AWP was previously tested under rain simulated conditions, and findings indicated that the marking was much more visible than its conventional counterpart, primarily due to supplemental optical elements that retroreflect light in rain conditions. This research study builds on these findings by employing studies at actual work zones. Four MOEs were utilized in studying five work zones in North Carolina and Ohio. These MOEs included retroreflectivity, speed, lane encroachment, and lateral lane placement.

Retroreflectivity measurements were taken to determine 1) differences in AWP applications across sites and 2) differences between AWP and standard pavement marking at each individual site. From the data, it can be concluded that AWP applications had statistically significant differences between the AWP and standard pavement marking; however, there was not consistency among the pavement marking contractors in the application of either pavement marking type. This suggests that pavement marking contractors likely have different application methods with very little consistency in application rates and element drop rates. The research team suggests that future studies look at the retroreflectivity values over time to see the degradation rates of individual pavement markings under actual field conditions.

Speed was utilized to supplement findings regarding lateral lane placement. The research team used speed as a supplemental finding because higher or lower speeds through work zones proved not to be a good indicator of improved safety. For instance, increased visibility may result in increased speed if there is more driver comfort. It is not known what affect this may have on driver safety. Nevertheless, analyses of speeds were instructive and interesting. Speeds should be interpreted with caution due to their variability and dependence on many factors, primarily the geometry of the crossovers being compared, rain intensity during different analysis periods, method of extraction, and overall light conditions. The research team took every possible measure to eliminate as much variability as possible when collecting data and making comparisons.

Speeds were analyzed at all five work zone sites. The overall findings indicated the following:

• Statistically significant differences in speeds at entry and exit portions of a single crossover. This indicates that drivers tend to decelerate into each crossover and speed up exiting the same crossover, regardless of the marking type used.

• Statistically significant differences in speeds at entry and exit curves (entering and exiting the entire work zone – two crossovers). The findings showed that drivers drive 3.8 to 5.9 miles per hour (6.1 to 9.5 kph) faster when exiting the work zone.

• Although inconclusive, it appears that speeds were more likely to be higher in similar entry or exit curves when AWP was used than when standard pavement marking was used. This was the case in three of four different sets of curves that were compared, and all were statistically significant except one. For all practical purposes, even this one sample was significant (p-value =0.057).

Lane encroachment and lateral lane placement were the primary MOEs used to determine if safety had improved using the new AWP, as these measures are easily correlated with driver’s lane-keeping ability. It was hypothesized that brighter lane markings would enhance the driver’s ability to distinguish lanes on the roadway and make safer driving maneuvers. A total of four sites were studied for lane-keeping, two each for lane encroachments and lateral lane placement.

Lane encroachments were defined as vehicles crossing the lane line in a crossover. Comparing both pavement marking types in similar curves provided contrasting findings. The first site studied indicated that a higher number of lane encroachments occurred at standard pavement marking curves; however, that study introduced several factors that could not be accounted for directly (such as rain intensity). The second study, which was more robust, found statistically insignificant findings that standard markings resulted in more lane encroachments than AWP.

An interesting finding did emerge at these two sites when comparing entry and exit curves for both marking types—the exit lane was found to have more lane encroachments. This correlates to earlier findings that speeds were higher at exit curves (3.5 miles per hour [5.6 kph] or greater). This finding indicates that more attention should likely be given to the exit curve of a work zone since higher lane encroachments and speeds likely lead to higher potential for a serious collision.

Lateral lane placement describes a vehicle’s proximity to a lane line, indicating the potential for a conflict or crash. Two sites were studied, one comparing an entry, within, and exit portions of two parallel work zones and the other comparing the entry and exit crossover for one direction of travel. At the first site, statistically significant differences were noted for daytime and nighttime rainy conditions within and exiting the lane shift; drivers maintained their lateral placement better within a work zone marked with standard markings but maintained their lateral placement better when exiting a work zone marked with AWP. During nighttime clear conditions, drivers maintained their lateral placement better throughout work zones marked with standard markings, while a statistically significant improvement in lateral lane placement was found with AWP when drivers were exiting the work zone.

At the second site, results also varied. During daytime clear conditions, drivers in the standard pavement marking sections maintained their lateral placement better. In contrast, during nighttime clear conditions, they maintained their lateral placement better in the sections marked with AWP. More importantly, nighttime rain conditions showed no statistically significant difference between the pavement markings.

In short, findings for lateral lane placement were similar to those regarding lane encroachments. Findings varied among sites and locations within the work zone; therefore, the findings were inconclusive.