Limitations in drivers' ability to recognize pedestrians at night.
Pedestrian fatalities are a major road safety issue. In the United States, 70,000 pedestrians were injured and 4747 pedestrians were killed in the year 2003; pedestrian fatalities accounted for 11% of all traffic fatalities in that year (National Highway Traffic Safety Administration [NHTSA], 2005). Fatal pedestrian collisions are overrepresented at night, when nearly two thirds (65%) of all fatal pedestrian collisions occur (NHTSA, 2005). One of the main factors responsible for these collisions is drivers' inability to detect pedestrians from a safe distance (Leibowitz, Owens, & Tyrrell, 1998; NHTSA, 2004; Rumar, 1990). The critical importance of pedestrian visibility is further supported by systematic analyses of the U.S. Fatality Analysis Reporting System database, which indicate that pedestrian fatalities increase as illumination decreases even when other factors are held constant (e.g., Owens & Sivak, 1996; Sullivan & Flannagan, 2002).
The conspicuity of pedestrians at night can be enhanced by visibility aids such as street lighting and improved headlighting systems, as well as conspicuity treatments such as reflectors, reflective clothing, and lights. Night vision systems have also been advocated as a potential solution, although their benefits have yet to be demonstrated (e.g., Gish, Shoulson, & Perel, 2002). Retroreflective markings, which are engineered to reflect light back in the direction of its source, have been incorporated in vests and jackets and are widely used by construction workers, police, and emergency personnel, who have greater nighttime exposure. Retroreflective material positioned on the movable joints to create the sense of "biological motion" has also been advocated as a means of enhancing the visibility and recognition of pedestrians (Blomberg, Hale, & Preusser, 1986; Luoma, Schumann, & Traube, 1996; Owens, Antonoff, & Francis, 1994).
A number of studies have sought to determine how the use of such conspicuity aids might enhance pedestrian safety. Many of these studies have been undertaken in the laboratory, using either slide- or video-based representations of pedestrians wearing different clothing combinations (e.g., Owens et al., 1994). Although such studies provide good baseline data on which to design field studies, their validity is unknown because they do not necessarily represent the lighting or environmental conditions, vehicle or pedestrian motion, or driver workload present in real-world driving tasks. Several studies have been undertaken under either closed or open road situations but have relied on the participants acting as passengers instead of drivers (Allen, Hazlett, Tacker, & Graham, 1970; Luoma et al., 1996). Other studies that did use participants as drivers on closed road circuits relied on distance measurement techniques with potential limitations--for example, techniques that are contaminated by the inclusion of participant and experimenter reaction times and assumptions about vehicle speed (Chrysler, Danielson, & Kirby, 1996). Other studies on pedestrian visibility have used participants recruited from members of universities, who are not necessarily representative of the general driving population (e.g., Blomberg et al., 1986).
The goal of the present study was to quantify the ability of young and older drivers to recognize the presence of pedestrians walking in place on the shoulder of a real road at night, in the presence and absence of glare. The effects of different pedestrian clothing configurations and headlight beams were also investigated. The drivers' response distances were recorded using a novel parallax-based measurement technique that provided accurate measurements without interfering with the flow of the driving task.
Volunteers were recruited in a variety of ways, including attendance at various presentations by the research team, a listing of patients at a local optometry clinic, membership in a local church group, and graduate students. None of the participants was familiar with the hypotheses under investigation. The total of 20 participants consisted of 10 younger drivers (mean age 27.8 [+ or -] 4.7 years, range 21-34 years; 3 men and 7 women), and 10 older drivers (mean age 67.9 [+ or -] 5.30 years, range 60-75 years; 7 men and 3 women). All participants were licensed drivers and all reported that they drove regularly. All participants passed the minimum drivers' licensing criteria for corrected binocular visual acuity of 6/12 (20/40). Participants wore the optical correction that they normally wear while driving, if any. The study was conducted in accordance with the requirements of the Queensland University of Technology Human Research Ethics Committee. All participants were given a full explanation of the experimental procedures, and written informed consent was obtained, with the option to withdraw from the study at any time.
A confidential questionnaire was administered to obtain an overall sense of the driving habits of the two participant groups as well as their perceptions of driving, particularly at nighttime. Only findings relevant to general driving are reported here. The older participants had 30 to 58 years of driving experience (M = 45.8 years), and the younger participants had 1.5 to 17 years of driving experience (M = 7.9 years). When questioned about nighttime driving experiences over the previous year, the older participants reported that a smaller percentage of their total driving was at nighttime, as compared with the younger participants (M = 12.7% vs. 34%, respectively), t(18) = 3.2, p = .005. On a 5-point scale from very comfortable to very uncomfortable, the older drivers also rated nighttime driving in good weather as less comfortable than did their younger counterparts, t(18) = -2.3, p = .035. Despite this, on a 3-point scale from always avoid to never avoid, the older participants did not report avoiding nighttime driving any more than the younger participants did, for either heavy traffic, t(17) = -1.7, p = ns, or rainy conditions, t(17) = -1.3, p = ns.
Experimental Vehicle and Closed Road Circuit
The experiment was conducted under nighttime conditions on the closed road circuit at the Mt Cotton Driver Training Centre, which has been used in previous studies of driving and vision (e.g., Wood & Troutbeck, 1994). The experiment was undertaken only on nights when there was no active precipitation and the road surface was dry. The circuit, which is representative of a rural road, consists of a two- to three-lane bitumen road surface and includes hills, curves, bends, and straight sections as well as standard road signs and road markings. A 1.8-km (1.1-mile) section of the circuit was used for this study. The circuit does not include any street lighting. Three sets of headlamps, each consisting of a pair of stationary battery-powered headlights mounted at the correct height and separation, were positioned at three locations along the circuit to simulate the effect of an oncoming sedan vehicle (Figure 1). These headlights were activated when the test vehicle drove through a series of remote sensors. A series of 50 traffic cones with retroreflective markers and three 1.1-m high retroreflective poles were also distributed along the circuit to provide a degree of visual clutter and to act as distractors.
[FIGURE 1 OMITTED]
The experimental vehicle was an instrumented 1997 Nissan Maxima that had been serviced (including headlamp alignment) immediately prior to the experiment. Two digital video cameras were mounted a fixed distance apart on the roof of the vehicle. This system recorded two overlapping images of the forward road scene and was linked to a fiberoptic marking system and touch pad, which recorded the exact moment that the participant recognized a pedestrian, thus identifying the relevant pairs of images. These images were analyzed off line to determine the positions of a corresponding point in the marked images. The change in position of these points is the parallax; this was measured and the recognition distance was thus calculated. In addition to having the advantage of not interfering with the flow of the driving task, the technique has a high level of accuracy and validity (Jones, Bentley, Wood, & Woolf, 1998; Jones, Wood, Wolf, & Bentley, in press). To minimize measurement error further, traffic cones with retroreflective markers were positioned strategically along the circuit. Thus when a participant driver pressed the touch pad to indicate that a pedestrian was recognized, the measurement system was required only to measure the distance from the test vehicle to the nearest traffic cone (always less than 58 m). The distance from that cone to the pedestrian was known in advance and was added to the measurement. The speed of travel at the time that the driver pressed the touch pad was also recorded.
Four experimenters were involved in each experimental session. Two experimenters were seated in the experimental vehicle (one in the passenger seat and one in the back), and two experimenters acted as pedestrians located at different positions on the circuit (see Figure 1). Both pedestrians walked in place on one shoulder of the roadway; the test vehicle was driven toward the pedestrians in the far lane. One pedestrian was situated at the end of a 400-m (1312-foot) straight section of three-lane roadway that started and ended at approximately the same elevation hut which included a dip roughly in the middle. This pedestrian was referred to as the primary pedestrian because the drivers encountered this pedestrian first. The maximum distance at which the primary pedestrian could be seen in daylight conditions (i.e., sight distance) was 413 m (1556 feet). The driver had been exposed to one of the sets of oncoming headlights, at a distance of 431.5 m (1416 feet) prior to encountering this pedestrian (Figure 1), and was therefore not completely dark adapted. The secondary pedestrian was positioned on the opposite portion of the circuit, on the far shoulder of a two-lane roadway. A pair of headlights was positioned 10.2 m (33.5 feet) in front of the secondary pedestrian: To the driver, these headlights simulated oncoming traffic and were a direct source of glare. Each pedestrian had a two-way radio, as did the experimenter, who was seated in the back of the vehicle. All communication was conducted between laps with the experimenter outside of the vehicle, so the participant could not hear the conversations.
Clothing and Headlight Beam Conditions
For each lap the pedestrians wore one of four clothing conditions. The black condition was a black cotton sweatshirt (2% reflectance), a pair of black cotton sweatpants, black gloves, and black shoe covers. The white condition was a large white cotton lab coat (68% reflectance), white gloves, white cotton leggings, and white shoe covers. The vest condition was the clothing from the black condition plus a white retroreflective (diamond grade; 3M) panel measuring 30 x 17.5 cm (525 [cm.sup.2]) worn on the chest. The biomotion condition was the clothing from the black condition with the addition of white retroreflective (diamond grade) straps (2.5 cm; 1 inch) around the wrists, elbows, shoulders, waist, knees, and ankles. The total area of visible retroreflective material was matched to the vest condition (525 [cm.sup.2]), and the same type of retroreflective material was used.
Each pedestrian wore each of the four clothing conditions twice, once for a low-beam lap and once for a high-beam lap. These eight combinations were spread across nine laps. The order of the two headlamp beam conditions was randomized, as was the order in which each of the two pedestrians wore the four clothing configurations within each beam condition. To reduce the drivers' expectancies, the primary pedestrian was always absent on Lap 3 and the secondary pedestrian was always absent on Lap 7. Pedestrians walked in place as the test vehicle approached; this allowed the inclusion of natural pedestrian motion while maintaining the pedestrians at a known location.
Each participant completed 10 laps of the test circuit. The primary purpose of each driver's first lap was to familiarize the driver with the vehicle and the circuit. The secondary purpose of this lap was to measure the drivers' reaction time (described later). All pedestrian recognition data were collected on Laps 2 through 10 (with the primary pedestrian always being absent on Lap 3). At the start of each of these laps the drivers were instructed to follow the prescribed route, to drive at a comfortable speed, and to press a large (6 x 12 cm) luminous dash-mounted touch pad (and to announce "pedestrian!") as quickly as possible each time they recognized that a pedestrian was present. They were instructed not to press the button until they were confident that what they saw was in fact a pedestrian. They were also informed that there would not always be pedestrians on the circuit. To increase driver workload, we also instructed participants to read aloud all road signs that they encountered, although performance on this task was not recorded.
Two primary dependent variables are described here, the first of which is the percentage of trials in which the driver correctly recognized the presence of the pedestrians. Pedestrian recognition was recorded as having occurred if the driver pressed the response button at any point along the approach to a pedestrian or immediately after having passed a pedestrian. This procedure is conservative in that mere recognition does not imply that the driver would have been able to initiate a successful avoidance maneuver. Recognition was measured for both the primary pedestrian (no glare present) and for the secondary pedestrian (glare present). The second dependent variable is the driver's response distance to the primary pedestrian, defined as the distance from the vehicle to the primary pedestrian (no glare) at the moment the touch pad was pressed. (Response distances were not recorded for the secondary pedestrian because of the limited sight distance.) Response distances were coded as zero for all trials in which the driver did not respond to the primary pedestrian or had passed the primary pedestrian before pressing the touch pad.
Reaction times were recorded during each driver's first lap. This was accomplished by instructing the drivers that at some point during their first lap that they would see a large bright red light appear somewhere on the left or right shoulder of the road and that when they saw this light they should press the touch pad as quickly as possible. The light was a battery-powered array of red LEDs (11.5 x 19.5 cm) that was positioned at a set point on the left shoulder of the straight section of the road. An experimenter, who was hidden from view, activated the LEDs when she could see both of the vehicle's headlights.
The reaction times recorded on the first lap were of a similar magnitude for the older (M = 1.54 s) and younger (M = 1.59 s) participants, t(18) = -0.22, ns. Similarly, the average speed recorded at the moment that the driver pressed the touch pad to indicate that he or she recognized the presence of a pedestrian was not significantly different between the younger (M = 70.7 km/ hr) and older (M = 67.7 km/hr) participants, t(63) = 1.7, ns. However, the mean time to complete a lap was significantly longer for the older (M = 147.6 s) than for the younger (M = 137.5 s) participants, t(18) = -2.5, p = .03.
Tables 1 and 2 present the percentage of drivers who correctly recognized the presence of the pedestrians as a function of pedestrian clothing, beam condition, and driver age. Only one false alarm occurred during testing. This occurred during one of the older drivers' low-beam laps when the driver passed one of the low-contrast wooden poles that held the trigger for the glare lights. Table 1 presents the data for the recognition of the primary pedestrian (no glare present) and Table 2 presents the data for the recognition of the secondary pedestrian (glare present). Overall, drivers correctly responded to only 61% of pedestrians when glare was present and to 76% in the absence of glare. Collapsed across the glare, clothing, and beam variables, older drivers responded to only 53.2% of the pedestrians, whereas the younger drivers responded to 84.4%.
The pedestrians' clothing also affected performance: Collapsed across the glare, age, and beam variables, recognition was 33.8% for the black condition, 63.8% for the vest condition, 83.8% for the white condition, and 93.8% for the biomotion condition. The drivers' beam setting also influenced performance: Overall, drivers responded to 63.2% of the pedestrians with low-beam illumination and to 74.4% of the pedestrians with high-beam illumination. Interestingly, the ability of drivers to recognize the pedestrians wearing white clothing was better than their ability to recognize the pedestrian wearing the retroreflective vest. When the same amount of retroreflective material was worn in the biomotion configuration, 100% of the young drivers and at least 70% of the older drivers correctly responded to the pedestrians, even when the glare challenge was present. In sum, the ability of drivers to recognize the presence of pedestrians varied from 5% (black clothing, low-beam illumination, and glare) to 100% (biomotion, low- or high-beam illumination, and no glare).
The percentage of pedestrians recognized was modeled as a function of pedestrian clothing, headlamp beam, presence or absence of glare, and driver age using a logistic regression model. The regression indicated that pedestrian clothing, [chi square](3) = 64.5, headlamp beam, [chi square](1) = 7.9, glare, [chi square](1) = 13.3, and driver age, [chi square](1) = 41.6, were all significantly (p < .005) associated with correctly identifying the presence of the pedestrian.
The response distances for the recognition of the primary pedestrian are represented in Table 3 as a function of driver age, pedestrian clothing, and beam condition. The camera system failed to record the response distance for 1 older driver's high-beam lap when the primary pedestrian was wearing white clothing. The driver did respond to the pedestrian for that lap. The missing data point was replaced by the mean response distance for that age group for that condition (76.1 m); this step did not alter the pattern of results in the analyses that follow. Averaged across all no-glare conditions, drivers responded to pedestrians at a mean distance of 76.5 m (251 feet), but these values ranged from 0.0 m (0.0 feet; older drivers, black clothing, low beam) to 220.0 m (721.8 feet; younger drivers, biomotion, high beam). Overall, black clothing resulted in the shortest response distances and biomotion resulted in the longest; collapsed across age and beam, drivers responded to black-clad pedestrians at only 12.8 m (42.0 feet) and to pedestrians marked with biomotion at 165.1 m (541.7 feet). Collapsed across age and beam, drivers responded to pedestrians wearing a retro-reflective vest at distances (55.5 m or 182.1 feet) similar to those when pedestrians wore white clothing (72.8 m or 238.9 feet). Overall, older drivers responded to pedestrians at 56.0 m (183.7 feet) and younger drivers responded to pedestrians at 97.0 m (318.2 feet). Changing from low-beam to high-beam illumination increased overall response distances from 59.4 to 93.6 m (194.9-307 feet).
An analysis of variance (ANOVA) with two within-subject factors (clothing and headlamp beam) and one between-subjects factor (driver age) demonstrated that the main effects of pedestrian clothing, F(3, 54) = 49.29, p < .001, and headlamp setting, F(1, 18) = 11.07, p = .004, were both significant. The main effect of driver age was also significant, indicating that overall, older drivers had shorter response distances than did younger drivers, F(1, 18) = 10.54, p = .004. None of the interactions was significant. (Because none of the older drivers responded to the primary pedestrian wearing black clothing in the low-beam condition, that cell had no variability and the ANOVA assumption of equal variances was clearly violated. We repeated the ANOVA twice, once without the black clothing condition and once without the older drivers' data. In both cases the pattern of results matched that of the original.) Model-based contrast analysis indicated that the differences in response distances for the different clothing conditions were all significant (p < .05), except for the comparison between the white clothing and vest conditions. Calculation of the effect sizes ([[omega].sup.2]) indicated that 41.8% of the variance in response distances was accounted for by clothing, 5.6% by beam setting, and 8.1% by driver age.
It is important to keep in mind that the response distances reported here were measured at the moment the driver pressed the touch pad and not necessarily when pedestrian recognition first occurred. The reaction time data recorded in the first lap, combined with the knowledge of the speed that the driver was traveling when he or she responded in each lap, can be used to estimate the distance at which recognition first occurred. These calculations, which are not presented here, did not result in changes to the overall patterns revealed by the ANOVA. Averaged across the participants in each age group, the mean distance traveled during the mean reaction time was 31.2 m (102.4 feet) for the younger drivers and 29.0 m (95.1 feet) for the older drivers. Adding these values to the response distances presented in Table 3 provides a rough estimate of the recognition distances.
The present experiment sought to quantify the ability of drivers to recognize the presence of roadside pedestrians at night. Young and older drivers drove a test vehicle around a closed road circuit and pressed a large dash-mounted touch pad when they first recognized that a pedestrian (positioned on the far shoulder of a two- or three-lane roadway) was present. A novel parallax-based system was used to measure response distances. The data strongly demonstrate that driver age, pedestrian clothing, and headlamp beam setting all significantly affect both the probability that a driver will recognize the presence of a roadside pedestrian and the distance at which drivers respond to pedestrians under realistic nighttime driving conditions.
The ability of older drivers to respond to the presence of pedestrians was consistently worse than that of the younger drivers. Whereas the young drivers recognized 84% of the pedestrians overall (94% in the absence of glare and 75% in the presence of glare), older drivers recognized only 53% of the pedestrians (59% with no glare and 48% with glare). The older drivers also identified the pedestrians at significantly shorter distances than did the younger drivers. Overall, the mean recognition distance for the older drivers was only 58% of that of the younger drivers. Depending on the clothing and beam conditions, older drivers' mean response distances ranged from 0% to 76% of those corresponding means of the younger drivers. These results are in general agreement with those of Chrysler et al. (1996), who reported that pedestrian detection distances (for a child-sized mannequin) were up to 75% shorter for older drivers than for younger drivers for some visibility conditions. Interestingly, Luoma et al. (1996) found smaller age-related decreases in pedestrian recognition distances, which may relate to the fact that their participants acted as passengers (seated in either the passenger seat or the back of the vehicle), not drivers. The participants in our study drove under realistic conditions while performing a sign detection task.
The age-related decline in the ability to respond to the presence of pedestrians is unlikely to be attributable to differences in driver reaction time, given that the reaction times measured in the first lap for each driver were not significantly different between age groups. The age effect also seems unlikely to be attributable to age-related differences in the drivers' response criterion, given that the only false alarm was from an older driver. Instead, age-related changes in visual function are likely to be responsible for the age effects reported here. Of particular relevance are age-related changes reported in contrast sensitivity (e.g., Elliott, Whitaker, & MacVeigh, 1990; Higgins, Jaffe, Caruso, & deMonasterio, 1988; Scialfa, Garvey, Tyrrell, Goebel, & Leibowitz, 1992), static (Frisen & Frisen, 1981; Wood & Bullimore, 1995) and dynamic visual acuity (Long & Crambert, 1990), glare sensitivity (Haegerstrom-Portnoy, Schneck, & Brabyn, 1999), motion sensitivity (Wood & Bullimore, 1995), and the useful field of view (Ball, Beard, Roenker, Miller, & Griggs, 1988).
Pedestrian clothing had a dramatic effect on the recognition of pedestrians by drivers. The drivers did not recognize the presence of most of the black-clad pedestrians. When the pedestrians wore black clothing, they were recognized on only 52.5% of the younger drivers' laps and on only 15% of the older drivers' laps. Further, when recognition did occur, it was at distances that were unlikely to be sufficient to allow drivers to avoid a collision. These results are in general agreement with those of previous researchers, who have found recognition distances of the order of 50 m or less for real pedestrians wearing black when the participants acted as passengers (Allen et al., 1970; Luoma et al., 1996) and drivers (Blomberg et al., 1986). Thus, even when drivers are alerted to the possibility of encountering pedestrians, they are unable to achieve a safe level of performance in recognizing the presence of low-contrast pedestrians. Drivers who do not expect to encounter a roadside pedestrian are likely to show even worse performance (Roper & Howard, 1938).
The drivers' ability to recognize pedestrians was greatly improved when pedestrian contrast was increased. When pedestrians wore white clothing, they were recognized on 97.5% of the younger drivers' laps and on 70% of the older drivers' laps, and response distances increased by as much as a factor of 25x (for the older drivers for the high-beam condition). The addition of retroreflective materials provided some unexpected results. Merely adding a white retroreflective vest on top of the black clothing did not enhance recognition as much as might be anticipated. Indeed, it was only when the same amount of retroreflective material was distributed in order to convey a perception of biological motion (with the material attached to the movable joints) that the effect of retroreflective material exceeded that of wearing white cotton clothing.
Such results are in general agreement with those of Owens, Wood, Whittam, and Woolf (in press), who showed that wearing retroreflective material in the biomotion configuration, rather than as a diagonal slash of retroreflective material on the chest, had significant advantages in terms of improving older drivers' ability to recognize pedestrians (although these improvements were not quantified in terms of recognition distances). In the present study, when the pedestrians wore the biomotion configuration in the absence of glare, both the young and the older drivers recognized them on 100% of the laps. The biomotion advantage was also relatively robust to glare: Even when standing behind a glare source, the secondary pedestrian wearing biomotion was recognized by 100% of the young drivers and by 75% of the older drivers. Response distances for the biomotion condition were also improved by up to a factor of 50x when compared with the black clothing condition.
These results support those of earlier demonstrations of the utility of biomotion (Blomberg et al., 1986; Luoma et al., 1996; Owens et al., 1994), but they suggest that the benefits of a biomotion configuration are even larger than was thought. Further, Luoma et al. (1996) showed that the benefits of retroreflective markers on the major joints had even greater effects (47% longer average distances) when the pedestrians were crossing the road rather than approaching the driver, as was the case in the present study. The benefits of wearing such materials may thus be even greater than those demonstrated here.
Headlamp beam also had a significant effect on pedestrian recognition. Overall, changing from low beams to high beams improved recognition distances from a mean of 59.4 m (194.9 feet) to 95.6 m (507.1 feet), with recognition distances increasing most dramatically (by a factor of 3.5x) for the black-clad pedestrian. The finding that recognition distances are longer under high-beam than under low-beam conditions is in general accord with the findings of Mortimer and Olson (1974) and also of Shinar (1984), whose participants acted as passengers driven along a rural road. However, the latter author also found that the effects of beam were significant only for the darkly clad pedestrians and not when the pedestrians wore a retroreflective tag (Shinar, 1984). In the present data, high beams increased response distances by a factor of 2.7x when no retroreflective material was present (the black and white conditions) and by a factor of 1.3x when retroreflective material was present (the vest and biomotion conditions).
Perhaps the most important finding of this study was the difficulty that drivers of all ages had in responding to the presence of roadside pedestrians. Overall, only 76% of the pedestrians were recognized in the absence of glare and only 61% were recognized when glare was present. Even more dramatically, when pedestrians wore dark clothing and were illuminated by low-beam headlamps (conditions that are, perhaps, most typical), pedestrians were seen on only 40% of the trials without glare and 5% of the trials with glare. Further, when recognition did occur in these conditions, it was at a mean distance of 5.6 m (18.4 feet) without glare. Thus even alerted drivers frequently fail to see pedestrians walking along the roadside, and when they do see the pedestrian it may be too late to initiate a successful avoidance maneuver (American Association of State Highway and Transport Officials, 1984; Leibowitz et al., 1998). These results are sufficiently striking to raise serious concerns for the safety of pedestrians at night and to explore methods of making pedestrians more conspicuous to drivers. Further, drivers need to be educated regarding their limited ability to recognize pedestrians at night. Drivers should also be encouraged to use their high beams when not facing oncoming traffic.
The pedestrian visibility problem is compounded by the tendency of pedestrians to overestimate their own visibility at night (Allen et al., 1970; Shinar, 1984). Pedestrians who mistakenly believe that they are easily recognized by approaching drivers may be more likely to behave in ways that increase their danger (e.g., crossing a road inappropriately or jogging along the shoulder of the road instead of on a sidewalk). Indeed, data collected in a recent experiment that parallels the one reported here quantified pedestrians' estimates of their own visibility using the same driving circuit and protocols as well as the same test vehicle, beam settings, and clothing configurations (Tyrrell, Wood, & Carberry, 2004). Comparison of the two data sets provides a unique opportunity to compare actual and estimated visibility distances and confirms that pedestrians dramatically overestimate their own conspicuity, particularly when wearing the black, white, and vest clothing configurations (Tyrrell, Wood, et al., 2004).
Given that pedestrians appear to appreciate neither the magnitude of the nighttime visibility problem nor the value of retroreflective treatments, one approach that could enhance pedestrian safety would be to educate pedestrians. In this regard, Tyrrell, Patton, and Brooks (2004) recently used an educational approach on both university students and high school driver education students. In these experiments, students' overestimates of their own nighttime visibility were significantly reduced and they gained an appreciation for the safety benefits of reflective clothing and retroreflective treatments (Tyrrell, Patton, et al., 2004). This finding suggests that informing pedestrians of specific issues regarding night visibility (e.g., contrast, retroreflection, biological motion) and of specific behaviors that can reduce the danger (e.g., avoiding interacting with nighttime traffic, using retroreflective markers, biological motion) may be more productive than simply offering vague admonitions to be careful.
This research was conducted while the second author was on sabbatical leave at Queensland University of Technology. The authors would like to express appreciation to Queensland Transport for allowing the use of the facilities at the Mt Cotton Driver Training Centre and to the staff of the Mt Cotton Centre for their generous cooperation and support. The authors are grateful for the assistance of Johnell Brooks, Tabitha Faulks, Fred Owens, Justin Owens, Rachel Pickering, and Mark Woolf during data collection. The authors are especially grateful to Kevin Jones for his role in data interpretation. This study was supported by grants from the Australian Research Council, Queensland University of Technology and, Clemson University. These data were presented in part at the 81st Annual Meeting of the Transportation Research Board, Washington, D.C.
Allen, M. J., Hazlett, R. D., Tacker, H. L., & Graham, B. V. (1970). Actual pedestrian visibility and the pedestrian's estimate of his own visibility. American Journal of Optometry and Archives of the American Academy of Optometry, 47, 44-49.
American Association of State Highway and Transport Officials. (1984). A policy on geometric design of highways and streets. Washington, DC: Author.
Bali, K. K., Beard, B. L., Roenker, D. L., Miller, R. L., & Griggs, D. S. (1988). Age and visual search: Expanding the useful field of view. Journal of the Optical Society of America: A, 5, 2210-2219.
Blomberg, R. D., Hale, A., & Preusser, D. F. (1986). Experimental evaluation of alternative conspicuity-enhancement techniques for pedestrians and bicyclists. Journal of Safety Research, 17, 1-12.
Chrysler, S. T., Danielson, S. M., & Kirby, V. M. (1996). Age differences in visual abilities in nighttime driving field conditions. In Proceedings of the Human Factors and Ergonomics Society 40th Annual Meeting (pp. 923 927). Santa Monica, CA: Human Factors and Ergonomics Society.
Elliott, D. B., Whitaker, D., & MacVeigh, D. (1990). Neural contribution to spatio-temporal contrast sensitivity decline in healthy aging eyes. Vision Research, 30, 541-547.
Frisen, L., & Frisen, M. (1981). How good is normal visual acuity? A study of letter acuity thresholds as a function of age. Albrecht von Graefes Archly fur Klinische und Experimentelle Ophthalmologie, 215, 149-157.
Gish, K. W., Shoulson, M., & Perel, M. (2002, January). Driver behavior and performance using an infrared night vision enhancement system. Presented at the 80th Annual Meeting of the Transportation Research Board, Washington, D.C.
Haegerstrom-Portnoy, G., Schneck, M. E., & Brabyn, J. A. (1999). Seeing into old age: Vision function beyond acuity. Optometry and Vision Science, 76, 141-158.
Higgins, K. E., Jaffe, M. J., Caruso, R. C. & deMonasterio, F. M. (1988). Spatial contrast sensitivity: Effects of age, test-retest, and psychophysical method. Journal of the Optical Society of America, 5, 2173-2180.
Jones, K., Bentley, B. E., Wood, J. M., & Woolf, M. I. (1998). Application of parallax for the measurement of visibility distances in the open-road environment. International Archives of Photogrammetry and Remote Sensing, 32(5), 74-79.
Jones, K., Wood, J. M., Woolf, M. I., & Bentley, B. E. (in press). A technique for on-road assessment of road sign visibility distances. In A. Gale (Ed.), Vision in vehicles VII. (Publisher n.a.)
Leibowitz, H. W., Owens, D. A., & Tyrrell, R. A. (1998). The assured clear distance ahead rule: Implications for traffic safety and the law. Accident Analysis and Prevention, 30, 93-99.
Long, G. M., & Crambert, R. F. (1990). The nature and basis of age-related changes in dynamic visual acuity. Psychology and Aging, 5, 138-143.
Luoma, J., Schumann, J., & Traube, E. C. (1996). Effects of retro-reflector positioning on nighttime recognition of pedestrians. Accident Analysis and Prevention, 28, 377-383.
Mortimer, R. G., & Olson, P. L. (1974). Evaluation of meeting beams by field tests and computer simulation (Report No. UM-HSRIHF-74-27). Ann Arbor, MI: Highway Safety Research Institute.
National Highway Traffic Safety Administration. (2005). Traffic safety facts 2003: Pedestrians (DOT HS 809 769). Washington, DC: U.S. Department of Transportation. Retrieved October 20, 2005, from http://www-nrd.nhtsa.dot.gov/pdf/nrd-30/NCSA/ TSF2005/809769.pdf
Owens, D. A., Antonoff, R., & Francis. E. L. (1994). Biological motion and nighttime pedestrian conspicuity. Human Factors, 36, 718-732.
Owens, D. A., & Sivak, M. (1996). Differentiation of visibility and alcohol as contributors to twilight road fatalities. Human Factors, 38, 680-689.
Owens, D. A., Wood, J. M., Whittam, D., & Woolf, M. (in press). Effects of age and illumination on driving performance. In A. Gale (Ed.), Vision in vehicles VIII. (Publisher n.a.)
Roper, V. J., & Howard, E. A. (1938). Seeing with motorcar headlamps. Illumination Engineering, 33, 412-458.
Rumar, K. (1990). The basic driver error: Late detection. Ergonomics, 32, 1281-1290.
Scialfa, C. T., Garvey, P. M., Tyrrell, R. A., Goebel, C. C., & Leibowitz, H. W. (1992). Age differences in dynamic contrast thresholds. Journal of Gerontology: Psychological Sciences, 47, P172-175.
Shinar, D. (1984). Actual versus estimated nighttime pedestrian visibility. Ergonomics, 27, 863-871.
Sullivan, J. M., & Flannagan, M. I. (2002). The role of ambient light level in fatal crashes: Inferences from daylight saving time transitions. Accident Analysis and Prevention, 34, 487-498.
Tyrrell, R. A., Patton, C. W., & Brooks, J. O. (2004). The effects of educational interventions, clothing reflectance, and headlamp beam setting on pedestrians' estimates of their own nighttime visibility. Human Factors, 46, 170-182.
Tyrrell, R. A., Wood, J. M., & Carberry. T. (2004). On-road measures of pedestrians' estimates of their own nighttime conspicuity. Journal of Safety Research, 35, 483-490.
Wood, J. M., & Bullimore, M. A. (1995). Changes in the lower displacement limit for motion with age. Ophthalmic and Physiological Optics, 15, 31-36.
Wood, J. M., & Troutbeck, R. (1994). The effect of visual impairment on driving. Human Factors, 36, 476-487.
Joanne M. Wood is an associate professor at the Centre for Eye Research at Queensland University of Technology. She received her Ph.D. in visual psychophysics from the University of Aston in Birmingham, U.K., in 1987.
Richard A. Tyrrell is a professor of psychology at Clemson University. He received a Ph.D. in experimental psychology from Pennsylvania State University in 1993.
Trent P. Carberry is pursuing a Ph.D. in vision and driving from the Centre for Eye Research at Queensland University of Technology, where he received his B.Psych. in 2000.
Date received: March 17, 2003
Date accepted: July 6, 2004
Joanne M. Wood, Queensland University of Technology, Brisbane, Australia, Richard A. Tyrrell, Clemson University, Clemson, South Carolina, and Trent P. Carberry, Queensland University of Technology, Brisbane, Australia
Address correspondence to Joanne M. Wood, Center for Eye Research, Queensland University of Technology, Brisbane, Queensland, Australia; email@example.com. HUMAN FACTORS, Vol. 47, No. 3, Fall 2005, pp. 644-653.
TABLE 1: Percentage of Drivers Who Recognized the Presence of the Pedestrian in the Absence of Glare (Primary Pedestrian) Black White Vest Biomotion Mean Young Drivers Low beam 70 100 90 100 90.0 High beam 90 100 100 100 97.5 Mean 80 100 95 100 93.8 Older Drivers Low beam 10 70 30 100 52.5 High beam 20 90 50 100 65.0 Mean 15 80 40 100 58.8 All Drivers Low beam 40 85 60 100 71.3 High beam 55 95 75 100 81.3 Mean 47.5 90 67.5 100 76.3 TABLE 2: Percentage of Drivers Who Recognized the Presence of the Pedestrian in the Presence of Glare (Secondary Pedestrian) Black White Vest Biomotion Mean Young Drivers Low beam 0 90 80 100 67.5 High beam 50 100 80 100 82.5 Mean 25 95 80 100 75 Older Drivers Low beam 10 60 30 70 42.5 High beam 20 60 50 80 52.5 Mean 15 60 40 75 47.5 All Drivers Low beam 5 75 55 85 55 High beam 35 80 65 90 67.5 Mean 20 77.5 60 87.5 61.3 TABLE 3: Mean Distances in Meters (and Standard Deviations) at Which Drivers Responded to the Presence of the Pedestrian in the Absence of Glare (Primary Pedestrian) Black White Vest Young Drivers Low beam 11.3(11.0) 52.3 (11.1) 59.4 (107.5) High beam 37.0(20.7) 134.5 (65.1) 76.4 (73.8) Mean 24.1(20.8) 93.4 (62.0) 67.9 (90.2) Older Drivers Low beam 0 (0) 28.3 (22.9) 27.4 (50.8) High beam 2.8(7.5) 76.1 (37.6) 58.5 (102.0) Mean 1.4(5.2) 52.2 (37.6) 43.0 (80.0) All Drivers Low beam 5.6(9.5) 40.3 (21.4) 43.4 (83.5) High beam 19.9(23.2) 105.3 (58) 67.5 (87.1) Mean 12.8(18.9) 72.8 (55.2) 55.5 (87.1) Biomotion Mean Young Drivers Low beam 185.4 (91.5) 77.1 (94.9) High beam 220.0 (107.6) 117.0 (99.4) Mean 202.7 (98.8) 97.0 (98.6) Older Drivers Low beam 111.0 (32.4) 41.7 (52.3) High beam 143.9 (98.3) 70.3 (87.0) Mean 127.4 (73.2) 56.0 (72.8) All Drivers Low beam 148.2 (76.9) 59.4 (78.2) High beam 182.0 (107.7) 93.6 (95.7) Mean 165.1 (93.9) 76.5 (88.8)
|Printer friendly Cite/link Email Feedback|
|Author:||Wood, Joanne M.; Tyrrell, Richard A.; Carberry, Trent P.|
|Date:||Sep 22, 2005|
|Previous Article:||Attentive navigation for viewpoint control in virtual environments.|
|Next Article:||Team task analysis: identifying tasks and jobs that are team based.|