July/August 1997· Vol. 61· No. 1

That’s very cool.
That’s made it much more visible.
Now it’s different; these are far superior to what we had.
I think that shows up much better than anything I’ve ever seen before.
Everything this light touches is brighter.

These are representative, direct quotes of participants in a test to determine the effectiveness of using ultraviolet headlights used in conjunction with fluorescent pavement markings for roadway delineation. The two-part study measured distance and general visibility of the fluorescent pavement markings as compared to standard roadway markings. It is notable that there were no negative comments from any participant.

This study was conducted using a 1993 Volvo equipped with add-on ultraviolet headlights.

Background

Ultraviolet (UV) light is not visible to the human eye. However, when ultraviolet light strikes certain materials, the wavelengths of the ultraviolet light become longer, creating light that is readily visible. This phenomenon, known as fluorescence, makes objects more visible and, therefore, offers great potential for improving safety. Combining ultraviolet headlights on vehicles and ultraviolet-activated fluorescent materials in roadway markings is a promising approach to improving drivers’ nighttime vision. Implementing this innovative treatment will involve collaboration between industry and government.

Visibility plays a major role in nighttime accidents and fatalities.⁽¹⁾ Unfortunately, low-beam headlights do not provide adequate visibility under many different driving conditions, and problems were found with brightness, contrast detection, glare, and the sharpness of central vs. peripheral vision.⁽²⁾

A view of the static test area

European research, especially that conducted by the Swedish National Road Administration, has examined the effect of UV headlights on highway visibility. UV headlights - with emission spectra between 320 nanometers and 400 nanometers - and fluorescent road markings were used to create a full-light effect in conditions of poor visibility without blinding the oncoming traffic. Pedestrians could be seen much more clearly, and the path of the roadway could be seen far beyond oncoming vehicles.⁽³⁾

Method

The field trials were conducted in dry weather during October and November 1995 between 7 p.m. and 11 p.m. on a section of the Clara Barton Parkway in Montgomery County, Maryland, near Washington, D.C. The Clara Barton Parkway is a four-lane, divided highway with a 6- to 9-meter grass median. The roadway has low concrete curbs and no shoulders. There are a few post-mounted reflectors. The posted speed limit is 50 miles per hour (80 kilometers per hour).

A 1993 Volvo, series 960, was equipped with three rectangular Ultralux UV lamps installed in front of the grill and on top of the front bumper. The UV lamps, when activated by a toggle switch in the cabin of the Volvo, were always used in addition to standard Society of Automotive Engineers (SAE) low-beam headlights.

Dynamic Test Site

Six 90-meter segments of roadway were selected for testing. Three different roadway marking materials were used for the tests: worn and faded standard white paint, new thermoplastic, and recently installed thermoplastic containing fluorescent material. Two relatively straight segments of roadway were chosen for each type of marking material.

Static Test Site

A work zone right-lane closure was set up each night using advance warning signs, a changeable message sign (CMS), an arrow board, and orange and white retroreflective barrels and cones. The static test took place within the closed lane using a relatively straight segment of roadway where fluorescent-paint roadway markings were installed. This site was within, but not part of, the dynamic test course.

Markers were placed in the lane closure so the test vehicle could be consistently positioned within the lane during the stationary portion of the experiment. All post-mounted reflectors in the vicinity were covered.

Four traffic cones with reflective collars were placed on the grass to the right of the roadway. They were spaced at 30-meter intervals beginning at 107 meters and ending at 198 meters from the parked Volvo. They were used as points of reference for the subjects when the static tests were conducted.

Test Subjects

Subjects were recruited from the Turner-Fairbank Highway Research Center (TFHRC) subject bank. Only licensed drivers who drove at night, at least occasionally, took part in the study. Two age groups were used in the experiment: drivers age 25 to 45 and drivers age 65 and older.

A total of 41 subjects were tested, including five for pilot testing. Data for 36 subjects were analyzed. The final subject pool included seven older males, eight older females, 13 younger males, and eight younger females.

Procedure

At TFHRC, subjects were given a brief visual acuity test. A minimum acuity of 20/40 was required. Subjects were also given an overview of the study and preliminary instructions.

Each subject, in turn, was escorted to the study vehicle; seated in the right-front passenger seat; and given a clipboard, data form, pen, and penlight flashlight. During the ride to the test location, the driver/controller again explained the focus of the study. After arriving at the test site, the driver read the instructions for the dynamic test to the subject and asked the subject to record his opinion of the roadway markings on the data form by using the pen and penlight provided. After answering any questions that the subject asked, the driver took the subject once through the entire course as a practice run.

Subjects used a five-point subjective scale to rate how well the markings indicated where to drive the vehicle, with a A1" meaning poor and a A5" meaning excellent. After exposure to each site, subjects circled the number that represented their opinion of the markings in comparison with all other roadway markings they had seen in the past.

Three orders of presentation were used so that one-third of the subjects were exposed to the new thermoplastic sites first, one-third to the worn paint sites first, and one-third to the UV sites first.

Each subject was driven through the test course three times (three loops). The practice loop was always conducted with only the low-beam headlights on. At the beginning of the second and third loop, the Volvo headlights were set for either low beam only or low beam plus UV lamps, and the headlights remained on that setting for the entire loop. The order of headlight use was also counterbalanced.

For each test exposure, the number of vehicles (in front of, behind, passing, and across the median oncoming) in the vicinity of the study vehicle during the exposure period was recorded by a second team member, who rode in the back seat. This individual also recorded a subjective distraction rating of these vehicles. Distractions included glare, other vehicle interference, and/or vehicles in the exposure area. The distraction rating was based on a five-point subjective scale, on which a A1" indicated no distraction and a A5" indicated very distracting conditions. Since it was not possible to control the other vehicles using the roadway, it was suspected that the presence of other vehicles might affect subject performance.

At the conclusion of the dynamic test, subjects were driven to the lane closure, and the car was parked in a predetermined location.

The static test consisted of three activities conducted first with low-beam illumination and then again with UV-headlight illumination. First, subjects were asked to use the same five-point scale, as used in the dynamic test, to rate the overall visibility of the roadway markings at the static test site. Next, subjects were asked to count the number of skip or dashed lane lines they could see. That number was recorded by the observer in the back seat. Finally, subjects were asked to specify the point at which the Volvo headlights no longer illuminated the right edge line. Traffic cones were placed on the shoulder as markers to help subjects respond to the question, and subjects could say that the headlights ended at the first, second, third, or fourth traffic cones or anywhere in between. Their response was recorded. The UV lamps were turned on, and the three questions were repeated.

Results

Dynamic Testing
Results of the dynamic testing are shown in table 1.

The UV-activated fluorescent markings received the highest mean rating when the UV headlights were on (4.40). When the UV headlights were off, the rating was only 3.46. Since the new thermoplastic and worn paint did not contain any fluorescent material, the UV headlights would not be expected to increase the visibility of these markings. This is shown in table 1 as the ratings for the new thermoplastic are 3.92 with UV on and 3.89 with UV off, and the ratings for the worn paint are 2.50 with UV on and 2.51 with UV off.

The distributions of the subjective ratings for each of the three pavement types with the UV lights on and with the UV lights off are shown in figure 1. The fluorescent markings with the UV lights on received the highest possible rating from 40 percent of the subjects. The same markings with the UV lights off received a similar rating from only 15 percent of the subjects.

The expected lack of effect of the UV lights on both the thermoplastic and the worn paint is also shown in figure 1. The relative effectiveness of the UV-activated fluorescent marking is indicated by the fact that 44 percent of the subjects rated the UV-activated fluorescent marking as a "5", while less than 10 percent gave the worn paint the same superior rating.

As described previously, a second researcher rode in the back seat of the test vehicle and recorded information about the degree of headlight glare from oncoming and following vehicles while the subject was evaluating the test segments. Since testing was conducted after the evening rush hour and traffic was relatively light, it is not surprising that glare conditions during most of the testing were rated as low. Attempts to identify a relationship between the degree of glare and the effectiveness of UV-activated delineations were unsuccessful due to the small number of test sessions that involved high-glare conditions. The performance of the UV-activated fluorescent materials in high-glare conditions will be addressed in future research.

Static Testing
During the static testing, subjects were asked to indicate how many center skip lines they could see and how far they could see the right-lane edge line. The testing location had UV-activated fluorescent pavement markings. Data for test sessions with the UV headlights on were compared with data for sessions with the UV headlights off. In addition, the subjects were asked to provide a subjective evaluation of the marking effectiveness, as was done in the dynamic testing.

Table 2 shows the means and standard deviations for each of the three measures. With the low beams, the subjects could see an average of 7.6 center skip lines. When the UV lamps were turned on, the mean value increased to 9.8. The amount of right edge line visible with the low beams only was 144 meters. This increased to 180 meters with the UV lamps. The mean subjective rating of visibility with the low beams only was 3.2. The UV lamps increased this value to 4.7, out of a possible 5.

The frequency distributions of the subject responses are shown in table 3. These data are graphically depicted in figure 2.

The superiority of the UV-activated fluorescent treatment is apparent. Without the UV lamps, 47 percent of the subjects could see seven or fewer skip lines. With the UV lamps, all of the subjects could see at least eight center skip lines. With the UV lamps, 58 percent of the subjects could see at least 200 meters of right edge line. Without the UV, only 8 percent of the subjects could see that far.

The subjective rating data are even more dramatic. While 75 percent of the subjects gave an excellent (A5") rating to the UV-activated treatments, only 8 percent of the subjects gave the delineation an excellent rating without the UV lamps. In statistical terms, the number of skip lines, the length of visible edge line, and the subjective ratings were significantly greater with the UV lamp than without it, using the Kolmogorov-Smirnov test at the 0.05 level of significance.

Conclusions

The UV headlights provided a very noticeable increase in delineation visibility. In dynamic testing, the mean subjective rating of the roadway delineation with the UV headlights was 19 percent higher than with regular low beams. In the static testing, with the UV headlights and UV-activated pavement markings, subjects were able to see an average of 25 percent farther along the edge line and 29 percent more of the center skip lines. In the static testing, the subjective rating of visibility increased 47 percent with the UV headlights.

The results of this preliminary testing are very encouraging. Additional testing of the UV headlights is currently underway. This testing involves determining the effect of UV headlights on the visibility of roadway delineation, post-mounted delineators, and pedestrians. The testing will take place on a closed testing track using an instrumented test vehicle.

References

1. R.C. Vanstrum and J.C. Landen. The Dark Side of Driving, Transportation Quarterly, Vol. 38 No. 4, Eno Foundation for Transportation, Westport, Conn., pp. 491-505, 1984.
2. P.L. Olson, M. Sivak, and D.L. Henson. Headlamps and Visibility Limitations in Nighttime Traffic, Journal of Traffic Safety Education, Vol. 27 No. 4, California Association for Safety Education, Anaheim, Calif., pp. 20-33, 1981.
3. C. Barrie. Invisible Light Helps Drivers to See in the Dark, Engineer, Vol. 268 No. 6938, Morgan-Grampian Publishers, Ltd., London, U.K., p. 45, 1989.
4. P.T. Fast. Field Test of UV Headlamps at a 100-km Test Site: Towards an Intelligent Transport System, Proceedings of the Fifth World Congress on Applications of Transport Telematics and Intelligent Vehicle-Highway Systems, Vol. 4, Paris, France, pp. 2172-2179.

Karen R. Mahach is a member of the research faculty staff of George Mason University. As a contractor in support of the human factors research program at TFHRC, she was directly involved with the study reported in this article. She received her bachelor’s degree in psychology from the University of Virginia. She received a master’s and a doctorate in human factors engineering from George Mason University.

Richard L. Knoblauch has been the director of the Center for Applied Research Inc. since 1980. He was the principal investigator on the research study upon which this article is based.

Carole J. Simmons was an engineering psychologist in the Traffic and Driver Information Systems Division of the Federal Highway Administration (FHWA) at the time this research was conducted. She was the manager of FHWA’s UV/Fluorescent Research Program. She is currently with the Human Factors Research Program of Public Policy Center at the University of Iowa. She received her doctorate in psychology from the University of Michigan.

Marsha Nitzburg is a project manager for the Center for Applied Research Inc. Since 1983, she has been active in transportation safety research with an emphasis in the human factors aspects of pedestrian and highway safety. She was the field manager for the study discussed in this article.

John B. Arens is the manager of the Photometric and Visibility Laboratory at TFHRC. He has been involved in the UV/Fluorescent Research Program for several years. He obtained the UV headlights and the delineation materials from Sweden. He holds a degree in electrical engineering from Cleveland State University. He is an active member of the Illuminating Engineering Society, serving on the Roadway Lighting Committee and on the Testing Procedures Committee, and of the International Commission on Illumination, Divisions 2 and 4. From 1992 to 1995, he was chairman of the Visibility Committee of the Transportation Research Board.

Samuel C. Tignor is the chief of the Traffic and Driver Information Systems Division of FHWA’s Office of Safety and Traffic Operations Research and Development. The research reported in this article was initiated in his division. He has been with FHWA since he received his bachelor’s degree in civil engineering from Virginia Polytechnic Institute and State University in 1958. He received his master’s and a doctorate degrees in civil engineering from the University of Michigan.

TFHRC Home | FHWA Home | Feedback United States Department of Transportation - Federal Highway Administration