• Federal Highway Administration >
• Publications >
• Research Publications >
• Human Factors >
• 04135 >
• 508txt.Cfm >
• Enhanced Night Visibility, Volume IV: Phase II—Study 2: Visual Performance During Nighttime Driving in Rain

Enhanced Night Visibility, Volume IV: Phase II—Study 2: Visual Performance During Nighttime Driving in Rain

PDF files can be viewed with the Acrobat® Reader®

Alternate Text

Figure 1. Photo. Pedestrian in black clothing. This photograph shows one of the objects used in the study. It is a side view of a person dressed in black standing against a plain background. The person is wearing a black ski cap with face mask, protective glasses, a black jacket, black pants, dark boots, and black gloves. Back to Figure 1.

Figure 2. Photo. Cyclist in white clothing. This photograph shows one of the objects used in the study. It is a side view of a person riding a mountain bike with a bright orange frame against a plain background. The person is wearing a white short-sleeved shirt over a black long-sleeved shirt, white pants, a black ski cap with face mask, protective glasses, dark boots, and black gloves. Back to Figure 2.

Figure 3. Photo. Pedestrian in white clothing. This photograph shows one of the objects used in the study. It is a side view of a person wearing a white short-sleeved shirt over a black long-sleeved shirt, white pants, black ski cap with face mask, protective glasses, dark boots, and black gloves. The person is standing against a plain background. Back to Figure 3.

Figure 4. Photo. Child’s bicycle. This photograph shows one of the objects used in the study. It is a side view of a child’s bicycle propped upright against a plain background. The bicycle has a yellow frame, pink tire rims, and white seat, tires, and pedals. Back to Figure 4.

Figure 5. Photo. Tire tread. This photograph shows one of the objects used in the study. It is an intact circular tire tread with the sidewall removed. The tire tread is lying on its side against a plain background.
Back to Figure 5.

Figure 6. Diagram. Data collection display screen. The graphic shows a sample display screen used to record the data collected during an onroad trial. The top section identifies the driver and passenger by participant ID, age, and gender. The middle section describes the trial setup, including the VES, target order, day, participant number, output file name, and the specific experiment being conducted. The bottom section allows the experimenter to see the object order used in the trial and record the driver and passenger detection and recognition distances. Back to Figure 6.

Figure 7. Photo. Five or three UV–A + HLB. This photo shows the front end of a white sport utility vehicle with the headlamps removed and a modular light rack attached. In front of the vehicle’s own headlamps are two HLB headlamps. Taking up almost all the space between them are three round, black UV–A headlamps. Two more UV–A headlamps hang beneath the HLB headlamps. Back to Figure 7.

Figure 8. Photo. HOH or HHB. This photo shows two high output halogen headlamps and two halogen high beam headlamps attached to the front of an experimental pickup truck on a modular light rack. One high output halogen headlamp and one halogen high beam headlamp are housed together on the right front of the vehicle, and the other pair is housed together on the left front of the vehicle. Back to Figure 8.

Figure 9. Photo. Hybrid UV–A + HID. This photo of two hybrid UV–A headlamps and two high intensity discharge headlamps shows how the headlamps are attached to the front of an experimental SUV on a modular light rack. The high intensity discharge headlamps are mounted on the far right and far left front, and the hybrid UV–A lights are mounted adjacent to them toward the center of the vehicle. Back to Figure 9.

Figure 10. Photo. HLB–LP with IR–TIS. This photo of the front of the experimental sedan used in the study shows that it is equipped with standard halogen low beam headlamps. A circular opening for the infrared camera in the center of the front grille indicates the presence of the infrared thermal imaging system. Back to Figure 10.

Figure 11. Photo. Smart Road. This daytime aerial photo shows a section of the Virginia Smart Road. The section of road is fairly straight with one gradual curve. The rural area has no light-producing structures in the immediate vicinity. Back to Figure 11.

Figure 12. Photo. Smart Road rain towers. This daytime photo of the Smart Road shows several rain towers located just outside the guard rail. The towers overhang the edge of the pavement. They are producing a heavy rainfall across the width of the road. An SUV can barely be seen in the midst of the rainfall. The photo is taken beyond the rainfall and shows dry pavement after the rainfall. Back to Figure 12.

Figure 13. Diagram. Locations where the objects were presented for the adverse weather condition. This diagram shows a map of the Smart Road used in the experiments for this study. The road is depicted as being fairly straight with one gradual curve up toward the top. There is a turnaround loop at each end of the road. The top turnaround is in the upper left corner of the diagram and the bottom turnaround is in the lower right corner, indicating that there is a grade. A legend contains an arrow labeled as “Object Location.” There are four arrows along the side of the road indicating the four locations where the objects were presented. Locations 1 and 2 are on one side of the road, indicating that a driver would see the objects placed there during the drive from the top to the bottom of the road. Locations 4 and 5 are on the opposite side of the road from the first two locations, indicating that a driver would see the objects placed there during the drive from the bottom of the road to the top of the road. Locations 1, 2, 4, and 5 are grouped toward the middle segment of the course, which is marked with the note: “Adverse Weather Testing Area Where Rain Was Generated.” Locations 1 and 5 are close together but on opposite sides of the road. Locations 2 and 4 are also close together but on opposite sides of the road. The diagram shows that all locations are on straight sections of the road. Back to Figure 13.

Figure 14. Bar graph. Results for the interaction: VES by Object by Age for IR–TIS. The graph is titled “Detection Distance by Age Group and Object Using IR–TIS in the Rain Condition.” The Y-axis shows the mean detection distance in feet. The X-axis includes each of the seven objects used in the study. Three bars are presented for each of the objects, representing the three age groups: young, middle, and older. Standard error bars are also provided for each graph bar. Following are the approximate ranges of detection distances for each age group. The younger drivers’ detection distances ranged from 110 feet (perpendicular pedestrian in black clothing) to 230 feet (child’s bicycle). The detection distances for the middle age group ranged from 105 feet (parallel pedestrian in black clothing and perpendicular pedestrian in black clothing) to 270 feet (parallel pedestrian in white clothing). The detection distances for the older drivers ranged from 105 feet (perpendicular pedestrian in black clothing) to 270 feet (cyclist in white clothing). Standard error bars for each set of object groups are comparable, with the detection distance of the child’s bicycle for the young age group producing a slightly greater standard error than other objects and age groups.
Back to Figure 14.

Figure 15. Bar graph. Results for the interaction: VES by Object by Age for HLB–LP. The graph is titled “Detection Distance by Age Group and Object Using HLB–LP in the Rain Condition.” The Y-axis shows the mean detection distance in feet. The X-axis includes each of the seven objects used in the study. Three bars are presented for each of the objects, representing the three age groups: young, middle, and older. Standard error bars are also provided for each graph bar. Following are the approximate ranges of detection distances for each age group. The younger drivers’ detection distances ranged from 110 feet (perpendicular pedestrian in black clothing) to 270 feet (parallel pedestrian in white clothing). The detection distances for the middle age group ranged from 105 feet (perpendicular pedestrian in black clothing) to 245 feet (parallel pedestrian in white clothing). The detection distances for the older drivers ranged from 100 feet (tire tread) to 275 feet (cyclist in white clothing). Standard error bars for each set of object groups are comparable.
Back to Figure 15.

Figure 16. Bar graph. Results for the interaction: VES by Object by Age for HOH. The graph is titled “Detection Distance by Age Group and Object Using HOH in the Rain Condition.” The Y-axis shows the mean detection distance in feet. The X-axis includes each of the seven objects used in the study. Three bars are presented for each of the objects, representing the three age groups: young, middle, and older. Standard error bars are also provided for each graph bar. Following are the approximate ranges of detection distances for each age group. The younger drivers’ detection distances ranged from 130 feet (perpendicular pedestrian in black clothing) to 290 feet (perpendicular pedestrian in white clothing). The detection distances for the middle age group ranged from 105 feet (perpendicular pedestrian in black clothing) to 275 feet (parallel pedestrian in white clothing). The detection distances for the older drivers ranged from 110 feet (tire tread) to 270 feet (cyclist in white clothing and parallel pedestrian in white clothing). Standard error bars for each set of object groups are comparable, with the detection distance of the cyclist in white clothing for the older age group producing a greater standard error than other objects and age groups.
Back to Figure 16.

Figure 17. Bar graph. Results for the interaction: VES by Object by Age for HHB. The graph is titled “Detection Distance by Age Group and Object Using HHB in the Rain Condition.” The Y-axis shows the mean detection distance in feet. The X-axis includes each of the seven objects used in the study. Three bars are presented for each of the objects, representing the three age groups: young, middle, and older. Standard error bars are also provided for each graph bar. Following are the approximate ranges of detection distances for each age group. The younger drivers’ detection distances ranged from 110 feet (tire tread) to 265 feet (perpendicular pedestrian in white clothing). The detection distances for the middle age group ranged from 110 feet (tire tread) to 255 feet (cyclist in white clothing). The detection distances for the older drivers ranged from 110 feet (parallel pedestrian in black clothing and perpendicular pedestrian in black clothing) to 260 feet (parallel pedestrian in white clothing). Standard error bars for similar type of objects are comparable. The cyclist in white clothing and child’s bicycle for the young age group have the smallest standard error followed by the pedestrians in black clothing for all age groups. Back to Figure 17.

Figure 18. Bar graph. Results for the interaction: VES by Object by Age for five UV–A + HLB. The graph is titled “Detection Distance by Age Group and Object Using Five UV–A + HLB in the Rain Condition.” The Y-axis shows the mean detection distance in feet. The X-axis includes each of the seven objects used in the study. Three bars are presented for each of the objects, representing the three age groups: young, middle, and older. Standard error bars are also provided for each graph bar. Following are the approximate ranges of detection distances for each age group. The younger drivers’ detection distances ranged from 140 feet (parallel pedestrian in black clothing) to 305 feet (parallel pedestrian in white clothing). The detection distances for the middle age group ranged from 125 feet (parallel pedestrian in black clothing) to 305 feet (perpendicular pedestrian in white clothing). The detection distances for the older drivers ranged from 120 feet (parallel pedestrian in black clothing) to 325 feet (cyclist in white clothing). Standard error bars for each set of object groups are not comparable. The detection distances for the middle age group for parallel pedestrian in white clothing, and parallel and perpendicular pedestrian in black clothing produced smaller standard errors than other objects and age groups. Back to Figure 18.

Figure 19. Bar graph. Results for the interaction: VES by Object by Age for three UV–A + HLB. The graph is titled “Detection Distance by Age Group and Object Using Three UV–A + HLB in the Rain Condition.” The Y-axis shows the mean detection distance in feet. The X-axis includes each of the seven objects used in the study. Three bars are presented for each of the objects, representing the three age groups: young, middle, and older. Standard error bars are also provided for each graph bar. Following are the approximate ranges of detection distances for each age group. The younger drivers’ detection distances ranged from 140 feet (parallel pedestrian in black clothing, perpendicular pedestrian in black clothing, and tire tread) to 310 feet (parallel pedestrian in white clothing). The detection distances for the middle age group ranged from 135 feet (parallel pedestrian in black clothing and perpendicular pedestrian in black clothing) to 295 feet (parallel pedestrian in white clothing). The detection distances for the older drivers ranged from 140 feet (parallel pedestrian in black clothing, perpendicular pedestrian in black clothing, and tire tread) to 300 feet (parallel pedestrian in white clothing). Standard error bars for each set of object groups are comparable, with the detection distance of the cyclist in white clothing for the middle age group and parallel pedestrian in white clothing for the older group producing a slightly greater standard error than other objects and age groups. Back to Figure 19.

Figure 20. Bar graph. Results for the interaction: VES by Object by Age for hybrid UV–A + HLB. The graph is titled “Detection Distance by Age Group and Object Using Hybrid UV–A + HLB in the Rain Condition.” The Y-axis shows the mean detection distance in feet. The X-axis includes each of the seven objects used in the study. Three bars are presented for each of the objects, representing the three age groups: young, middle, and older. Standard error bars are also provided for each graph bar. Following are the approximate ranges of detection distances for each age group. The younger drivers’ detection distances ranged from 140 feet (parallel pedestrian in black clothing and perpendicular pedestrian in black clothing) to 295 feet (cyclist in white clothing). The detection distances for the middle age group ranged from 120 feet (parallel pedestrian in black clothing) to 295 feet (cyclist in white clothing). The detection distances for the older drivers ranged from 110 feet (perpendicular pedestrian in black clothing) to 305 (cyclist in white clothing). Standard error bars for each set of object groups are comparable, with exception of the cyclist in white clothing for the old age group producing a greater standard error and the perpendicular pedestrian in black clothing for the middle and old age groups producing a smaller standard error than other objects and age groups. Back to Figure 20.

Figure 21. Bar graph. Results for the interaction: VES by Object by Age for HLB. The graph is titled “Detection Distance by Age Group and Object Using HLB in the Rain Condition.” The Y-axis shows the mean detection distance in feet. The X-axis includes each of the seven objects used in the study. Three bars are presented for each of the objects, representing the three age groups: young, middle, and older. Standard error bars are also provided for each graph bar. Following are the approximate ranges of detection distances for each age group. The younger drivers’ detection distances ranged from 140 feet (parallel pedestrian in black clothing, perpendicular pedestrian in black clothing, and tire tread) to 280 feet (perpendicular pedestrian in white clothing). The detection distances for the middle age group ranged from 115 feet (perpendicular pedestrian in black clothing) to 270 feet (perpendicular pedestrian in white clothing). The detection distances for the older drivers ranged from 120 feet (parallel pedestrian in black clothing) to 270 feet (perpendicular pedestrian in white clothing). Standard error bars for each set of object groups are comparable, with the exception of some of the detections for the younger age group having a slighter greater standard error.
Back to Figure 21.

Figure 22. Bar graph. Results for the interaction: VES by Object by Age for five UV–A + HID. The graph is titled “Detection Distance by Age Group and Object Using Five UV–A + HID in the Rain Condition.” The Y-axis shows the mean detection distance in feet. The X-axis includes each of the seven objects used in the study. Three bars are presented for each of the objects, representing the three age groups: young, middle, and older. Standard error bars are also provided for each graph bar. Following are the approximate ranges of detection distances for each age group. The younger drivers’ detection distances ranged from 105 feet (perpendicular pedestrian in black clothing) to 305 feet (cyclist in white clothing). The detection distances for the middle age group ranged from 110 feet (perpendicular pedestrian in black clothing) to 300 feet (parallel pedestrian in white clothing). The detection distances for the older drivers ranged from 100 feet (parallel pedestrian in black clothing and perpendicular pedestrian in black clothing) to 280 feet (perpendicular pedestrian in white clothing). Standard error bars for each set of object groups are not comparable. The detection distances of the parallel pedestrian and perpendicular pedestrian in black clothing for the three age groups have smaller standard errors than the other objects. Among young participants, detection distance for the perpendicular pedestrian in white clothing has a larger standard error than the other objects and age groups.
Back to Figure 22.

Figure 23. Bar graph. Results for the interaction: VES by Object by Age for three UV–A + HID. The graph is titled “Detection Distance by Age Group and Object Using Three UV–A + HID in the Rain Condition.” The Y-axis shows the mean detection distance in feet. The X-axis includes each of the seven objects used in the study. Three bars are presented for each of the objects, representing the three age groups: young, middle, and older. Standard error bars are also provided for each graph bar. Following are the approximate ranges of detection distances for each age group. The younger drivers’ detection distances ranged from 130 feet (perpendicular pedestrian in black clothing) to 275 feet (parallel pedestrian in white clothing). The detection distances for the middle age group ranged from 105 feet (parallel pedestrian in black clothing) to 275 feet (parallel pedestrian in white clothing). The detection distances for the old age group ranged from 105 feet (parallel pedestrian in black clothing) to 280 feet (cyclist in white clothing). Standard error bars for each set of object groups are comparable, with the detection distance of the cyclist in white clothing for the old age group producing a greater standard error than other objects and age groups. Back to Figure 23.

Figure 24. Bar graph. Results for the interaction: VES by Object by Age for hybrid UV–A + HID. The graph is titled “Detection Distance by Age Group and Object Using Hybrid UV–A + HID in the Rain Condition.” The Y-axis shows the mean detection distance in feet. The X-axis includes each of the seven objects used in the study. Three bars are presented for each of the objects, representing the three age groups: young, middle, and older. Standard error bars are also provided for each graph bar. Following are the approximate ranges of detection distances for each age group. The younger drivers’ detection distances ranged from 105 feet (perpendicular pedestrian, black clothing) to 295 feet (parallel pedestrian, white clothing). The detection distances for the middle age group ranged from 105 feet (parallel pedestrian, black clothing) to 260 feet (parallel pedestrian, white clothing). The detection distances for the old age group ranged from 100 feet (perpendicular pedestrian, black clothing) to 295 (perpendicular pedestrian, white clothing). Standard error bars for each set of object groups are comparable; however, detection distances of the cyclist in white clothing for the old age group, parallel pedestrian in white clothing for the young age group, perpendicular pedestrian in white clothing for the old age group, and child’s bicycle for the old age group have a greater standard error than other objects and age groups. Back to Figure 24.

Figure 25. Bar graph. Results for the interaction: VES by Object by Age for HID. The graph is titled “Detection Distance by Age Group and Object Using HID in the Rain Condition.” The Y-axis shows the mean detection distance in feet. The X-axis includes each of the seven objects used in the study. Three bars are presented for each of the objects, representing the three age groups: young, middle, and older. Standard error bars are also provided for each graph bar. Following are the approximate ranges of detection distances for each age group. The younger drivers’ detection distances ranged from 110 feet (parallel pedestrian, black clothing and perpendicular pedestrian, black clothing) to 240 feet (cyclist, white clothing and parallel pedestrian, white clothing). The middle age group had detection distances ranging from 110 feet (perpendicular pedestrian, black clothing) to 240 feet (cyclist, white clothing). The older age group had detection distances ranging from 110 feet (perpendicular pedestrian, black clothing) to 270 feet (parallel pedestrian, white clothing). Standard error bars for each set of object groups are comparable; however, the detection distance of the parallel pedestrian in white clothing for the old age group has a greater standard error than other objects and age groups. Back to Figure 25.

Figure 26. Bar graph. Results on detection distances for the VES by Object interaction for pedestrians and cyclist with white clothing. The graph is titled “Detection Distances for the VES by Object Interaction for the Rain Condition.” The Y-axis shows the mean detection distance in feet. The X-axis includes the 12 different VESs evaluated in the study. Three bars are presented for each VES, representing the following three objects: cyclist with white clothing, parallel pedestrian with white clothing, and perpendicular pedestrian with white clothing. Standard error bars are also provided for each graph bar. Following are the approximate ranges of recognition distances for each object: cyclist with white clothing ranged from 230 feet with the HLB–LP and HID to 300 feet with the five UV–A + HLB; parallel pedestrian with white clothing ranged from 240 feet with the IR–TIS to 305 feet with the three UV–A + HLB; perpendicular pedestrian with white clothing ranged from 220 feet with the IR–TIS to 300 feet with the five UV–A + HLB. Standard error bars for each set of object groups are comparable. Back to Figure 26.

Figure 27. Bar graph. Results on detection distances for the VES by Object interaction for pedestrians with black clothing and other objects. The graph is titled “Detection Distances for the VES by Object Interaction for the Rain Condition.” The Y-axis shows the mean detection distance in feet. The X-axis includes the 12 different VESs evaluated in the study. Four bars are presented for each VES, representing the following four objects: parallel pedestrian with black clothing, perpendicular pedestrian with black clothing, child’s bicycle, and tire tread. Standard error bars are also provided for each graph bar. Following are the approximate ranges of detection distances for each object: parallel pedestrian with black clothing ranged from 120 feet with the hybrid UV–A + HID to 145 feet with the three UV–A + HLB; perpendicular pedestrian with black clothing ranged from 105 feet with the hybrid UV–A + HID to 145 feet with the five UV–A + HLB; child’s bicycle ranged from 175 feet with the HHB to 225 feet with the five UV–A + HLB; tire tread ranged from 120 feet with the IR–TIS and the HHB to 155 feet with the five UV–A + HLB. Standard error bars for each set of object groups are comparable. Back to Figure 27.

Figure 28. Bar graph. Results on recognition distances for the VES by Object interaction for pedestrians and cyclist with white clothing. The graph is titled “Recognition Distances for the VES by Object Interaction for the Rain Condition.” The Y-axis shows the mean recognition distance in feet. The X-axis includes the 12 different VESs evaluated in the study. Three bars are presented for each VES, representing the following three objects: cyclist with white clothing, parallel pedestrian with white clothing, and perpendicular pedestrian with white clothing. Standard error bars are also provided for each graph bar. Following are the approximate ranges of recognition distances for each object: cyclist with white clothing ranged from 195 feet with the HID to 260 feet with the five UV–A + HLB; parallel pedestrian with white clothing ranged from 210 feet with the HLB–LP to 275 feet with the three UV–A + HLB; perpendicular pedestrian with white clothing ranged from 200 feet with the IR–TIS and HID to 270 feet with the five UV–A + HLB. Standard error bars for each set of object groups are comparable. Back to Figure 28.

Figure 29. Bar graph. Results on recognition distances for the VES by Object interaction for pedestrians with black clothing and other objects. The graph is titled “Recognition Distances for the VES by Object Interaction for the Rain Condition.” The Y-axis shows the mean recognition distance in feet. The X-axis includes the 12 different VESs evaluated in the study. Four bars are presented for each VES, representing the following four objects: parallel pedestrian with black clothing, perpendicular pedestrian with black clothing, child’s bicycle, and tire tread. Standard error bars are also provided for each graph bar. Following are the approximate ranges of recognition distances for each object: parallel pedestrian with black clothing ranged from 100 feet with the hybrid UV–A + HID and the HID to 120 feet with the three UV–A + HLB; perpendicular pedestrian with black clothing ranged from 85 feet with the hybrid UV–A + HID to 130 feet with the five UV–A + HLB; child’s bicycle ranged from 155 feet with the HHB to 200 feet with the five UV–A + HLB; tire tread ranged from 100 feet with the IR–TIS to 135 feet with the five UV–A + HLB, three UV–A + HLB, and hybrid UV–A + HLB. Standard error bars for each set of object groups are comparable. Back to Figure 29.

Figure 30. Bar graph. Bonferroni post hoc results for the main effect: VES. The graph is titled: “Detection and Recognition Distances by VES for the Rain Condition.” The Y-axis shows the mean distance in feet. The X-axis includes the 12 VESs evaluated in the study. There are two bars for each VES, one for detection distance and one for recognition distance. Letters above each bar note the groupings for the means (same letter are not significantly different) based on Bonferroni results. Detection distances ranged from approximately 180 feet with IR–TIS, HID, and HLB–LP to 225 feet with five UV–A + HLB. Five UV–A + HLB was significantly different from all the VESs, except hybrid UV–A + HLB and three UV–A + HLB. Detection distances with IR–TIS, HID, and HLB–LP were significantly different from all HLB configurations and five UV–A + HID. Recognition distances ranged approximately from 155 feet with IR–TIS to 195 feet with five UV–A + HLB. The average recognition distance for five UV–A + HLB was not significantly different from three UV–A + HLB and hybrid UV–A + HLB, but was significantly different from the other VESs. Recognition distances for IR–TIS, HID, and HLB–LP were significantly different from all recognition distances for the HLB configurations. Back to Figure 30.

Figure 31. Bar graph. Bonferroni post hoc results for the main effect: Object. The graph is titled: “Detection and Recognition Distances by Object for the Rain Condition.” The Y-axis shows the mean distance in feet. The X-axis includes the seven objects used in the study. There are two bars for each object, one for detection distance and one for recognition distance. Letters above each bar note the groupings for the means (same letter are not significantly different) based on Bonferroni results. The detection distances for the cyclist and pedestrians in white clothing ranged approximately from 260 feet to 270 feet. There were no significant differences among them. The detection distances for the perpendicular and parallel pedestrian in black clothing were approximately 120 feet and 125 feet, respectively, and this difference was not significant. Child’s bike and tire tread with approximate corresponding detection distances of 200 feet and 135 feet, were significantly different from each other and from the other objects. The recognition distances for the cyclist and the pedestrians in white clothing ranged approximately from 225 feet to 240 feet and this graph demonstrates that they differ significantly from all the other objects. Recognition distances for the cyclist and parallel pedestrian in white clothing differed significantly from each other. There was no significant difference in recognition distance between perpendicular pedestrian and cyclist or parallel pedestrian (all in white clothing). Recognition distances for pedestrians in black clothing differed significantly from all the other objects except parallel pedestrian and tire tread that did not differ significantly from each other. The recognition distance of child’s bicycle (approximately 180 feet) was significantly different from all the other objects. Back to Figure 31.

Figure 32. Bar graph. Bonferroni post hoc results on the ratings evaluating detection for the main effect: VES. The graph is titled: “Statement 1: This vision enhancement system allowed me to detect objects sooner than my regular headlights.” The Y-axis shows the mean rating for the statement and starts at 1 (“strongly agree”) and ends at 7 (“strongly disagree”). A lower rating indicates a more desirable system. The X-axis includes the 12 VESs evaluated in the study. Five UV–A + HID (approximately 1.9) and three UV–A +HID (approximately 2.1) had the lowest ratings, and were significantly lower than IR–TIS (approximately 4.8), HHB (approximately 4.0), and HLB–LP (approximately 3.3), which had the highest ratings. IR–TIS differed significantly from every other VESs, except HHB. The mean rating for HLB (baseline) was approximately 2.7.
Back to Figure 32.

Figure 33. Bar graph. Bonferroni post hoc results on the ratings evaluating recognition for the main effect: VES. The graph is titled: “Statement 2: This vision enhancement system allowed me to recognize objects sooner than my regular headlights.” The Y-axis shows the mean rating for the statement and starts at 1 (“strongly agree”) and ends at 7 (“strongly disagree”). A lower rating indicates a more desirable system. The X-axis includes the 12 VESs evaluated in the study. Five UV–A + HID (approximately 1.8) and three UV–A +HID (approximately 2.1) had the lowest ratings, and were significantly lower than IR–TIS (approximately 4.7), HHB (approximately 4.0), and HLB–LP (approximately 3.3), which had the highest ratings. IR–TIS differed significantly from every other VESs, except HHB. The mean rating for HLB (baseline) was approximately 2.7. Back to Figure 33.

Figure 34. Bar graph. Bonferroni post hoc results on the ratings evaluating lane-keeping assistance for the main effect: VES. The graph is titled: “Statement 3: This vision enhancement system helped me to stay on the road (not go over the lines) better than my regular headlights.” The Y-axis shows the mean rating for the statement and starts at 1 (“strongly agree”) and ends at 7 (“strongly disagree”). A lower rating indicates a more desirable system. The X-axis includes the 12 VESs evaluated in the study. Five UV–A + HID (approximately 2.0) and three UV–A +HID (approximately 2.1) had the lowest ratings, and were significantly lower than IR–TIS (approximately 5.1), HHB (approximately 4.3), and HLB–LP (approximately 3.5), which had the highest ratings. IR–TIS differed significantly from every other VESs, except HHB. The mean rating for HLB (baseline) was approximately 3.0. Back to Figure 34.

Figure 35. Bar graph. Bonferroni post hoc results on the ratings evaluating roadway direction for the main effect: VES. The graph is titled: “Statement 4: This vision enhancement system allowed me to see which direction the road was heading (i.e., left, right, straight) beyond my regular headlights.” The Y-axis shows the mean rating for the statement and starts at 1 (“strongly agree”) and ends at 7 (“strongly disagree”). A lower rating indicates a more desirable system. The X-axis includes the 12 VESs evaluated in the study. Five UV–A + HID (approximately 2.1) had the lowest rating and was significantly lower than IR–TIS (approximately 4.9), HHB (approximately 4.2), and HLB–LP (approximately 3.2), which had the highest ratings. IR–TIS differed significantly from every other VESs, except HHB. Five UV–A + HID is also significantly lower than HLB–LP, although not different than any other VES. The mean rating for HLB (baseline) was approximately 3.1. Back to Figure 35.

Figure 36. Bar graph. Bonferroni post hoc results on the ratings evaluating visual discomfort for the main effect: VES. The graph is titled: “Statement 5: This vision enhancement system did not cause me any more visual discomfort than my regular headlights.” The Y-axis shows the mean rating for the statement and starts at 1 (“strongly agree”) and ends at 7 (“strongly disagree”). A lower rating indicates a more desirable system. The X-axis includes the 12 VESs evaluated in the study. Five UV–A + HID (approximately 1.5) had the lowest rating and was significantly lower than IR–TIS (approximately 4.3) and HHB (approximately 3.7) which had the highest ratings. IR–TIS differed significantly from every other VESs, except HHB. The mean rating for HLB (baseline) was approximately 2. Back to Figure 36.

Figure 37. Bar graph. Bonferroni post hoc results on the ratings evaluating overall safety for the main effect: VES. The graph is titled: “Statement 6: This vision enhancement system makes me feel safer when driving on the roadways at night than my regular headlights.” The Y-axis shows the mean rating for the statement and starts at 1 (“strongly agree”) and ends at 7 (“strongly disagree”). A lower rating indicates a more desirable system. The X-axis includes the 12 VESs evaluated in the study. Five UV–A + HID (approximately 1.8) and three UV–A + HID (approximately 1.8) had the lowest ratings, and were significantly lower than IR–TIS (approximately 4.9), HHB (approximately 4.1), and HLB–LP (approximately 3.2), which had the highest ratings. IR–TIS was significantly different from all the other VESs except HHB. The mean rating for HLB (baseline) was approximately 2.7. Back to Figure 37.

Figure 38. Bar graph. Bonferroni post hoc results on the overall rating for the main effect: VES. The graph is titled: “Statement 7: This is a better vision enhancement system than my regular headlights.” The Y-axis shows the mean rating for the statement and starts at 1 (“strongly agree”) and ends at 7 (“strongly disagree”). A lower rating indicates a more desirable system. The X-axis includes the 12 VESs evaluated in the study. Five UV–A + HID (approximately 1.7) had the lowest rating and was significantly lower than IR–TIS (approximately 4.8) and HHB (approximately 3.9), which had the highest ratings. IR–TIS differed significantly from every other VES, except HHB. The mean rating for HLB (baseline) was approximately 2.6. Back to Figure 38.

Figure 39. Equation. Braking distance. d subscript BD equals V squared divided by the following: the sum of f and uppercase G, end of sum, that sum multiplied by lowercase g, multiplied by 2. Back to Figure 39.

Figure 40. Equation. Total stopping distance for brake reaction time plus braking distance. d equals the following sum: 2.5 multiplied by V, that product plus the quotient of V squared divided by the product of 2 multiplied by g multiplied by f. Back to Figure 40.

Figure 41. Equation. AASHTO calculation of coefficient of friction for wet pavement. f equals a divided by g, which equals 11.2 feet per second squared divided by 32.2 feet per second squared, which equals 0.35. Back to Figure 41.

Figure 42. Bar graph. Participants’ visual acuity divided by age group. The graph shows the participants’ acuity by age group. The Y-axis shows the number of participants from 0 to 10. The X-axis shows visual acuity ranging from 20/13 to 20/40. Each age group of participants, young, middle, and older, is represented by a different-style bar. Young participants’ visual acuity ranged from 20/13 to 20/25, with half of them falling at 20/15 or better. Nine middle age participants had visual acuity of 20/20 and one participant in the middle age group had 20/25 visual acuity. Older age participants’ visual acuity ranged from 20/15 to 20/30. Two participants in the older group had a visual acuity of 20/15. Four participants in the older group had a visual acuity of 20/20. Three in the older group had 20/25 visual acuity, and one participant had a 20/30 visual acuity. Back to Figure 42.

Figure 43. Bar graph. Participants’ contrast sensitivity at 1.5 cpd (cycles per degree) divided by age group. The Y-axis shows the number of participants from 0 to 7. The X-axis is split in half, with percentage of contrast values for the left eye on one side and the percentage of contrast values for the right eye on the other. The percentage of contrast values listed on the X-axis range from 0.83 (Good) to 2.86 (Poor) for each eye. Each age group of participants, young, middle, and older, is represented by a different-style bar. Scores for participants in the young and middle age groups were similar in both the left and right eye tests, with a majority falling within 1.43 percent category, and fewer in the .83 and 2.86 percent categories, resembling a bell curve distribution. Older participants scored 1.43 and greater for left and right eye tests. Fifty and forty percent of older participants, in the left and right eye contrast tests, respectively, were in the 2.86 percent category. Back to Figure 43.

Figure 44. Bar graph. Participants’ contrast sensitivity at 3.0 cpd divided by age group. The Y-axis shows the number of participants from 0 to 7. The X-axis is split in half, with percentage of contrast values for the left eye on one side and the percentage of contrast values for the right eye on the other. The percentage of contrast values listed on the X-axis range from 0.45 (Good), to 2.27 (Poor). Each age group of participants, young, middle, and older, is represented by a different-styled bar. Scores for participants in the young age group were similar in left and right eye tests, resembling a bell curve distribution, having no score higher than 1.13 percent with a majority falling within .59 percent category. Results for the middle age group resemble an approximate positive linear distribution for both left and right eye tests, with very few falling in the .45 percent category and most falling in the 1.43 percent category. In the older group most participants fell in the .59 and 1.13 percent categories, with 10 and 20 percent of older participants falling within the 2.27 percent category, for left and right eye tests, respectively. Back to Figure 44.

Figure 45. Bar graph. Participants’ contrast sensitivity at 6.0 cpd divided by age group. The Y-axis shows the number of participants from 0 to 7. The X-axis is split in half, with percentage of contrast values for the left eye on one side and the percentage of contrast values for the right eye on the other. The percentage of contrast values listed on the X-axis range from 0.38 (Good), to 4.76 (Poor). Each age group of participants, young, middle, and older, is represented by a different-styled bar. The distribution of participants in the percentage of contrast values differed between the left and the right eye for the three age groups. For the left eye, all but one participant in the young age group fell between the 0.54 and 0.8 percent contrast categories, with one falling into the 4.76 percent category. The majority of participants of the middle age group fell between the 0.54 and 0.8 percent categories for the left eye test, with one falling in the 0.38 percent category and one falling into the 2.2 percent category. The older age group participants fell between the 0.54 and 2.2 percent categories in left eye tests, with no apparent trend. For the right eye, the participants in the young age group were dispersed across all the percent contrast categories, resembling a bell curve distribution skewed to the left. Participants in the middle age group fell between the 0.54 and 1.43 percent categories for right eye tests not following any apparent trend. Although no older age participants fell in the .38 percent category for right eye tests, participants in the older age group were dispersed across the remaining contrast percentage categories, resulting in a resemblance to a bell curve distribution skewed to the left.
Back to Figure 45.

Figure 46. Bar graph. Participants’ contrast sensitivity at 12.0 cpd divided by age group. The Y-axis shows the number of participants from 0 to 7. The X-axis is split in half, with percentage of contrast values for the left eye on one side and the percentage of contrast values for the right eye on the other. The percentage of contrast values listed on the X-axis range from 0.59 (Good), to 12.5 (Poor). Each age group of participants, young, middle, and older, is represented by a different-styled bar. For left eye tests, a majority of young participants fell between the 0.8 to 3.13 percent categories, with one falling into the 12.5 percent category. Middle aged participants fell between the 0.8 to 1.82 categories for left eye tests, with one falling in the 12.5 percent category. The majority of older participants fell between 1.14 and 6.67 percent category in left eye tests, with one falling into the 0.59 category. For the right eye, all young and middle age participants fell between the 0.8 to 6.67 percent categories, with no apparent trends. Older aged participants fell between the 1.14 to 6.67 percent categories for right eye tests, again with no apparent trends. Back to Figure 46.

Figure 47. Bar graph. Participants’ contrast sensitivity at 18.0 cpd divided by age group. The Y-axis shows the number of participants from 0 to 7. The X-axis is split in half, with percentage of contrast values for the left eye on one side and the percentage of contrast values for the right eye on the other. The percentage of contrast values listed on the X-axis range from 1.11 (Good), to 25 (Poor). Each age group of participants, young, middle, and older, is represented by a different-styled bar. For left eye tests, a majority of young participants fell between the 1.54 to 6.67 percent categories, resembling a bell curve distribution, with one falling into the 14.25 percent category. Middle age participants fell between the 2.5 to 6.67 percent categories in left eye tests, with one falling into the 25 percent category. Older aged participants fell between the 1.54 to 25 percent categories in left eye tests, resembling a bell curve distribution. For the right eye, most young participants fell between the 2.5 to 3.85 percent categories, with one each falling into the 1.11, 10 and 25 percent categories. Middle age participants fell between the 2.5 to 10 percent categories in right eye tests. Older aged participants fell between the 2.5 to 14.29 percent categories for right eye tests, with a majority in the higher percent categories. Back to Figure 47.

Figure 48. Photo. Smart Road testing facility. The photograph shows a paved road with a large turnaround at one end extending down into a valley at the other end. Back to Figure 48.

Appendix C: Contrast Sensitivity Test Diagram. The diagram shows the form the experimenters used to document the results for each participant’s contrast sensitivity exam. The diagram shows a graph with contrast sensitivity from 3 to 300 on the left Y-axis, contrast threshold from .0 to .003 on the right Y-axis, and spatial frequency (cycles per degree) from .5 to 6 on the X-axis. There is a column of circles above each cycle per degree. The experimenter documents the participant’s response by filling in the appropriate circle. There are two diagrams, one for recording the results of the right eye and one for the results of the left eye.
Back to Diagram.

• Schedule
• Training
  • Driver’s License Verification
  • Informed Consent
  • Forms and Questionnaires
  • Vision Tests
  • Laboratory Training
  • In-vehicle Familiarization
• Night 1 and 2
  • On Road Study

• What is enhanced night visibility?
• Why is your help important?
• Vehicles:
  • Car
  • Pick-up
  • SUV
• Scenario:
  • Smart Road test facility
  • Nighttime
  • Weather: Clear, Rain, Snow, or Fog

• Lab Training
• This training will help orient you to:
  • the Thermal Imaging System
  • the definition of terms we will use
  • the procedures
  • the objects
  • what we will ask from you

The Smart Road
For this research effort, you will be driving on the Smart Road test facility.
The Smart Road will be closed off to all traffic other than research vehicles. As a result, there will be at most two vehicles moving on the road, including the one you are driving.

Appendix E. Training Slide 6. The slide shows an aerial photo of a portion of the Smart Road. Back to Slide 6.

• Experimental Vehicles
• Vision Enhancement Systems
  • The Night Vision System
  • Prototype Headlamps

• Detection and Recognition
• Your primary task is to drive safely
  • Training; 15 mph in gravel lot, 25 mph on paved road
  • Night 1&2; 10 mph on Smart Road
• Your job will be to detect and recognize different objects on the Smart Road
• You will be required to press a button when you both detect and recognize objects

• Detection of Objects
• Detection is when you can just tell that something is on the road in front of you.
  • Detection is important while driving in that it prepares you to possibly make an evasive action
• When you detect an object, push the button as soon as you know something is in the road.

• Recognition of Objects
• Recognition is when you can say for sure what the object is.
  • This provides you with more information so you can adequately react to the object
• When you can recognize the object, you must push the button and, at the same time, recognize the object to the experimenter by saying, “I see a _____.”
• In case of an Unsuccessful Recognition press the push button again as soon as you notice what the right object is and tell the experimenter.

• Types of Objects
• Dynamic Objects
  • Pedestrians: People will be walking along side or across the road.
  • Cyclists: People will be riding bicycles across the road.

• Types of Objects
• Static Objects
  • Bicycle: A children’s bicycle will be laying on the right side of the road.
  • Tire Tread: A vehicle tire tread will be laying on the right side of the road.
  • Static pedestrian: A pedestrian will be standing still on the right side of the road.

Appendix E. Training Slide 13. The slide is titled “Dynamic Objects” and contains two separate photographs. One photograph is of a person wearing white clothing riding a bicycle. The other photograph is of a person wearing black clothing riding a bicycle. A single caption of “Bicyclists” is located under the photographs. Back to Slide 13.

Appendix E. Training Slide 14. The slide is titled “Dynamic Objects” and contains two separate photographs. One photograph is the side view of a person walking and wearing black clothing. The other photograph is of a person walking and wearing white clothing. A single caption of “Walking Pedestrians” is located under the photographs. Back to Slide 14.

Appendix E. Training Slide 15. The slide is titled “Static Objects” and contains two separate photographs. One photograph is of a child’s bicycle propped upright against a plain background. The bicycle has a yellow frame, pink tire rims, and white seat, tires, and pedals. The other photograph is of an intact circular tire tread with the sidewall removed. The tire tread is lying on its side against a plain background.
Back to Slide 15.

• Questionnaires
• You will be asked to respond to a questionnaire after each VES
  • Headlamp configuration questionnaire: You will provide a numbered rating of each headlight on a scale from 1 to 7.
  • Show questionnaires and different rating scales.

What we need from you

Driving is the primary task, so use safe driving practices

Maintain the specified speed limit

Immediately push the button when you Detect and/or Recognize an object

Verbally identify all objects as you press the button for the Recognition portion

Respond to the questionnaires

Ask questions whenever you need to

Appendix E. Training Slide 18.

Questions?

Back to Slide 18.

Appendix J. Diagram of hotspot location 1 for headlamp alignment. The diagram depicts a set of crosshairs with a circle centered in the lower right-hand corner. The caption above the diagram says “Hotspot Location: The circle represents the target hotspot location with respect to the target crosshairs. The center of the circle is the center of the hotspot.” Return to Diagram.

Appendix J. Diagram of hotspot location 2 for headlamp alignment. The diagram depicts a set of crosshairs centered on a large circle. Two smaller circles are centered horizontally inside the large circle. Two smaller circles are also centered vertically inside the larger circle, but portions of them are overlapped by the horizontal circles. The caption above the diagram says “Hotspot Location: The large outer circle represents the overall target area. The center of the large circle is the target hotspot location.” Return to Diagram.

Appendix K. Diagram. Protocol for running rain. The diagram shows the entire course of the Smart Road driven by participants during the study. The road is depicted as being fairly straight with one gradual curve up towards the top. There is a turnaround loop at each end of the road. The top turnaround is in the upper left corner of the diagram and the bottom turnaround is in the lower right corner, indicating that there is a grade. There are eight cones along the side of the road indicating the eight locations where participants would park or stop the test vehicle. Cones A, B, C, and D are located before the top turnaround. Cone A is located on one side of the road, and cones B, C, and D are located on the opposite side of the road, indicating that a driver would see the cones placed there after the drive from the bottom to the top of the road. Cones E, F, G, and H are located in the bottom turnaround. Cones F and G are located on one side of the road, and cones E and H are located on the opposite side of the road indicating that a driver would see the cones placed there after the drive from the top to the bottom of the road. Return to Diagram.