U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590
202-366-4000


Skip to content
Facebook iconYouTube iconTwitter iconFlickr iconLinkedInInstagram

Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations

Report
This report is an archived publication and may contain dated technical, contact, and link information
Publication Number: FHWA-HRT-04-140
Date: December 2005

Enhanced Night Visibility Series, Volume IX: Phase II—Characterization of Experimental Objects

PDF Version (716 KB)

PDF files can be viewed with the Acrobat® Reader®

CHAPTER 3—RESULTS

The results from this measurement activity are presented below. The measured values of headlamp illuminance, object luminance, and background luminance are presented followed by the calculated metrics of reflectance, fluorescence, contrast, and VL.

ILLUMINANCE FROM THE HEADLAMPS

The illuminance from the headlamps was measured at points corresponding to the luminance measurement points. For the analysis, an analysis of variance (ANOVA) that considered the effect of the UV–A systems on the illuminance was conducted. Then the data were analyzed in terms of the pedestrian objects (which included the cyclist), the child’s bicycle, and the tire tread.

UV–A Illuminance

Because UV–A is outside of the human visual range, the UV–A sources should not contribute to the illuminance on the object; however, the UV–A systems used, particularly the hybrid source, did have some output in the visible range. Thus a comparison of the measurement results with and without the UV–A systems was used to investigate their effect on the illuminance. The object position was considered as a factor because the illuminance contribution from the UV–A sources may have varied across the roadway. That is to say, the parallel and static objects on the shoulder of the road might have received a different level of illuminance from the UV–A sources than did the objects in the center of the road such as the cyclist and the perpendicular objects. The results of this comparison are shown in figure 18.

Bar graph. Object illuminance by UV–A level. Click here for more detail.
Figure 18. Bar graph. Object illuminance by UV–A level.

This interaction of position and UV–A shows that there is a slight increase in illuminance on all the objects, with the greatest increase on the parallel object at the side of the road; however, the increase remains within the standard error of the measurements made without UV–A sources. Because neither effect was significant, the UV–A sources were not considered in the analysis of the headlamp illuminance.

Pedestrian and Cyclist Objects

The parallel pedestrian, perpendicular pedestrian, and cyclist were considered together in this group. The results for the black-clothed objects and the white-clothed objects were averaged because the illuminance on an object is independent of the object itself. The results were then summarized by measurement height, position of the object on the roadway, and the distance from the VES to the object.

Figure 19 shows illuminance results for the different VESs and the measurement height. It is interesting that with the exception of the HHB light source, the illuminance decreased as the measurement height increased, meaning that the headlamps primarily lit the lower portions of objects, which were closer to the roadway. This is to be expected because headlamps are designed to limit the projected light above the horizon that could cause glare for oncoming drivers. The HHB light source illuminated the higher portions of the object better than they did the lower portions. This is also to be expected because the HHB typically is aimed to provide illumination farther down the roadway.

Bar graph. Illuminance from each VES by measurement height for the pedestrian object types. Click here for more detail.
Figure 19. Bar graph. Illuminance from each VES by measurement height for the pedestrian object types.

The next aspect of the illuminance results is the performance for the three object positions on the roadway, shown in figure 20. Through review of the standard error, this relationship shows that the HLB and the
HLB–LP similarly illuminated all three objects. The noteworthy result is that the HID and the HOH headlamps lit the pedestrian at the side of the road (parallel pedestrians) more than they lit those in the center of the road (cyclists and perpendicular pedestrians). Conversely, the HHB headlamps lit objects in the center of the road more than it did the objects on the side of the road.

Bar graph. Illuminance from each VES by object position for the pedestrian object types. Click here for more detail.
Figure 20. Bar graph. Illuminance from each VES by object position for the pedestrian object types.

The other aspect of interest for the illuminance is the relationship of the illuminance to the distance between the object and the vehicle as shown in figure 21. This relationship should follow the inverse square law, which is shown as the line on the figure. The difference between the HID and HLB headlamp types as well as the relationship between the chest and the waist measurement appears to be consistent, and the difference followed the inverse square trend.

Bar graph. Illuminance and distance relationship by lamp and measurement location with the inverse square law trend for the pedestrian object types. Click here for more detail.
Figure 21. Bar graph. Illuminance and distance relationship by lamp and measurement
location with the inverse square law trend for the pedestrian object types.

Child’s Bicycle and Tire Tread Objects

The results for the tire tread and the child’s bicycle are shown in figure 22. Both of these objects were at the side of the road and close to the road surface, and as a result, the illuminance on these objects followed the relationship of the ankle measurement for the pedestrian objects as seen in figure 19. Generally, the illuminance on the two object types was the same for all of the headlamps except HLB and HID. It is not clear what caused the difference between the tire tread and the child’s bicycle for the HLB and the HID light sources.

Bar graph. Illuminance on the child’s bicycle and the tire tread for each VES type. Click here for more detail.
Figure 22. Bar graph. Illuminance on the child’s bicycle and the tire tread for each VES type.

OBJECT LUMINANCE

For analysis, the object luminance was broken into three categories. As with the illuminance data, the pedestrians and cyclists, the child’s bicycle, and the tire tread were all analyzed separately. The pedestrians and cyclists were further broken into white- and black-clothed groupings for analysis.

The measurement method for both the object luminance and the background luminance was designed such that a first measurement series was performed with 61.0 m (200 ft) between the experimental object and the vehicle to establish the object luminance. The next series of measurements was performed at 91.5 m (300 ft), 152.5 m (500 ft), and 244 m (800 ft) to establish the effect of distance on the luminance.

Pedestrian and Cyclist Objects

The overall effect of VES types on the object luminance is shown in figure 23 for the black- and the white-clothed pedestrians and cyclist. In this case the HID, HOH, and HLB–LP provided a lower level of luminance than the HHB, which had the highest value, and HLB, which had an intermediate level. Figure 23 also shows the effect of the UV–A source on the luminance of the white-clothed objects; as the UV–A level increased from none to five UV–A, the luminance of the objects also increased. This is evident for both the UV–A + HID and UV–A + HLB combinations; however, the most obvious significant difference is the effect on object luminance of the white and black clothing, which was analyzed separately to review this difference in more detail. Results follow the figure.

Bar graph. Object luminance by VES for white- and black-clothed pedestrians. Click here for more detail.
Figure 23. Bar graph. Object luminance by VES for white- and black-clothed pedestrians.

Black-Clothed Objects

The black-clothed objects included the cyclist and the perpendicular and parallel pedestrians. Figure 24 shows the object luminance for VES for each of these object classes.

Bar graph. Object luminance by VES for black-clothed pedestrians and cyclist by object position. Click here for more detail.
Figure 24. Bar graph. Object luminance by VES for black-clothed pedestrians
and cyclist by object position.

All of these measurements were taken from a distance of 61.0 m (200 ft). With the exception of the HID class of headlamps, the object luminances were not statistically different between the black-clothed objects. For the HID class of headlamps, the parallel position had a consistently higher luminance than the perpendicular position and the cyclist. This probably indicates that the beam distribution of the HID headlamps is wider than the other headlamp types. A wider beam will illuminate the object at the side of the road to a greater extent than narrower beams. The contribution of the UV–A to the luminance of the black-clothed objects appears to have been minimal. There was slight contribution of the three and five UV–A designs with the HLB headlamp, but there was no contribution from the hybrid UV–A with HLB or any of the UV–A headlamps with the HID headlamp. Finally, the HHB provided the highest values for object luminance; the values for the perpendicular pedestrian and cyclist were slightly greater than the luminance for the parallel pedestrian, which also is indicative of beam pattern. The HHB headlamps are designed to cast more light straight down the roadway rather than to the side. This design would illuminate the objects at the center of the road more clearly than it would those on the shoulder. The HOH shows the opposite characteristics; it provided more light to the side than the center. In terms of the non-UV–A systems, the HOH and the
HLB–LP provided equivalent lighting levels, with the HID series being the lowest, the HHB being the highest, and the HLB being between these values.

The influence of the measurement height on the objects is shown in figure 25.

Bar graph. Object luminance by VES for black-clothed perpendicular pedestrians by measurement height. Click here for more detail.
Figure 25. Bar graph. Object luminance by VES for black-clothed perpendicular pedestrians
by measurement height.

For most VESs, the luminance at the chest measurement height was lower than the luminance at the other measurement heights. The exceptions to this trend were the HLB and the hybrid UV–A + HLB, which showed no significant differences between the measurement heights, and HHB, which showed greater luminance at the chest and waist than at the knee and ankle. These results are to be expected because the HLB and HHB headlamps do not have a significant intensity cutoff; however, the similarity in object luminance with height for HLB and the HHB VES types might have been exacerbated by aiming issues associated with these headlamps. For more information about this issue, please refer to ENV Volume XVII, Characterization of Experimental Vision Enhancement Systems. The UV–A headlamps showed no significant effect on the black clothing and the measurement height. There were also no significant differences among the other VES types.

White-Clothed Objects

The luminance of all the white-clothed objects was first measured at 61.0 m (200 ft), and then the luminance of the parallel and perpendicular pedestrians was measured at 91.5 m (300 ft), 152.5 m (500 ft), and
244 m (800 ft) to establish the influence of distance on the luminance. The analysis of these white-clothed objects was first performed at 61.0 m (200 ft), and then the effect of distance was considered.

As with the black-clothed object, the white-clothed object appeared in perpendicular, parallel, and cyclist configurations. As mentioned, in addition to these, the white-clothed pedestrian was also presented in a static configuration in which the pedestrian stood on the shoulder of the road, approximately 0.3 m (1 ft) from the white roadway edgeline, and faced the approaching participant vehicle. For characterization purposes, this object was considered to be the same as the parallel object, so the static condition was not explicitly considered in any of the analyses. The influence of the object position on the measurements is shown in figure 26. In this relationship, the influence of the UV–A sources on the luminance of the white-clothed objects is evident. The luminance for the five UV–A + HLB VES was the highest, followed by the
three UV–A + HLB VES. The hybrid UV–A also indicated a slight increase in luminance over the HLB with no UV–A condition. The HHB and the three UV–A + HLB VESs were not statistically different. The lowest-performing VESs were the HOH and the HLB–LP, which had similar luminance levels. One of the more interesting results is the luminance of the parallel pedestrian under the HID-based VESs. In the HID configuration with no UV–A contribution, the parallel pedestrian had the highest luminance, which is indicative of the beam pattern of the HID luminaires. However, as the UV–A contribution is added, the influence of the UV–A on the cyclist and the perpendicular pedestrian was greater than on the parallel pedestrian, which had the lowest luminance of the three configurations at the five UV–A level.

Bar graph. Object luminance by VES for white-clothed pedestrians by object position. Click here for more detail.
Figure 26. Bar graph. Object luminance by VES for white-clothed pedestrians by object position.

The influence of the measurement height on the object luminance is shown in figure 27. This influence was very similar to that on the black-clothed objects. The headlamps that have a sharp cutoff, such as the HID headlamps, had a much lower luminance on the chest of the object than the other locations. The HHB headlamp showed a higher luminance on the chest than the ankle because of its higher aiming point.

Bar graph. Object luminance by VES for white-clothed perpendicular pedestrian by measurement height. Click here for more detail.
Figure 27. Bar graph. Object luminance by VES for white-clothed perpendicular pedestrian by measurement height.

The effect of the measurement distance on the object luminance is shown in figure 28. As expected, the luminance decreased as the distance increased. Because the headlamps were the only source of illumination, this would follow the same relationship as the object illuminance and the inverse square law. The change in the luminance for the VESs with the UV–A contribution also showed a decrement with distance. For the HLB configurations, the luminance levels appeared to become statistically similar at 152.5 m (500 ft). For the HID configurations, the luminance appeared to become statistically similar at 91.5 m (300 ft).

Bar graph. Object luminance by VES for white-clothed pedestrians by measurement distance. Click here for more detail.
Figure 28. Bar graph. Object luminance by VES for white-clothed pedestrians by measurement distance.

Child’s Bicycle

The luminance of the child’s bicycle was measured at 61.0 m (200 ft). The results are presented in figure 29. This relationship shows that the UV–A configurations provided a higher object luminance than their baseline counterparts; however, unlike the pedestrian and cyclist objects, the HID configurations outperformed the HLB–LP. The HLB and the HID performed equivalently, and the HOH performed slightly higher. Because the child’s bicycle was placed at the edge of the road and close to the road surface and because the HID has a wider beam than the HHB, this would result in a higher object luminance, as was found.

Bar graph. Object luminance by VES for the child's bicycle. Click here for more detail.
Figure 29. Bar graph. Object luminance by VES for the child’s bicycle.

Tire Tread

The tire tread object was measured at 61.0 m (200 ft), and the results are shown in figure 30. Similar to the child’s bicycle, the HLB, HHB, and HID conditions were all similar, with the HLB–LP having a lower luminance and the HOH having a higher luminance. Again, this appears to result from the nature of the beam pattern for these headlamp types. The HID beam is wider and would illuminate the tire tread at the side of the road better than the beam profile of the HLB–LP. There was no significant contribution to the luminance of the tire tread by the UV–A sources except for the three and five UV–A configurations with HLB, which showed a significantly higher luminance level. This contribution of the UV–A to the tire tread luminance was not expected.

Bar graph. Object luminance by VES for the tire tread. Click here for more detail.
Figure 30. Bar graph. Object luminance by VES for the tire tread.

Cyclist Bicycles

The bicycle used by the black-clothed cyclist was a dark color, while the bicycle used by the white-clothed cyclist was painted with light-colored fluorescent paint. The wheel rims of the two bicycles also differed. The dark bicycle had shiny chrome rims, and the fluorescent bicycle had black rims. The measurement points were the brightest locations, so on the dark bicycle, the chrome rims were used for the measurement. The luminance relationship for these objects is shown in figure 31. The luminance performance was similar to the other objects, with the HHB providing a high level of luminance as compared to the other non-UV–A configurations. The important issue is the overall luminance level as compared to the cyclist alone. For the black-clothed cyclist, the luminance level ranged from 0.2 to 1.2 cd/m², and the bicycle for the same object ranged from approximately 1 to close to 6 cd/m². For the white-clothed cyclist, the luminance ranged from 0.5 to 3.5 cd/m², and the bicycle ranged from 0.25 to 2.25 cd/m². This indicates that the bicycle for the black-clothed cyclist was much brighter than the cyclist associated with it, which was not the case for the white-clothed cyclist’s bicycle. During the ENV visual performance experiments, the brightness of the bicycle, rather than the luminance of the cyclist, may have provided the visibility to the participant. Unfortunately, this cannot be ascertained from the experimental results of the visibility investigation for the ENV clear condition.

Bar graph. Object luminance by VES for the cyclists' bicycles. Click here for more detail.
Figure 31. Bar graph. Object luminance by VES for the cyclists’ bicycles.

OBJECT BACKGROUND LUMINANCE

The object background luminance was measured for all VESs and objects at stations 2 and 4 at a distance of 61.0 m (200 ft). The effects of the measurement distance and the different station locations were also assessed during the measurement process. As with the object illuminance, the effects of the UV–A sources on the background measurements were assessed first. The analysis then continued with the assessment of the objects themselves.

Two factors should not have had an influence on the measurements made of the background luminance: the UV–A sources and the color of the object’s clothing. For UV–A to affect photometric measurements, the radiation must come into contact with an object that fluoresces. In a typical background scene, fluorescent behavior is not evident. The object type, in particular the color of the pedestrian’s or cyclist’s clothing, also should not have an influence on the background measurement because the background is not influenced by the reflectance of the object. To verify this, the means of the data measured at the 61.4-m (200-ft) distance were compared in figure 32. This analysis included all of the VES configurations that included UV–A and their baseline headlamps as well as the background measurement for the black- and white-clothed pedestrian and cyclist objects.

Bar graph. Background luminance for white- and black-clothed pedestrians by UV–A based VES. Click here for more detail.
Figure 32. Bar graph. Background luminance for white- and black-clothed pedestrians by UV–A based VES.

Figure 32 shows that the UV contribution does not influence the background luminance measurement. This is particularly true for the HID-based VESs. The HLB with three UV–A and five UV–A show an increase over the other measurements, but the differences are not statistically significant. This trend is similar to the trend for object luminance of the black-clothed perpendicular pedestrian. The results also show very little difference between the backgrounds of the two clothing colors. These results show that the UV–A sources and the object clothing color were not significant in the background luminance results and so were not included in any other analysis of the background luminance.

Pedestrian and Cyclist Objects

For the pedestrian and cyclist objects, eight luminance measurements of the background were made at varying heights. The heights matched those used for the object luminance measurements. These measurements were made for all of the VESs at each of the object presentation stations. The analysis of the data consisted of first summarizing the effect of the VES on the background measurements. Figure 33 shows this relationship by object position; the mean background luminance was not influenced by the position of the object on the roadway. The HLB, HHB, and HOH sources appear to have provided similar background luminance levels, with the HLB–LP and the HID providing slightly less.

Bar graph. Influence of VES on background luminance by pedestrian position. Click here for more detail.
Figure 33. Bar graph. Influence of VES on background luminance by pedestrian position.

The relationship of the measurement height to the object background is interesting. Figure 34 shows this relationship for each VES. In this figure, the background luminance is shown to increase as the measurement height decreases. This is to be expected because the road surface came into the view of the photometer at low measurement heights. The other interesting increase is that of the HOH. In this case, the pavement marking may have been included as part of the background, which explains why the HOH VES showed the highest background luminance for the parallel pedestrian.

Bar graph. Influence of VES on background luminance by measurement height. Click here for more detail.
Figure 34. Bar graph. Influence of VES on background luminance by measurement height.

To investigate this further, the interaction of the measurement height and the object position is shown in figure 35. The parallel pedestrian shows a higher background luminance for the ankle measurement than all of the other measurement heights. This is particularly pronounced for the HOH and the HHB measurements. Here again, the background for the parallel pedestrian may have included the pavement marking, whereas the perpendicular pedestrian and the cyclist locations included only the pavement surface itself.

Bar graph. Influence of VES on background luminance by pedestrian type and measurement height. Click here for more detail.
Figure 35. Bar graph. Influence of VES on background luminance by pedestrian type and measurement height.

The influence of the measurement distance on the background luminance for the pedestrian objects is shown in figure 36. Here the high measurements of the HOH and HHB VESs are significantly higher than the other VESs for the 61.4-m (200-ft) measurement. For the 91.5-m (300-ft), 152.5-m (500-ft), and 244-m (800-ft) measurements, the influence of the VES was not significant. An example inverse square relationship has been added to this figure also. Because the illuminance from the headlamps followed the inverse square relationship, it would be expected that the background luminance would have followed the same relationship; however, from the figure, this does not seem to be the case because there was no significant difference between the measurements at 61.0 m (200 ft) and 91.5 m (300 ft) for the HLB, HID, and HLB–LP VES types, and for all of the VESs, there was no significant difference between the 152.5 m (500-ft) and 244-m (800-ft) measurement distances. The background luminance seems to have reached a minimum ambient luminance at these distances, and the inverse square law did not apply.

Bar graph. Influence of VES on background luminance by measurement distance. Click here for more detail.
Figure 36. Bar graph. Influence of VES on background luminance by measurement distance.

The influence of the object station on the background luminance is shown in figure 37. For most VESs, there were no significant differences among the stations; however, the HOH measurement did show a higher result at station 4, which again probably was a result of the pavement marking influencing the measurement. The HOH measurement was made from the pickup truck vehicle that was used for this VES and represents the highest observation location with one of the shorter observation distances, which may have allowed the pavement marking to play a more significant role in the measurement with this VES than the other VESs.

Bar graph. Influence of station on background luminance for pedestrians by VES. Click here for more detail.
Figure 37. Bar graph. Influence of station on background luminance for pedestrians by VES.

Child’s Bicycle

For the child’s bicycle, the background luminance was considered two ways. The first was the influence of the measurement height, and the second was the influence of the object station. Figure 38 shows the influence of the measurement height on the background luminance for this object. For this relationship, the primarily pavement background behind the bicycle’s bottom-measurement height had the greatest luminance, with the middle- and the top-measurement heights showing a lower luminance. This trend was the same for all VESs, with the HOH being the highest, followed in order by the HLB, HID, HHB, and
HLB–LP. In terms of significant differences among the VESs, some of these may be grouped as having differences; however, there were no meaningful differences among all of the VESs. The background luminance of the child’s bicycle is shown in figure 39 for each of the object stations. It should be noted that the child’s bicycle never appeared at station 3. Neither the VES nor the station had an influence on the background luminance for this object.

Bar graph. Influence of measurement height on background luminance of child’s bicycle by VES. Click here for more detail.
Figure 38. Bar graph. Influence of measurement height on background luminance of child’s bicycle by VES.


Bar graph. Influence of station on background luminance for the child’s bicycle by VES. Click here for more detail.
Figure 39. Bar graph. Influence of station on background luminance for the child’s bicycle by VES.

Tire Tread

As with the child’s bicycle, the tire tread’s background luminance was analyzed first by measurement height and then by object station; the tire tread did not appear at station 3. The tire tread showed the same trends in all of the relationships shown. The influence of the measurement height is shown in figure 40. Just below the object, or the bottom-measurement height, had the highest measured luminance, followed by the middle- and the top-measurement heights. Similarly, the HOH had the highest result, followed in order by the HLB, HID, HHB, and the HLB–LP. The relationship for the background luminance and the object station is shown in figure 41. Here again, neither the VES nor the station influenced the background luminance results.

Bar graph. Influence of measurement height on background luminance for the tire tread by VES. Click here for more detail.
Figure 40. Bar graph. Influence of measurement height on background luminance for the tire tread by VES.


Bar graph. Influence of station on background luminance for the tire tread by VES. Click here for more detail.
Figure 41. Bar graph. Influence of station on background luminance for the tire tread by VES.

OBJECT REFLECTANCE

The reflectance was calculated for all of the objects used in the experimental conditions. The calculation was performed only for VES conditions with no UV–A contribution. In the case of the human objects, all black-clothed and white-clothed objects were grouped, and the reflectance values calculated for these conditions were averaged. These calculations were performed for the objects when they were dry, as they were in the clear and fog conditions, as well as wet, as they were in the rain and snow conditions. White clothing became significantly less reflective under the wet condition, while the black clothing became slightly more reflective. The tire tread object developed specular reflective attributes, as evidenced by an increase in its reflectivity for the wet condition. The reflectance of all the objects is summarized in table 7. Figure 42 also shows these values.

Table 7. Object reflectance summary.
Object Dry Wet
Black Clothing 3.02 3.83
Tire Tread 4.00 6.76
Child’s Bicycle 18.00 18.00
White Clothing 50.30 21.80


Bar graph. Reflectance of all objects both dry and wet. Click here for more detail.
Figure 42. Bar graph. Reflectance of all objects both dry and wet.

The other reflective characteristic for the cyclists were the bicycles themselves. The black-clothed cyclist rode a bicycle of a dark color, which was a mix of burgundy and black, but the wheel rims were chrome. The white-clothed cyclist rode a bicycle painted with a fluorescent orange paint, and the wheel rims were painted black. For these objects, the specular (non-Lambertian) reflection was measured. These results are presented in figure 43 for the dry conditions. As this figure indicates, the bicycle associated with the black-clothed cyclist had a much higher reflectance than that associated with the white-clothed cyclist. As discussed under object luminance, this might indicate that the bicycle used by the dark-clothed cyclist was more visible than the cyclist for this object presentation.

Bar graph. Specular reflection of all bicycle objects for both black-clothed and white-clothed cyclist. Click here for more detail.
Figure 43. Bar graph. Specular reflection of all bicycle objects for both black-clothed and white-clothed cyclist.

OBJECT FLUORESCENCE

The object fluorescence was calculated for all VES configurations that included UV–A radiation, using the equation in figure 11. These included hybrid UV–A + HLB, three UV–A + HLB, five UV–A + HLB,
hybrid UV–A + HID, three UV–A + HID, and five UV–A + HID. As a comparison, the base conditions of HID and HLB were used as the nonfluorescent comparison conditions. The means of the results are plotted for black-clothed and the white-clothed objects in figure 44 by their position on the roadway and in figure 45 by the VES type. Both figures indicate that the black-clothed objects showed no significant fluorescent behavior; however, the white-clothed objects showed fluorescence that increased with the higher levels of UV–A, as seen in figure 45.

Bar graph. Fluorescence for the black-clothed and white-clothed objects by roadway position. Click here for mroe detail.
Figure 44. Bar graph. Fluorescence for the black-clothed and white-clothed objects by roadway position.


Bar graph. Fluorescence for the black-clothed and white-clothed objects by VES type. Click here for more detail.
Figure 45. Bar graph. Fluorescence for the black-clothed and white-clothed objects by VES type.

The fluorescence level also varied by the position on the roadway. The parallel pedestrian object showed a significant difference from the perpendicular pedestrian object and the cyclist object. This is to be expected because the UV–A systems are aimed toward the direction of travel of the vehicle, and therefore, they do not provide as high an irradiance level on objects located on the side of the road as they do on objects in the middle of the roadway. Figure 46 shows the interaction of the VES type and the object position for the white-clothed objects only. In this figure, the effect of increasing the UV–A content, such as adding two more UV–A sources to the three UV–A configuration, was smaller than expected. Similarly, the performance of the hybrid source was not significantly different than having no UV–A contribution at all.

Bar graph. Fluorescence for the white-clothed objects by VES type and position on the roadway. Click here for more detail.
Figure 46. Bar graph. Fluorescence for the white-clothed objects by VES type and position on the roadway.

Because the child’s bicycle was painted with fluorescent paint, that object’s fluorescence could also be calculated; the results of this calculation are shown in figure 47 by VES type. This figure shows the response of the paint to the increasing levels of UV–A, with each level being significantly different than all of the other levels. The child’s bicycle was placed at the side of the roadway in a similar location to that of the parallel pedestrian, but it showed a more consistent increase than the pedestrian. This response is likely due to the paint used on the child’s bicycle, which was a fluorescent paint that would have had a linear response to the UV–A contribution of the VES as compared to the clothing.

Bar graph. Fluorescence of the child’s bicycle by VES type. Click here for more detail.
Figure 47. Bar graph. Fluorescence of the child’s bicycle by VES type.

The tire tread was not considered in this analysis because it was black rubber and, much like the black-clothed pedestrian, it would not fluoresce.

LUMINANCE DIFFERENCE

The difference in the luminance between each object and its background (delta symbolL) was calculated for all of the objects at all of the measured distances. The luminance difference was calculated by subtracting the background luminance from the object luminance (this is the same value as delta symbolLActual, used in the calculation of VL). Thus, delta symbolL can be positive or negative, depending on if the background luminance is the lesser or greater of the two values. Contrast, which is defined as the ratio of the luminance difference to the background luminance, will be considered later. Note that a positive delta symbolL indicates positive contrast, while a negative delta symbolL indicates a negative contrast between the object and its background.

The change in the object luminance as a result of the change in the measurement distance for the white-clothed objects was applied to the black-clothed objects to calculate the change in their luminance because the influence of distance was not measured for the black-clothed objects. Similarly, the effect of the measurement distance on the background measurements for the white-clothed objects was used for the black-clothed pedestrians to establish their background luminance. For these background calculations, there was no effect of the UV–A system applied to the calculation.

As in the object luminance analysis, the delta symbolL analysis first considered the pedestrian and cyclist objects and then the child’s bicycle and the tire tread.

Pedestrian and Cyclist Objects

Based on the differences between object luminance of the white-clothed and the black-clothed objects, delta symbolL also should have been significantly different. For this reason, the two clothing colors were investigated individually.

Black-Clothed Objects

The first relationship considered in this analysis was the influence of object position on delta symbolL, which is shown in figure 48. In general, the position had very little effect on delta symbolL. There were some transitions from positive to negative values for the headlamps that incorporated HID. In these transitions, the cyclist had a negative delta symbolL, whereas the parallel and perpendicular pedestrians had positive values of delta symbolL. The other major transition was that of the HOH VES. With that VES, all of the objects had negative values of delta symbolL because the VES provided a high background luminance and a low object luminance.

Bar graph. Luminance difference by VES for black-clothed pedestrians by object position. Click here for more detail.
Figure 48. Bar graph. Luminance difference by VES for black-clothed pedestrians by object position.

The next relationship is the change in delta symbolL with the measurement height. Shown in figure 49, the consistent relationship shown is the negative values for delta symbolL measured at the ankle height of the object; however, this resulted from the appearance of the object against the immediate background, which in this case was the road surface. It is to be expected that the luminance difference at this level would have been negative, with the black object against the gray pavement. For the measurement height and VES combinations, few significant changes were shown by the calculations.

Bar graph. Luminance difference by VES for black-clothed pedestrians by measurement height. Click here for more detail.
Figure 49. Bar graph. Luminance difference by VES for black-clothed pedestrians by measurement height.

The final aspect of the black-clothed pedestrian object is the effect of distance on delta symbolL. This relationship is shown in figure 50. Because the UV–A did not affect the black-clothed object, this relationship is shown for the visible-light-based VESs only. The important issue in this relationship is the apparent lessening of delta symbolL as the vehicle approached the object. From farther away, the object would have appeared dark against the light pavement, thus providing a negative contrast situation. As the vehicle approached, more light was cast on the front vertical side of the object, and the object luminance increased, thereby reducing the luminance difference. However, as the vehicle approached the object, the visual size got bigger, reducing the threshold difference and possibly offsetting the effect on visibility due to the reduction of the difference in luminance. For the HID-based VESs and HHB, delta symbolL actually changed from negative to positive as the vehicle approached, which means that the object would have gone through a transitional point where it had zero luminance difference and would have been invisible at that distance.

Bar graph. Luminance difference by VES for black-clothed pedestrians by measurement distance. Click here for more detail.
Figure 50. Bar graph. Luminance difference by VES for black-clothed pedestrians by measurement distance.

White-Clothed Objects

The luminance difference for the white-clothed objects was evaluated in the same manner as for the black-clothed objects. The first relationship considered is shown in figure 51. In this relationship, the luminance of the white-clothed object was so dominant as compared to the background at 61.0 m (200 ft) that delta symbolL was driven by object luminance value. This would also have applied to the values of delta symbolL by measurement height relationship, so it is not shown here.

Bar graph. Luminance difference by VES for white-clothed pedestrians by object position. Click here for more detail.
Figure 51. Bar graph. Luminance difference by VES for white-clothed pedestrians by object position.

The other relationship of interest is the change in delta symbolL with distance, which is shown in figure 52. Like the other relationships, delta symbolL was dominated by the object luminance; however, at greater distance, the object luminance fell closer to the background luminance, reducing the magnitude of delta symbolL. For the HLB, hybrid UV–A + HLB, HHB, and the HOH lamps, delta symbolL transitioned to negative values, which caused the same through-zero transition mentioned for the black-clothed object. It is noteworthy that delta symbolL for the HID-based systems did not fall as much as for the other VESs. It is not clear if this resulted from the headlamp cutoff or the strength of the HID light source.

Bar graph. Luminance difference by VES for white-clothed pedestrians by measurement distance. Click here for more detail.
Figure 52. Bar graph. Luminance difference by VES for white-clothed pedestrians by measurement distance.

Child’s Bicycle

As with the white-clothed object, delta symbolL for the child’s bicycle was primarily dependent on the object luminance, as seen in figure 53. In this case, the background luminance from each of the VESs did not show a dramatic difference. This lack of change resulted in delta symbolL of the bicycle following that of the object luminance. All of the delta symbolL values for the VESs were positive, with the HLB- and HID-based VESs providing the highest values of delta symbolL and the HLB–LP providing the lowest values.

Bar graph. Luminance difference by VES for the child's bicycle. Click here for more detail.
Figure 53. Bar graph. Luminance difference by VES for the child’s bicycle.

Tire Tread

The delta symbolL of the tire tread is shown in figure 54. In this relationship, the tire tread appeared with both positive and negative values for delta symbolL based on the performance of the VES. The HLB had the highest negative delta symbolL, with the HHB providing the highest positive delta symbolL.

Bar graph. Luminance difference by VES for the tire tread. Click here for more detail.
Figure 54. Bar graph. Luminance difference by VES for the tire tread.

VISIBILITY LEVEL

As mentioned, the visibility level (VL) metric considers all of the measurements already mentioned; however, the VL is also dependent on the age of the participant. For this analysis, the age of the observer was considered for a young (23-year-old) observer, a middle-aged (45-year-old) observer, and an older (70-year-old) observer. The VL was analyzed in the same manner as the other measurements; the pedestrians were considered first, followed by the child’s bicycle and then the tire tread.

Pedestrian and Cyclist Objects

The mean VL for the pedestrian and cyclist objects are shown in figure 55. All of these measurements were taken at the 61.4-m (200-ft) measurement distance. The VL trend followed that of the object luminance. The addition of the UV–A increased the VL, with the effect on the halogen-based VESs being higher than that on the HID-based VESs. The resulting VL for the HHB was lower than that of the five UV–A + HLB, and it was not significantly different than the five UV–A + HID. The HOH provided the lowest VL, with the HLB–LP performing on par with the HLB. The age of the observer reduced the VL by approximately one-third for each age group. That is to say, the middle-age group had about two-thirds the VL of the younger group, and the older group had about one-third the VL of the younger group.

Bar graph. Visibility level by age and VES for the pedestrian objects. Click here for more detail.
Figure 55. Bar graph. Visibility level by age and VES for the pedestrian objects.

The VL analysis then considered each of the pedestrian types individually. According to the Adrian model, the ratio with respect to age will be the same for all of the VESs considered, and so it was not included in the remaining VL analyses.(1)

Black-Clothed Objects

For the black-clothed objects, the addition of the UV–A sources did not make a significant difference to the object luminance, so only the non-UV–A VESs were considered as part of this analysis. Figure 56 shows the VL by pedestrian type for all of the black-clothed objects. In this relationship, the parallel object showed the highest VL for HID, HHB, and the HLB–LP VESs, with the perpendicular object having lower values and the cyclist having the lowest values. For the HOH VES, there was no difference among the object locations, and for the HLB, there was no difference between the parallel and perpendicular objects, with the cyclist being slightly lower.

Bar graph. Visibility level by VES for the black-clothed pedestrian objects by position. Click here for more detail.
Figure 56. Bar graph. Visibility level by VES for the black-clothed pedestrian objects by position.

The effect of the distance on the VL is shown in figure 57. As the vehicle approached the object, the visual size of the pedestrian increased, so the object became easier to see. In terms of the VL, the increase in the size would have caused the VL to increase. This was seen in the HLB and the HLB–LP VESs, but the other VESs did not show the same trend. For the HID VES, the VL actually decreased for the 91.5-m
(300-ft) and 152.5-m (500-ft) distances and then increased again at the 61.4-m (200-ft) distance. Compare these values to the contrast values shown in figure 50; there was a negative to positive transition of contrast with distance that impacted the calculated value of VL. This means that the object could be seen at 244 m (800 ft), then was more difficult to see at 152.5 m (500 ft) and 91.5 m (300 ft), and then was more easily seen at 61.0 m (200 ft). The HOH VES showed the opposite behavior, where the VL was higher at 91.5 m (300 ft) and 152.5 m (500 ft). The HHB VES performed in a much less predictable manner with a low point at 91.5 m (300 ft) and reached the highest point at 152.5 m (500 ft). This was not expected because the contrast for the same VES is a linear change with distance.

Bar graph. Visibility level by VES for the black-clothed pedestrian objects by distance. Click here for more detail.
Figure 57. Bar graph. Visibility level by VES for the black-clothed pedestrian objects by distance.

White-Clothed Objects

The VL for the white-clothed objects followed the same path as the object luminance, with the influence of the UV–A contribution increasing the VL over the base condition as shown in figure 58. The change for the white-clothed objects as compared to the black-clothed objects was that the cyclist had a higher VL than the parallel and perpendicular pedestrians for almost all of the tested VESs. The other interesting result is that the VESs with a UV–A component were as high as or higher in VL than the HHB. The HOH continued to perform the worst, and the HLB and HLB–LP conditions performed the same. The other point is that in both conditions that included the hybrid UV–A headlamp, the gain in VL was just barely significant as compared to the contributions from the other UV–A sources.

Bar graph. Visibility level by VES for the white-clothed pedestrian objects by position. Click here for more detail.
Figure 58. Bar graph. Visibility level by VES for the white-clothed pedestrian objects by position.

For the influence of the distance, as shown in figure 59, all of the VESs performed in a similar manner, with the VL increasing as the distance decreased. This is to be expected when comparing the contrast results in figure 52. There is a constant rise in the contrast as the distance decreases with only a few negative to positive transitions at the 152.5 m (500 ft) to 244 m (800 ft) distance. This trend for contrast and the increase in the object size as the distance closes will result in an increase in the VL.

Bar graph. Visibility level for the white-clothed pedestrians by distance and VES. Click here for more detail.
Figure 59. Bar graph. Visibility level for the white-clothed pedestrians by distance and VES.

Child’s Bicycle

As with the white-clothed objects, the VL for the child’s bicycle followed the object luminance (figure 60). The HID system with the sharp cutoff had the best performance, followed by the HLB-based systems. The HHB, HOH, HLB, and HLB–LP showed no significant differences between the VESs. The effect of age, about one-third for each age step, was the same as for the pedestrian objects.

Bar graph. Visibility level for the child's bicycle by age and VES. Click here for more detail.
Figure 60. Bar graph. Visibility level for the child’s bicycle by age and VES.

Tire Tread

The VL for the tire tread at 61.0 m (200 ft) is shown in figure 61. The UV–A effect is not shown because the tire tread was not influenced by the UV–A. In this relationship, the HHB showed the highest VL, followed by the HOH and HID VESs, and then the HLB and HLB–LP configurations. Comparing these values to the contrast results shown in figure 54, the VL generally followed the contrast, with the highest contrast belonging to the HHB system. As with the other objects, the effect of age followed the same trend of one-third increase or decrease per age group.

Bar graph. Visibility level for the tire tread by age and VES. Click here for more detail.
Figure 61. Bar graph. Visibility level for the tire tread by age and VES.

 

Previous    Table of Contents    Next
Federal Highway Administration | 1200 New Jersey Avenue, SE | Washington, DC 20590 | 202-366-4000
Turner-Fairbank Highway Research Center | 6300 Georgetown Pike | McLean, VA | 22101