Geometric Design - USD
Highway Research to Enhance the
Safety and Mobility of Older Road Users
Draft Version 2.0
June 21, 2000
by
Frank Schieber
Heimstra Human Factors Laboratories
University of South Dakota
414 E. Clark Street
Vermillion, SD 57069
schieber@usd.edu
UNDER FINAL REVIEW
NOT TO BE QUOTED WITHOUT PERMISSION
Contents
1.0 Introduction
2.0 Highway Geometric Design
1. AASHTO Models
2. Gap Acceptance Models
3. Age-Appropriate Models for Highway Geometric Design
3.0 Traffic Operations
3.1 Intersections
3.2 Freeway Operations
3.3 Work Zones
4. Pedestrian Crossings
5. Traffic Calming and Roundabouts
6. Older Driver Highway Design Handbook
4.0 Traffic Control Devices
4.1 Roadway Delineation
4.1.1 Lane Markings
4.1.2 Hazard Delineation
4.1.3 Ultraviolet-Activated Treatments
2. Highway Signing
1. Sign Conspicuity
2. Sign Legibility
1. Letter Height
2. Font Characteristics
3. Minimum Brightness/Retroreflectivity Requirements
3. Symbol Signs
4. Traffic Signal Lights
5.0 Highway Lighting
6. Summary Recommendations
1. Recommendations for Immediate Implementation and Practice
2. Research Recommendations
7.0 References
1.0 Introduction
Published in 1988, the Transportation Research Board’s (TRB) Special Report 218 was a milestone in the history of transportation research. This two-volume report summarized our knowledge about the special needs of older persons, speculated about how those emerging needs might impact future mobility and safety, and served as a “blue print” for a decade of research sponsored primarily by the United States Federal Highway Administration (FHWA) and the National Highway and Traffic Safety Administration (NHTSA). TRB (1988) Special Report 218 placed an especially heavy emphasis upon research and development needs in the area of highway design and operations. One reason for this emphasis stemmed from the simple fact that the lag between discovery and implementation in this area is often so very long. Changes to concrete, signing and lighting, when necessary, often require a generation to implement in a cost-effective manner. Thus, if changes to our highway infrastructure can be made to better accommodate the wave of older travelers who will reach retirement age during the early decades of the third millennium, we need to know about them now. Otherwise, those needed changes will not have sufficient time to be implemented to meet our appointment with the demographic destiny of a graying traveling populace.
There are several general assumptions that guided this review of the literature of which the reader should be made aware: First, old was arbitrarily presumed to include those 65 years of age or older. Second, in many cases a critical judgment had to made concerning whether or not a particular highway design factor “accommodated the needs of older travelers” (usually drivers). The criterion used to make such judgments throughout the current review was the 85th percentile level of performance demonstrated by the designated “older” sample or group. For example, if 85 percent of the older drivers in a study could read a highway sign at a given letter height, then that letter height was judged to accommodate the needs of older drivers in general. The third assumption made by the reviewer was that the older volunteers in most of the studies were representative of the 80 percent of older persons who are living independently in the community and are capable of successfully performing the general suite of tasks typically referred to as the “activities of daily living” (Sterns and Gray, 1999). There is no doubt that some “volunteer bias” is present in most studies of older travelers. Given the assumption that better educated and healthier people are more likely to volunteer as participants in research projects, logic would dictate that our samples tend to be slightly biased toward the “cream of the crop”. Some investigators have attempted to implement their own correction for volunteer bias by using either the 90th or 95th percentile performance levels as a criterion for judging the “age appropriateness” of various highway designs and/or experimental treatment levels (instead of the de facto 85th percentile group performance level used in traditional engineering studies). Yet, there is no direct empirical rationale to support the use of these inflated criterion levels. Without rigorous empirical support, research conclusions or recommendations based upon these inflated values would likely be rejected as being overly conservative, and, hence, not cost effective. Since the analyses regarding age appropriate design made in this chapter were based upon the adoption of the 85th percentile performance criterion, the conclusions reported here sometimes conflict with those reached by the author(s) of the original report.
A major development that has occurred since TRB Special Report 218 was published in 1988 has been the astronomical growth in intelligent transportation systems (ITS) research and development. Many of the anticipated mobility and safety problems resulting from a poor fit between the older roadway user and the highway infrastructure may be amenable to remediation through the strategic application of ITS technology. Since page space is limited, however, coverage of this topic is relegated to another chapter of this volume which focuses specifically upon ITS-related issues. Yet, the author is compelled to issue the following caveat: ITS technology offers not only the promise to assist older travelers but also the potential to increase their burden if the information processing demands of in-vehicle interfaces are engineered without special regard for their emerging needs and/or changes in capacity. Special care is needed to assure that older road users are not engineered out of next generation vehicle-highway systems.
2.0 Highway Geometric Design
Several major categories of research in the area of highway geometric design and operations have received considerable attention during the decade following TRB (1988) Special Report 218. Perhaps the most critical work in this area dealt with research that evaluated whether or not commonly applied models of human driving performance accommodated the capabilities of representative older drivers. If not, roadway elements and intersections designed on the basis of these models might not accommodate the needs of older road users. This work and other significant research involving intersection design, freeway design and operations, work zones and pedestrian crossings are covered in this section of the report.
2.1 AASHTO Models
The most widely used guidelines for highway design in the U.S. are those published in the American Association of State Highway and Transportation Officials’ (AASHTO)
A Policy on Geometric Design of Highways and Streets (universally referred to as the AASHTO Green Book). Many of the design formulas in the Green Book are based upon assumptions about the perception-reaction time of the “design driver”. TRB (1988) Special Report 218 presented evidence suggesting that the perception-reaction time assumptions of the models of driver behavior used in the AASHTO Green Book may not allow sufficient time to accommodate the general behavioral slowing seen among older drivers. If this were the case, the mobility of older drivers (and the safety of all drivers) would be compromised. Recently, a number of important studies have been conducted to investigate whether or not the AASHTO driver performance models accommodate the needs of older drivers.
One of the most fundamental design concepts contained within the AASHTO Green Book is that of stopping sight distance. The AASHTO stopping sight distance model specifies the minimum sight distance required by a driver to detect a criterion target on the roadway and then bring his or her vehicle to a safe stop. Stopping sight distance is the sum of two distances: the distance traversed by the vehicle from the instant the driver sights an object necessitating a stop until the moment braking is initiated (brake reaction time) plus the distance required to stop the vehicle from the instant the brake application begins (braking distance). The stopping sight distance model assumes a brake reaction time of 2.5 seconds. This value was selected because it accommodated the performance of the 85th-90th percentile driver in Johansson and Rumar’s (1971) classic field study of unalerted braking behavior. The braking distance component is based upon a simple physical model using the speed of the vehicle and the coefficient of friction between the tires and roadway as parameters (but see Fambro, Fitzpatrick and Koppa, 1997, who have recently proposed a reformulation of the braking distance equation based upon empirically determined deceleration behavior instead of the coefficient of friction). The minimum stopping sight distance requirement specified by the model is computed via Equation 1:
Stopping Sight Distance (ft) = 1.47PV + (V2 / 30(f ( G))
where: P = brake reaction time (2.5 s)
V = vehicle speed (mi/h)
f = coefficient of friction between tires and roadway
G = grade (%/100)
Equation 1
AASHTO (1994) Stopping Sight Distance Model
Lerner, Huey, McGee and Sullivan (1995) conducted a study designed to evaluate the validity of the 2.5 s brake reaction time assumption in the AASHTO stopping sight distance model. A sample of 253 drivers ranging in age from 20 to 70+ drove their own cars along a predetermined route. The drivers were given the “distraction task” of rating the quality of the road surface along various segments of the course. The last segment required the experimental participants to enter a closed section of highway. While traveling at a speed of approximately 40 mi/h (64 km/h) the driver reached a predetermined point on the closed section of highway where an unexpected emergency event requiring an avoidance response was triggered. A construction barrel, initially hidden from view, was rolled down a hill along the side of the road and apparently into the path of the participant’s vehicle (A safety chain was attached to the barrel to prevent it from actually making contact). The barrel emerged into view approximately 200 ft (61 m) ahead of the vehicle, thus, providing a time-to-collision of 3.4 s. Since the critical element of surprise would be difficult to replicate on subsequent trials, each participant experienced the “emergency event” only once. Because of technical problems and difficulty unambiguously scoring the behavioral responses, only 116 of the 253 participants yielded data that were judged “useful” for analysis. Of these 116 participants, 30 were young (ages 20-40), 43 were young-old (ages 65-69) and 43 were coded as old-old (70+ years-old). Fifty-six of these participants made an avoidance response that involved braking (The remaining participants executed a “steering only” or “other” avoidance response). The brake reaction time data for the emergency event are summarized in Table 1. Statistical analysis of this data revealed that there were no significant effects of age. All of the unalerted brake reaction times, except for a single case, fell within the 2.5 s value assumed by the AASHTO stopping sight distance model. The single exception demonstrated a brake reaction time of 2.54 s. These results have recently been confirmed in field studies conducted by Knoblauch, et al., (1995b) and Fambro, Fitzpatrick and Koppa (1997). Kloeppel, Peters, James, Fox and Alicandri (1995) also reported no age differences in the time needed to perform “emergency maneuvers” in an interactive driving simulator.
Table 1
Brake Reaction Times (seconds) for an Unexpected Driving Event
(from Lerner, Huey, McGee and Sullivan, 1995)
|Age (Years) |N |Mean |Median |85th Percentile |
|20-40 |14 |1.44 |1.35 |1.97 |
|65-69 |18 |1.59 |1.47 |1.92 |
|70+ |24 |1.49 |1.52 |1.72 |
|All |56 |1.51 |1.46 |1.85 |
At first glance, Lerner, et al.’s (1995) conclusion appears surprising. Laboratory studies of decision-making time consistently reveal a marked reduction in information processing time with advancing adult age. The finding that drivers over the age of 70 demonstrate the same median and 85th percentile brake reaction times for an unexpected event as 20-40 year-olds suggests that emergency braking for a roadway hazard is a highly over-learned or “automatized” behavior. Indeed, recent research has suggested that highly practiced behaviors involving consistent stimulus-response pairings tend to become “automatic” or proceduralized. According to Hasher and Zacks (1979), such over-learned or automatized operations are not only faster than “controlled” behavioral responses but consume very little of an operator’s attentional resource capacity, while responses to situations that are not over-learned require “controlled” or “executive” decision-making processes that are more “effortful” insofar as they require more attentional resources and often more time to be completed. Hence, it would appear that driving behaviors that have become automatized across a lifetime of experience may become somewhat protected from the harmful effects of normal human aging. This dichotomy between automatic versus controlled processes may ultimately play a significant role in our understanding of age-related changes in driving capacity and how we might intervene to mitigate the deleterious effects of such changes (see Schieber, 1994b).
A related and very important feature of highway geometric design is the concept of intersection sight distance.
Intersection Sight Distance (ft) = 1.47 V (J + ta)
where: V = highway design speed (mi/h)
J = perception-reaction time (usually 2.0 s)
ta = acceleration time required to clear intersection or
to reach 85% of design speed when turning onto cross road
Equation 2
AASHTO Case III Intersection Sight Distance Model
The AASHTO Green Book specifies a model for determining the minimum safe sight distance for several classes of intersections (see Table 2). All of these models have a perception reaction time component to represent the abilities of the design driver. A perception reaction time of 2.0 s is assumed for all of the intersection sight distance models except for Case II - which assumes a value of 2.5 s based upon its similarity to the scenario used to model stopping sight distance (described above).
TABLE 2.
AASHTO Intersection Sight Distance Model Classifications
|Intersection | |
|Classification |Description |
|I |No controls |
|II |Yield control |
|III-A |Stop controlled; crossing major road |
|III-B |Stop controlled; turning left onto major road |
|III-C |Stop controlled; turning right onto major road |
|IV |Signal controlled |
|V |Turning left from major roadway |
Earlier studies by Hofstetter, McGee, Crowley, Sequin and Dauber (1986) quantified the perception reaction time required to negotiate an intersection as the “time between the first head movement following a stop at an intersection and the first application of the accelerator”. Using this definition, field studies revealed a mean intersection perception reaction time of 1.8 s for the AASHTO Case III-A scenario. However, the estimated value for the 85th percentile reaction time was 2.6 s – suggesting that the AASHTO model’s 2.0 s estimate was not allowing sufficient time for many drivers to manage the information processing demands of many intersections. The data obtained for the Case III-B and III-C intersection scenarios demonstrated an even greater mismatch between the AASHTO assumption of 2.0 s and the empirically determined 85th percentile perception reaction time requirements. Based primarily upon the results of this investigation, TRB (1988) Special Report 218 recommended that a systematic series of field studies be conducted to evaluate whether or not the perception reaction time assumptions of the AASHTO intersection sight distance models were accommodating the needs of the older driver population.
Several studies have systematically attempted to evaluate the “age appropriateness” of the perception-reaction time estimates used by the AASHTO intersection sight distance models. These studies have focused upon the AASHTO Case III and Case V scenarios since older drivers appear to be “over represented” in automobile crashes at stop controlled intersections and while executing left-turn maneuvers (Schieber, 2000).
Lerner, Huey, McGee and Sullivan (1995) developed a technique to “open” the visual search loop of a driver stopped at an intersection; thus, simplifying the measurement of the perception-reaction time needed to negotiate an AASHTO Case III maneuver. That is: When drivers encountered a stop sign they came to a full stop. Once stopped, they were required to look down at an indicator lamp (situated low on the instrument panel) until it signaled them to proceed. Once the signal lamp was activated, they raised their heads, searched the intersection and then completed their maneuver through the intersection. Perception-reaction time was defined at the interval between the instant the driver’s head was raised from the indicator monitoring task until the moment that the vehicle began to accelerate through the intersection. Using this definition, daytime perception-reaction times needed to negotiate AASHTO Case III A, B and C intersection maneuvers were collected from 25 young (ages 20-45), 27 young-old (ages 65-69) and 29 old-old (70+ years-old) volunteer drivers. The most noteworthy finding was that no age-related slowing in perception-reaction time was noted for Case III maneuvers (collapsed across 14 intersections). The median perception-reaction time for all participants was 1.3 s while the 85th percentile interpolated from the cumulative response function was approximately 2.0 s. Based upon these findings, Lerner, et al. (1995) concluded that the current AASHTO models for specifying minimum intersection sight distance successfully accommodate the information processing time needs of older drivers as well as younger ones.
2.2 Gap Acceptance Models
Several recent studies have concluded that the minimum safe sight distances specified by the AASHTO models appear to be overly conservative. That is, drivers (at least younger ones) may not really need as much sight distance as demanded by these models. For example, the AASHTO model for minimum safe sight distance required for executing a Case III-B intersection maneuver yields values ranging from 290 ft to 1340 ft (88 m to 408 m) as highway design speed is increased from 25 mi/h to 65 mi/h (40 km/h to 104 km/h). When these same intersection sight distances are expressed as times-to-arrival of approaching vehicles, one must conclude that entering onto a road with a speed limit of 65 mi/h (104 km/h) requires a critical gap in the traffic stream of at least 14.1 s (i.e., the time it takes to travel 1340 ft at a speed of 65 mi/h). Yet, observational field studies reveal that the average gap acceptance times demonstrated by drivers at historically safe intersections are typically only a fraction of the equivalent temporal safety cushions demanded by the AASHTO models. Drivers, on average, select gaps that are 9.2 s in duration – regardless of the speed limit (Harwood, Mason, Brydia, Pietrucha and Gittings, 1996). Based upon findings such as these, some investigators have argued that geometric design based upon gap acceptance models is both valid and more cost-effective than the overly conservative AASHTO approach (see Smiley, 1999). Indeed, such a conservative tendency might help explain Lerner, et al.’s (1995) unexpected conclusion that the AASHTO stopping sight distance and intersection sight distance models – without modification – already accommodate the perception-reaction time requirements of the 85th percentile older driver.
The rationale for the use of the gap acceptance measurement as a criterion for intersection sight distance is given by Harwood, et al. (1996) as follows:
…if drivers will accept a specific critical gap, such as 7.5 sec, in the major
road traffic stream when making a turning maneuver, and if such maneuvers
are routinely completed safely, then sufficient intersection sight distance should
be provided to enable drivers to identify that critical gap. (p. 39)
Based upon this rationale, driver gap acceptance times derived from observational studies of highway sites with good safety records can be used to develop intersection sight distance criteria. This approach has the advantages of being relatively easy to measure while at the same time avoiding the conceptual difficulties that plague the AASHTO models (e.g., serial/sequential versus parallel/simultaneous processing stages). Harwood, et al. (1996) have completed a series of analytic and field studies as part of a comprehensive National Cooperative Highway Research Program (NCHRP) project that has recommended the adoption of models for highway geometric design based upon this gap acceptance approach. Several U.S. states (California, Oklahoma) and national governments (France, Sweden) have already adopted highway geometric design guidelines based upon gap acceptance models. Only recently, however, have data become available to allow one to evaluate whether or not these models accommodate the needs of older drivers.
Lerner, Huey, McGee and Sullivan (1995) had 29 young (ages 20-40), 23 young-old (ages 65-69) and 22 old-old (70+) volunteer participants sit in a static test vehicle positioned perpendicularly to a busy roadway and asked them to make “yes” versus “no” judgments about potential maneuvers during daytime lighting conditions. Approximately half of the participants in each age group made judgments for a low speed road (30 mi/h)(48 km/h) while the other half were exposed to a high speed condition (50 mi/h)(81 km/h). They were asked to imagine that they were awaiting the opportunity either to proceed straight through an intersection, turn left or turn right (AASHTO Case III maneuvers A, B or C, respectively). Participants signaled their judgments by depressing a button whenever it was considered safe to initiate the target maneuver and released the button whenever it was considered unsafe. The test vehicle was equipped with microwave sensors and a battery of video cameras that allowed accurate off-line calculation of the vehicle gap times that were either accepted as safe or rejected as unsafe. The resulting data were quantified by constructing functions that expressed the cumulative probability of gap acceptance as a function of gap size (in seconds). Critical gap acceptance parameters interpolated from these functions, collapsed across speed condition and type of maneuver, are summarized in Table 3.
Table 3.
Gap Acceptance Times (seconds) for
Daytime Case III Maneuvers as a Function of Age
(Lerner, Huey, McGee and Sullivan, 1995)
| | | |Age Group | |
| | |20-40 |65-69 |70+ |
| |50% acceptancea |6.7 |7.2 |8.2 |
| |85% acceptanceb |9 |11 |11 |
a Daytime values reported in Table 16
b Interpolated from Figure 18
The critical performance value for the gap acceptance paradigm is typically taken as the time at which 50 percent of the traffic gaps are accepted by the driver. Gap intervals shorter than this critical gap are less likely to be accepted than rejected while gaps longer than this interval are more likely to be accepted. Statistical analysis of the 50 percent gap acceptance parameters revealed a significant age effect. The oldest drivers demonstrated a critical gap acceptance time that was one full second greater than that of their younger counterparts. This difference climbed to approximately 2 s in the case of the 85 percent gap acceptance interval. Lerner, et al’s (1995) demonstration that gap acceptance times must be longer in order to be judged as “safe” by older observers suggests that proposed highway geometric design guidelines based upon 50 percent gap acceptance models may not accommodate the needs of the older driver population.
Staplin, Harkey, Lococo and Tarawneh (1997) conducted a field study to investigate age differences in gap acceptance times during the performance of an AASHTO Case V maneuver (left turn from a major roadway). Thirty-three young (20-45), 37 young-old 65-74) and 30 old-old (75+) drivers made left-turns during the permissive phase of a traffic signal while positioned within a left-turn lane of varying offset geometry (the effects of varying offset geometry are discussed in a subsequent section of this report). Cumulative curves for the probability of executing a left-turn as a function of the size of the gaps in the approaching traffic stream were then calculated. The critical (50 percent) gap acceptance parameter was estimated for each age group using logistic regression. These values are reported in Table 4 along with the equivalent gap sizes recommended by the AASHTO and Harwood, et al. (1997) gap acceptance models for the Case V scenario. Reference to Table 4 reveals that the critical gap size observed for drivers over 75 years of age is almost 1 sec longer than that observed for younger drivers. The 6.0 s of equivalent highway sight distance recommended by the Harwood, et al. (1996) gap acceptance model exceeds the needs of the younger drivers but fails to accommodate the oldest drivers. Unlike the proposed gap acceptance model, however, the 8.0 s equivalent gap size generated by the AASHTO Case V intersection sight distance model appears to accommodate all of the drivers in the left-turn study. Unfortunately, the Staplin, et al. (1997) study did not report 85% gap acceptance values for the three age groups.
Table 4.
Critical (50 percent) Gap Acceptance Times (seconds) for Left-Turns from the
Staplin, et al. (1997) Field Study with Equivalent Gap Time Recommendations
from the AASHTO and Harwood, et al. (1996) Case V Models
Age Group
| |25-45 |65-74 |75+ |
|Staplin, et al., 1997 Data |5.9 |5.9 |6.6 |
|Harwood, et al., 1996 Case V Modela |6.0 |6.0 |6.0 |
|AASHTO Case V Equivalent Gap Sizeab |8.0 |8.0 |8.0 |
a Harwood, et al., 1996; p. 81 [G=6.0 s; 2 lanes of traffic]
b Staplin, et al., 1997; Table 64 [speed=35 mi/h; acceleration distance=84 ft; accel time = 5.8 s]
2.3 Age-Appropriate Models for Highway Geometric Design
TRB (1988) Special Report 218 fostered a considerable amount of research regarding the appropriateness of the AASHTO perception-reaction time models for the older driver population. The data from several studies suggest that the minimum information processing time requirements of “typical” older drivers are accommodated by the perception-reaction time assumptions (2.0-2.5 s) of the AASHTO sight distance models. Recently, however, there has been considerable activity in the traffic research community claiming that the AASHTO models are too conservative – resulting in highway geometric design standards that are unnecessarily costly in their realization. As an alternative, a new generation of models based upon field surveys of driver gap acceptance behavior has been proposed. There is clear evidence, however, that the critical time parameters of currently proposed gap acceptance models do not accommodate the processing time requirements of drivers over the age of 70. Additional research will be required to establish the validity and age appropriateness of gap acceptance models for highway geometric design. A major question that remains to be determined with respect to gap acceptance models is whether or not the “critical” gap should be at the 50 percent level. Does it really make sense to design an intersection based upon a value that is judged to be “unsafe” 50 percent of the time? It is interesting to note that if an 85 percent safe gap criterion were used for the critical gap size then the differences between the sight distances recommended by the gap acceptance models and AASHTO models would be considerably reduced. Indeed, Harwood, et al.’s (1996) own data for the Case III-B (their Fig. 25) and Case III-C (their Equation 41) scenarios appear to support such a conclusion.
3.0 Traffic Operations
3.1 Intersections
Knoblauch, et al. (1995b) conducted a study to determine how well drivers understand the various configurations of protected and permitted phases of left-turn traffic signals Sixteen different scenarios from the Manual of Uniform Traffic Control Devices (MUTCD) and related state traffic design manuals were tested. Results of the study clearly revealed that the 126 older drivers (ages 65 and greater) did not comprehend the operational meaning of either the protected or permitted phases of left-turn traffic signalization as well as the 121 younger drivers (those less than 65 years of age). These results are consistent with the findings of other recent studies demonstrating age-related declines in traffic signal comprehension (e.g., Williams, Ardekani and Asante, 1992; Bonneson and McCoy, 1994). However, it should be noted that Knoblauch, et al. (1995b) concluded that neither age group demonstrated an “acceptable level of comprehension associated with left-turn signalization” (p.52). Furthermore, they recommended that future efforts be directed toward improving the comprehension level of all drivers – rather than developing efforts targeted solely to the older driver population. Knoblauch and other investigators have repeatedly suggested that more uniform and consistent application of the MUTCD specifications would contribute to improvements in driver comprehension of traffic control devices.
Numerous studies have reported that older drivers are more likely to be involved in a crash involving a left-turn maneuver than any other category of motor vehicle accident (e.g., Knoblauch, et al., 1995b). Joshua and Saka (1992) reported that the relative alignment of opposing left-turn lanes contributed significantly to the probability of a crash while turning left – especially during the permitted phase of traffic signalization. They found that the relative alignment of opposing left-turn lanes influenced the intersection sight distance of drivers when both of the opposing turning lanes were occupied by waiting vehicles. When there was negative or zero offset in the relative alignment of the opposing left-turn lanes, sight distance was reduced and the number of crashes observed was elevated. However, intersections designed with opposing left-turn lanes having a positive offset in relative alignment were found to afford turning drivers with greater sight distance and (presumably as a result) a decreased incidence of left-turn crashes (see Figure 1 for turning lane offset examples). In their search for countermeasures to reduce the high proportion of left-turn crashes among older drivers, Staplin, Harkey, Lococo and Tarawneh (1997) identified the optimization of left-turn lane geometry as a likely source of remediation via improved highway design. Their primary argument was that older drivers were more likely to suffer due to limited intersection sight distance while turning; and, that appropriate offset of opposing left-turn lanes would serve to minimize exposure to such conditions. In order to evaluate this claim, they conducted a field study examining age differences in left-turn maneuvers as a function of intersection geometry. A total of 100 drivers participated in the study. There were approximately equal numbers of participants in each of three age groups: young (25-45 years-old), young-old (65-74) and old-old (75 years of age or over). Each participant drove around a predetermined course so as to pass through each of the experimental intersections four times. The intersections selected for study were signalized and
(a) partial positive offset (b) no offset (i.e., aligned)
(c) partial negative offset (d) full negative offset
Figure 1.
Left turning lane offset geometries investigated by Staplin, et al. (1997)
characterized by four levels of opposing left-turn lane geometry: (1) full negative offset, (2) partial negative offset, (3) aligned, or “zero” offset, and (4) partial positive offset (see Figure 1). The sight distances afforded at these intersections during the performance of the criterion left-turn maneuver (i.e., turning during the permitted phase of the traffic signal while another vehicle occupied the opposing left-turn lane) increased as lane offset moved from “full negative” to the “partial positive” condition. Video equipment positioned at each intersection recorded driver behavior as well as the opposing traffic stream for off-line analysis. The primary measure of driving performance used in this study was the size of the “critical gap” in the opposing traffic stream that was accepted by drivers during the performance of the criterion left-turn maneuver. The results of the study are depicted in Figure 2. They confirmed the working hypothesis that positive offset of opposing left-turn lanes would result in improved driving performance – especially for older drivers. Another aspect of driving performance measured in this study concerned the relative positioning of the driver’s vehicle just prior to executing the criterion left-turn maneuver. Most drivers tended to “encroach” slightly into the intersection (i.e., self-position) prior to turning, presumably to compensate for the obstruction in sight distance caused by the vehicle(s) in the opposing left-turn lane. However, the proportion of drivers demonstrating such compensatory self-positioning prior to executing a left turn declined significantly with increasing adult age. Fully 92 percent of the young drivers and 84 percent of the young-old drivers self-positioned their vehicles while waiting to turn left, while only 68 percent of the old-old drivers engaged in such behavior. Although such a difference in behavior appears to be subtle, it could be indicative of an age-related “passivity” that is rooted either in decreased “situation awareness” or some other systematic change in the nature of the driving task among the oldest road users. Regardless of the cause, the Older Driver Highway Design Handbook (see below) suggests that such age-related differences in vehicle positioning prior to turning represent a critical consideration for intersection geometric design. That is, if opposing left-turn lanes have a positive offset, there is no need to self-position (Hence, reducing the potential impact of this negative behavior observed among older drivers).
[pic]
Figure 2. Driving performance as a function of driver age
and Left-turn lane geometry (Staplin, et al., 1997)
Staplin, et al. (1997) also conducted a study to examine the effects of intersection lane geometry upon age differences in driving behavior while performing a right-turn. This study focused upon the right-turn on red (RTOR) maneuver since previous investigations had demonstrated that older drivers were especially vulnerable to experiencing a crash while executing such turns (e.g., Hu, Young and Lu, 1993). Research participants negotiated a predetermined test route that included four experimental intersections that varied along two design dimensions: (1) the right-turn lane was either channelized or not channelized, and (2) the receiving road either had an acceleration lane or did not have an acceleration lane. Approximately 100 drivers equally distributed across three age groups participated in the study. The age groups included: young (25-45), young-old (65-74) and old-old (75+) drivers. The results of the study indicated that the implementation of an exclusive right-turn lane (established through lane channelization) contributed significantly to the mobility of the younger drivers. Participants from the young and young-old age groups executed right-turns at speeds from 3 to 5 mi/h (4.8 to 8.0 km/h) faster at intersections with channelized right-turn lanes than at intersections without the channelization treatment. However, this mobility benefit of channelization was not observed among drivers in the old-old group. The young drivers took advantage of 83 percent of the opportunities to execute a RTOR maneuver. Young-old drivers, however, completed only 45 percent of such opportunities while the old-old drivers attempted to turn right while the traffic signal was red only 16 percent of the time. Also of great interest was the finding that channelization significantly increased the probability of a driver’s taking advantage of completing a RTOR maneuver without the need of first coming to a complete stop. This phenomenon was robust for both the young and the young-old groups but almost completely absent among the old-old drivers who came to a complete stop on 19 of the 20 RTOR maneuvers that were observed. Finally, the results indicated that the presence of an acceleration lane on the receiving road had little effect on right-turn behavior regardless of driving age.
3.2 Freeway Operations
Reilly, Pfefer, Michaels, Polus and Schoen (1989) conducted an observational field study of 35 freeway sites in order to investigate the relationship between driver gap acceptance and acceleration/deceleration behavior while entering/exiting the highway. They developed a model of driver freeway entry/egress behavior and concluded that the AASHTO Green Book guidelines for freeway speed-change lanes did not provide sufficient distance for either merge or diverge maneuvers. This finding suggests that freeway entry/egress lanes may fall far short of meeting the needs of many older drivers. Survey results from licensed drivers over the age of 65 support this conjecture (Benekohal, Resenda, Shim, Michaels and Weeks, 1992).
Knoblauch, Nitzburg and Seifert (1997) conducted a major analytic study of older driver freeway operations. The objectives of the investigation included: (1) identify the characteristics of older drivers that affect their ability to drive on freeways, (2) identify the characteristics of freeway driving that cause the greatest difficulties for older drivers, (3) conduct problem identification research to define difficulties experienced by older drivers on freeways, and (4) recommend further research needed to develop guidelines for countermeasures designed to accommodate the needs and capabilities of older drivers. These objectives were achieved through the application of human factors task analysis techniques, focus groups, a large-scale survey and detailed analysis of freeway crash records. Their focus group studies included 44 men and 44 women ranging from 65 to 88 years-old (median age: 70) recruited from 4 major metropolitan areas (Phoenix, AZ; San Diego, CA; Tampa, FL; Washington, DC). Results from the focus groups suggested that older drivers are highly concerned about traffic congestion, inconsistent signing and sign placement, entrance ramps that did not allow enough time to merge onto the mainline of a freeway, work zones and inadequate rest areas. Problems with freeway signing included inadequate advance warning for right-turn and exit-only lanes and difficulty reading shoulder-mounted signs. In fact, well-lit overhead signs were preferred to shoulder-mounted signs. Design issues raised by the focus groups also included the need for longer acceleration (i.e., merging) lanes, increased use of high concrete median barriers to promote safety and reduce glare, more shoulder rumble strips to help maintain driver alertness and a distinct dislike for combination exit/entrance ramps. Perhaps the most interesting finding from the focus groups was that there was no direct evidence that older drivers avoided using freeways. However, their survey of 1,400 members of the American Association of Retired Persons, ranging from 50 to 97 years of age (mean age: 72.2), revealed that approximately one-quarter of the respondents deliberately avoided freeway driving. Lane-changing difficulties were common as a large number of drivers reported that they would rather slow down than pass a slow moving vehicle. Nearly half of those sampled reported problems with fatigue while driving. A significant number of drivers reported difficulties entering or exiting freeways: 25 percent reported actually coming to a full stop before merging into the mainline traffic stream and 52 percent slow down prior to reaching the exit lane. Approximately half of these older drivers reported that they had problems staying in their lanes because of faded lane markings and fully three-quarters reported that more light was needed at various freeway locations such as exit ramps and interchanges. These drivers also indicated that they had a great appreciation for the benefits provided by various sources of roadway delineation, including painted lane markings, raised pavement markings (RPM’s) and post-mounted reflectors. Finally, the survey respondents concurred with the focus groups’ preference for overhead, as opposed to shoulder-mounted, freeway signs. The accident analyses conducted as part of the Knoblauch, et al. study revealed that older drivers were over involved in single-vehicle run-off-road crashes, possibly due to the effects of fatigue (almost one-quarter were reported to be fatigued or asleep). Although older drivers were not more likely to be involved in multi-vehicle freeway crashes, when they were involved in a multi-vehicle crash they were more likely than young drivers to be merging or changing lanes. Older drivers were also more likely to be struck by vehicles that were traveling faster than they were (although not necessarily speeding). Table 5 represents a modified version of Knoblauch, et al.’s (1997) summary of their results along with recommendations for future research in areas related to freeway operation and geometric design.
Table 5.
Older Drivers’ Problems Encountered during Freeway Driving with
Recommended Research Needs (Knoblauch, et al., 1997)
|Research Area |Problem(s) |Needed Research |
|Navigation/ |Difficulties navigation. Over involve-ment in |Further explore problem. Develop and test |
|Wayfinding |crashes in unfamiliar areas. |alternative traffic control device designs. |
|Freeway Ramp |Merging onto mainline from ramps. |Identify ramp geometry and mainline |
|Merging | |characteristics that contribute to the |
| | |problem. Develop and test new designs. |
|Illumination |Reduced visibility while driving at night. |Identify critical factors associated with |
|Requirements | |highway lighting; i.e., lamp placement, |
| | |intensity, etc. |
|Speed/Lane |Inappropriate lane selection. |Identify relevant design parameters: |
|Selection |Inappropriate speed selection. |horizontal/vertical curvature; lane and |
| | |shoulder width; median type and proximity; |
| | |guard rail type and proximity. |
|Work Zones |A major concern and reason for avoiding freeways. |Identify characteristics of construction and|
| | |maintenance areas that are especially |
| | |troublesome for older drivers. Develop and |
| | |test treatments to improve older driver |
| | |performance in work zones. |
|Fatigue/ |Fatigue was identified as a major contributory |Identify interventions that could offset |
|Medication. |factor in single-vehicle, freeway crashes. |effects of fatigue in older freeway users. |
| |Self-reports of fatigue problems also very common.| |
|Lane-Changing/ |Many crashes related to lane-change maneuvers. |Conduct detailed behavioral analysis of |
|Passing Behavior |Self-reported problems with lane-changing and |lane-changing and passing behavior. |
| |passing. |Determine adequacy of exit and advanced exit|
| | |signing relative to the time needed to |
| | |complete these maneuvers. |
|Roadway |Heavy reliance on delineation, RPM’s and post |Determine minimum and optimal size/width and|
|Delineation |mounted delineators at night and during |retroreflectivity of lane markings, RPM’s |
| |challenging visibility conditions. Run-off-road |and post-mounted delineators for older |
| |and lane-changing crashes may be related to need |drivers. |
| |for improved delineation. | |
|Roadway |Strong preference for overhead signing. |This preference is unexpected and may not |
|Signing | |generalize to unilluminated signs at night. |
| | |Additional investigation needed. |
3.3 Work Zones
During recent years there has been a significant increase in the number of traffic-related fatalities in and around roadway maintenance and construction areas. Much of this increase in work zone fatalities is no doubt due to an increase in the amount of maintenance required as the U.S. highway infrastructure itself begins to reveal the effects of aging. As a result of these trends, work zone safety has become a high priority issue for the Federal Highway Administration which recently established the National Work Zone Safety Information Clearinghouse [] at the Texas Transportation Institute (Dewar and Olson, in press).
Since work zones are associated with numerous visual, attentional and cognitive challenges, one would expect that crash frequency among older drivers would be elevated in the area of roadway construction and maintenance projects. However, little work has been done to investigate this potential problem among older drivers. Chiu, Mourant and Bond (1997) conducted a study to investigate the interacting influences of driver age and roadway illumination upon driving performance in work zones. Young (less than 35 years-old) and older (over age 58) participants operated an interactive driving simulator under six different conditions defined by two types of work zones (all lanes shift versus single lane closure) and three illumination conditions (day, dusk and night). Steering errors (lane departures) were greater in the dusk and nighttime condition relative to daytime for all drivers. In addition, all drivers reduced their driving speed in the nighttime condition, although young drivers drove slightly faster than older drivers regardless of the level of illumination. Older drivers made lane changes in response to work zone shifts in lane geometry at a much later time than their younger counterparts – especially in the simulated nighttime condition. When asked about the helpfulness of simulated traffic control devices, older drivers reported that lane markings with “reflectors” were beneficial while some young drivers reported that these same lane markings were “distracting” because of their excessive brightness.
Given the increased workload and safety hazards imposed by roadway work zones, it is surprising that little is known about age differences in driving behavior and crash history in such situations. The simulation-based work of Chiu, et al. (1997), although quite preliminary, strongly suggests that older drivers have increased difficulties negotiating roadway work zones despite the fact that they are more likely to slow down and obey signs posted before and in the work zone area itself. It is likely that the complexity of work zone environments and their tendency to violate driver expectations may overload the functional capacities of many older drivers. Additional studies based upon detailed crash analysis and observational field studies appear to be warranted.
3.4 Pedestrian Crossings
TRB (1988) Special Report 218 identified the high incidence of older pedestrian injuries and fatalities occurring at intersection as a high priority area for research. Subsequent epidemiological studies have verified these findings (e.g., Garber and Srinivasan, 1991; Knoblauch, et al., 1995a). A specific recommendation of the report was the need to evaluate the walking speed parameters used in the design of signal lights and related traffic control devices at intersections. The walking speed for pedestrians crossing a street is assumed to be 4 ft/sec (1.2 m/sec) in the U.S. DOT (1988) Manual of Uniform Traffic control Devices. TRB (1988) Special Report 218 urged that the age appropriateness of this walking speed parameter be evaluated; and, if necessary, updated for future reference in the MUTCD and related highway design guidelines.
Knoblauch, Nitzburg, Dewar, Templer and Pietrucha (1995a) conducted a comprehensive study of the characteristics of older pedestrians related to street crossing behavior. Using the conventional tools of a literature review, task analysis of pedestrian crossing behavior, focus groups and functional analysis of pedestrian crash records, they developed a taxonomy of the problems faced by many older pedestrians attempting to negotiate busy urban intersections. In order to translate these findings into design guidelines, they conducted field studies to establish parameters needed to describe street crossing behavior of older adults (viz., walking speed, stride length and latency between signal change and the initiation of a crossing response). Data from 3458 young (appeared to be less than 65 years-old) and 3665 old (appeared to be 65 years-old or greater) were collected from urban intersections in four U.S. East Coast cities. Mean walking speeds for the young and old pedestrians were 4.79 ft/sec (1.43 m/sec) and 3.97 ft/sec (1.18 m/sec), respectively. This age difference in average walking speed of nearly 1 ft/sec was statistically significant. In order to establish a “design value” for walking speed that would accommodate 85 percent of the pedestrian population the 15th percentile values for each age group were also calculated. The resulting values were 3.97 ft/sec (1.18 m/sec) and 3.08 ft/sec (0.92 m/sec) for the young and old pedestrians, respectively. The slower walking speed of older pedestrians appeared to be due to the fact that the length of their average stride was significantly shorter than that of their younger counterparts. It was concluded that the traditional design speed of 4 ft/sec could result in traffic signal timing and related pedestrian crossing implementations that failed to accommodate the emerging needs of a significant proportion of the older population. On the basis of these results, Knoblauch, et al. (1995a) recommended that the MUTCD and the AASHTO Green Book be modified to reflect a pedestrian walking speed that accommodated 85 percent of older adults observed – namely, 3 ft/sec (0.9 m/sec). Similar conclusions have also been reached by other recent investigations (e.g., Hoxie and Rubenstein, 1994; Coffin and Morrall, 1995). As a result of this work, appropriate changes to both the MUTCD and Green Book have been proposed.
Significant age differences in the latency of initiating a walking response following the onset of an appropriate signal from a traffic control device were also observed in the Knoblauch, et al. (1995a) study. Mean startup time latencies of 1.93 and 2.48 s were computed for young and old pedestrians, respectively. The respective 85th percentile startup times for the young and old groups were 3.06 and 3.76 s. Hence, the “design pedestrian” accommodating the older group would require approximately 3.75 s to begin crossing the street once the light turned green and/or the “WALK” signal was initiated. This long startup time value suggests that there is much that remains to be learned about pedestrian behavior at signalized intersections (see Dewar, 1992). Treatments aimed at improving pedestrian “situation awareness” during the “DON’T WALK” phase of pedestrian crossing control would appear to hold great promise for improving the safety and efficiency of urban intersections. However, it should be remembered that decrements in basic visual capacities (such as acuity and contrast sensitivity) may also make it more difficult for older pedestrians to read traffic control devices as well as detect and judge approaching vehicles. For example, approximately 25 percent of older pedestrians sampled in a busy urban area reported experiencing difficulty seeing the pedestrian crossing traffic control from the other side of the street (Bailey, et al., 1992).
In an effort designed to improve pedestrian “situation awareness” in a targeted and cost-effective manner, Blomberg and Cleven (1998) developed and verified the efficacy of the “zoning” technique for reducing pedestrian crossing fatalities in the older population. Accident analyses in Phoenix, AZ revealed that 7 “hot” zones (6 circular zones with 1 mile radius; 1 linear stretch of roadway extending 2 miles) accounted for 54% of the pedestrian accidents (but only 5% of the land mass of the metropolitan area). An intensive educational and informational campaign focused only in these identified zones (emphasizing behavioral phenomena such as “daytime conspicuity” and “wait for a fresh green before crossing”) resulted in significant reductions in pedestrian accidents. The authors argued that “zoning” technique for focusing traffic safety countermeasures might be similarly effective in other domains of traffic and safety engineering.
3.5 Traffic Calming and Roundabouts
Traffic calming devices date back to the introduction of the “speed bump” in the Netherlands. Since 1970, there has been increasing development and deployment of traffic calming systems throughout western Europe. These approaches have focused upon maintaining the “quality of life” for neighborhood residents and drivers alike (Schlabbach, 1997). Traffic calming approaches are beginning to be planned and implemented with increasing frequency in the United States (Lockwood, 1997). Upon initial inspection, the traffic calming approach appears to support the needs of older road users. For example, slower traffic flows with smaller speed variances would be expected to result in decreased time pressures for drivers and pedestrians. Yet, this may not always be the case. Traffic calming engineering devices such as narrowed lanes and roundabouts may also be associated with unanticipated increases in workload demands upon drivers - especially frail older drivers. The increased complexity of decision-making required to negotiate a busy roundabout may be particularly problematic. However, recent guidelines and specifications for traffic calming devices such as roundabouts appear to have ignored important issues regarding the needs of older road users (see Institute of Transportation Engineers, 2000).
3.6 Older Driver Highway Design Handbook
In January, 1998, the Federal Highway Administration released a comprehensive set of guidelines that attempts to translate our knowledge about human aging into principles of highway geometric design and operations. These guidelines are collectively known as the Older Driver Highway Design Handbook (Staplin, Lococo and Byington, 1998). The Handbook is “must” reading for anyone interested in highway design and older road users. Based on an exhaustive review of the literature, this volume makes specific recommendations about roadway and traffic control device design. The topical areas covered in the Handbook are listed in Table 6 and summarized in a separate chapter in this volume. The comprehensive nature of this work merits follow-up by the Federal Highway Administration. Perhaps a standing committee could be formed to continuously monitor and validate the guidelines. There is little doubt that given this level of commitment that many of the guidelines would eventually evolve to the point where they might be formally accepted into the MUTCD.
Table 6.
Contents of the Older Driver Highway Design Handbook (1998)
[pic]
4.0 Traffic Control Devices
4.1 Roadway Delineation
Safe and efficient driving depends upon adequate roadway delineation. Both the boundaries of the roadway and nearby off-road appurtenances need to be clearly visible to the driver of a motor vehicle; otherwise drivers who unintentionally deviate from their lane face risks that could lead to collisions with other vehicles or stationary hazards such as unprotected bridge supporting columns or drainage culverts. Although naturally occurring changes in brightness and contrast are available to provide for much of the visual requirements of driving during daylight conditions, most of the visual information available to the driver at night comes from engineered roadway and hazard delineation treatments using specially engineered retroreflective materials. Without the widespread use of such retroreflective materials light from vehicle headlights would reflect away from the driver. Retroreflective materials significantly enhance the revealing power of headlamps by selectively steering their light output back to the eye of the driver.
Numerous studies have revealed that adequate roadway delineation must support two classes of driver visual needs. First and foremost, roadway delineation must support the driver’s immediate needs of continuous lane tracking – the so-called short-range visual requirements of steering. Drivers require approximately 2-3 s of delineation lead time in order to maintain smooth and efficient lane keeping (see Rumar and Marsh, 1998). Secondarily, roadway delineation treatments must provide for the long-range visual needs of the driver. Rumar and Marsh (1998) have shown that drivers need approximately 5 s of advance warning to adequately prepare for significant changes in roadway curvature and the appearance of intersections or nearby off-road hazards.
TRB (1988) Special Report 218 acknowledged that improvements in roadway delineation treatments might be needed to accommodate the changing visual and attentional capacities of older adults; but concluded that there was insufficient data available to guide the formulation of specific recommendations (Deacon, 1988). Since then, several studies have been conducted to quantify the magnitude of age differences in visibility distance afforded by existing delineation treatments as well as to identify improvements in delineation that could compensate for these differences. Owing to the special problems experienced by older drivers under low illumination conditions, most of this work has focused upon retroreflective delineation treatments designed to improve the nighttime visibility of the roadway and nearby hazards.
4.1.1 Lane Markings
Given the significant reductions in retinal illuminance and contrast that accompanies normal adult aging, it would be expected that older drivers would experience difficulty seeing pavement markings while driving at night. Indeed, Zwahlen and Schnell (1998) have reported that the average nighttime visibility distance for retroreflective pavement markings demonstrated by older drivers (mean age: 68.3 years-old) was only about half that achieved by younger drivers (mean age: 23.2) under identical field conditions. Simulator-based studies have demonstrated similar results (e.g., Freedman, Staplin, Gilfillan and Byrnes, 1988; Staplin, Lococo and Sim, 1990). Such age differences are probably exacerbated under wet weather conditions.
Current guidelines and recommendations for the brightness of retroreflective road markings do not specifically address the emerging requirements of older drivers and range from a low of 90 mcd/m2/lux (Allen, O’Hanlon and McRuer, 1977) to a high of 400 mcd/ m2/lux (Ethen and Woltman, 1986). The recently proposed European standard (Comité Européen de Normalisation, 1997) specifies a replacement retroreflectivity level of 300 mcd/ m2/lux, but does not recommend an optimal value for new applications. Zwahlen and Schnell (in press) used the well-validated C.A.R.V.E. computer model to determine the minimum pavement marking retroreflectivity levels needed to fully accommodate short-range steering performance at night. Their calculations were based upon a constant preview time of 3.65 s and the visual needs of the average 62 year-old observer (i.e., 85th percentile licensed driver). Their recommended minimum values for white edge line pavement markings are presented in Figure 3.
Figure 3.
Minimum pavement marking retroreflectivity requirements as a function of speed.
An examination of Zwahlen and Schnell’s recommendations reveals that the proposed European replacement level of 300 mcd/m2/lux would accommodate the short-range visual guidance needs of the average 62 year-old observer for driving speeds below 65 mi/h (105 km/h). Note, however, that the European minimum would need to be approximately doubled to meet the minimum retroreflectivity demands of driving at speeds of 75 mi/h (120 km/h).
The expected retroreflectivity levels of newly installed roadway markings are summarized in Table 7. When newly applied, all of the state-of-the-art materials depicted in Table 7 appear capable of meeting the minimum visibility requirements specified by Zwahlen and Schnell’s model at 65 mi/h (105 km/h). However, engineering estimates of the half-life durability (i.e., time required for retroreflectivity to degrade to half of its initial value) for these materials suggest that they would have to be replaced every year or two in order to remain in compliance with the recommended minimum levels. The peak retroreflective performance of modern roadway marking systems has improved dramatically over the last decade. However, additional research is needed to significantly improve their operational durability with the ultimate goal of meeting and maintaining the visibility needs of older drivers in a cost-effective manner.
Table 7.
Retroreflectivity Levels for Newly Applied Road
Marking Materials with Half-Life Durability Estimates
| |Initial |Estimated Half-Life |
| |Value |Durability |
|Pavement Marking Material |mcd/m2/lx | |
|Paint/Glass beads |300-500 | < 1 year |
|Thermoplastic |350-400 |< 2 years |
|Durable Traffic Tape |700-800 |2 years |
|Ceramic |1200 |unknown |
Pietrucha, Hostetter, Staplin and Obermeyer (1996) conducted a series of engineering studies designed to evaluate the efficacy of currently available retroreflective treatments for improving the nighttime lane visibility distance of older drivers. Based upon a review of the literature and a preliminary simulation-based study, they identified 11 treatment combinations that showed some potential for improving the nighttime visibility of older drivers without negatively affecting the performance of young drivers [Note: Despite prior indications that older drivers would benefit from such a manipulation (see Deacon, 1988), none of the candidate treatment combinations involving an extra-wide edge line
(8 in)(20 cm) survived the preliminary evaluation phase of this study]. The experimental treatments included variations in the width and brightness of center and edge lines, raised pavement markings (RPM’s) and a variety of post-mounted delineators and chevrons (using both standard and wide spacing). These retroreflective delineation treatments (plus a control condition) are summarized in Table 8.
The controlled field study was conducted on a closed course that could be configured for each of the 12 conditions outlined in Table 8. Experimental participants consisted of 33 younger (mean age: 34.7; mean acuity: 20/20) and 33 older drivers (mean age: 70.2; mean acuity: 20/25). The efficacy of the treatments were assessed using two measures: a static curve recognition distance measure and a dynamic visual occlusion technique. Since the visual occlusion technique resulted in no systematic age differences and large within age group variations, only the results of the recognition distance study are presented. The results of this study are presented in Figure 4.
Table 8.
Delineation Treatments Examined in Field Study of Age Differences in
Nighttime Visibility Distance Conducted by Pietrucha, et al. (1996).
| | | |Relative |
|No. |Treatment Description |Grade |Cost |
|1 |4-in (10.16 cm) yellow centerline (baseline control) |F |- |
| |4-in (10.16 cm) yellow centerline and a 4-in (10.16 cm) | | |
|2 |high-brightness white “profiled” edge line |F |- |
| |4-in (10.16 cm) yellow centerline with widely spaced | | |
|3 |yellow RPM’s |F |- |
| |4-in (10.16 cm) yellow centerline with widely spaced yellow | | |
|4 |RPM’s and widely space white edge line RPM’s |F |- |
| |4-in (10.16 cm) yellow centerline with high-intensity chevrons at standard height| | |
|5 |and spacing |B |1.15 |
| |4-in (10.16 cm) yellow centerline, 4-in (10.16 cm) white edge-line, and | | |
|6 |high-intensity chevrons at standard height and spacing |A |1.52 |
| |4-in (10.16 cm) yellow centerline and standard flat post | | |
|7 |delineators at standard spacing |C |1.63 |
| |4-in (10.16 cm) yellow centerline, white edge line and standard post delineators | | |
|8 | |C |2.00 |
| |4-in (10.16 cm) yellow centerline and fully retroreflectorized flat post | | |
|9 |delineators at standard spacing |A |1.05 |
| |4-in (10.16 cm) yellow centerline and high-intensity T-post delineators at | | |
|10 |standard spacing |A |1.00 |
| |4-in (10.16 cm) yellow centerline with yellow RPM’s at standard spacing and | | |
|11 |high-intensity T-post delineators at standard spacing |A |2.72 |
| |4-in (10.16 cm) yellow centerline, 4-in (10.16 cm) white edge-line and | | |
|12 |engineering-grade T-post delineators at stand. spacing |A |2.03 |
[pic]
Delineation Treatments
Figure 4.
Visibility distance of nighttime delineation treatments as a function of age.
Two criteria will be used to evaluate this pattern of results: the magnitude of the age difference observed for a given treatment and whether or not the observed recognition distance met the 5 s visibility requirements of long-range steering performance (Assuming a top nighttime driving speed of 65 mi/h, 5 sec of visibility distance would correspond to 477 ft). With the exception of the control condition (treatment #1), all of the age differences in curve recognition distance were statistically significant. Six treatments resulted in the greatest curve recognition distances and were statistically indistinguishable from one another: namely, delineation treatments 5, 6, 9, 10, 11 and 12. All of these treatments afforded a visibility distance that exceeded the 5 s criterion (i.e., 477 ft at 65 mi/h) and merit the highest relative rating (“A”). However, since treatment 5 demonstrated a relatively large age difference it is assigned a grade of “B”. Treatments 7 and 8 are statistically different from all the other treatments but not from one another and are assigned a “passing” grade of “C” since the visibility distances afforded by both exceed the 5 sec sight distance criterion established by Rumar and Marsh (1998). Treatment conditions 1, 2, 3 and 4, however, all are assigned a “failing” grade of “F” since they yielded visibility distances well below the criterion level. These treatments would provide 5 s of visibility distance only for drivers traveling at speeds below 30 mi/h. Pietrucha, et al. (1996) also performed a cost-effectiveness analysis for the top performing delineation treatments (see Table 8 for a summary of the relative costs of applying each treatment). On the basis of the performance results as well as the costs of applying the manipulations, they recommended that delineation treatments 10 and 12 be selected as the best treatments to improve the performance of older drivers. However, there is an important difference between treatments 10 and 12. Although much less expensive, treatment 10 contains no edge line and, therefore, may fail to provide adequate support for the short-range lane keeping requirements of steering. Indeed, several of the top performing treatments examined in this study contain no edge lines. Unfortunately, given the “nested” experimental design employed, there is no way of determining the relative contribution of the edge line to any given treatment combination.
4.1.2 Hazard Delineation
Lerner, et al. (1997) performed a series of laboratory and field studies to evaluate the conspicuity of MUTCD object markers used to delineate near off-road hazards (such as ditches, bridge supports, etc). Age differences in the comprehension and visibility of currently prescribed hazard delineation treatments were assessed along with an evaluation of potential conspicuity enhancements for a series of ad hoc manipulations in marker size, color, shape and symbology. These studies were conducted under both day and nighttime viewing conditions across a range of roadway geometry. Although reliable age differences in the viewing distances afforded by the hazard delineators were observed, few of the engineering manipulations explored in the service of increased visibility were found to yield operationally meaningful improvements in performance.
4.1.3 Ultraviolet-Activated Delineation Treatments
Recently, the FHWA has sponsored a series of studies to evaluate the potential effectiveness of using “ultraviolet-activated” materials to improve the effectiveness of nighttime roadway delineation treatments. Experimental high-beam headlamps that emit ultraviolet (UV) light represent little or no glare hazard since UV light is virtually invisible to the human eye. Yet, when this UV energy strikes materials treated with special pigments, the UV energy is transformed into visible light by a process known as fluorescence. Roadway delineation treatments that fluoresce when illuminated by UV headlamps appear to “glow” and are visible from a much greater distance than would be possible under normal low-beam only operation conditions. Mahach, Knoblauch, Simmons, Nitzburg, Arens and Tignor (1997) conducted a preliminary field study to evaluate the potential effectiveness of UV headlamp-fluorescent roadway delineation treatments with 36 observers ranging in age from 25 to 65+. Three 90 m sections of roadway were treated with edge lines constructed from either (1) worn and faded retroreflective white paint, (2) new white thermoplastic material or (3) new white thermoplastic treated with fluorescent photopigments. Observers of all ages rated the fluorescent delineation treatment as significantly more visible than either of the other treatments when the UV headlamps of the test vehicle were activated. Static performance measurements revealed that the UV headlamp-fluorescent delineation combination increased visibility distance by an average of nearly 40 m. A more extensive follow-up study of the visibility of fluorescent delineation treatments was conducted by Turner, Nitzburg and Knoblauch (1998). They measured the visibility distance benefits of UV headlamp-fluorescent delineation treatments under static and dynamic conditions on a closed test track. Static visibility distances afforded by various lane delineation, post-mounted delineation and off-road hazard delineation treatments were measured in 8 young (16-25), 14 middle-aged (26-59) and 6 older (60+) observers. Statistically significant increases in both detection and recognition distance were observed when the fluorescent delineation treatments were activated with UV headlamps. For example, recognition distances for fluorescent-activated thermoplastic lane markings used to delineate right curves, no passing zones and cross walks were improved by 27, 57 and 55 percent, respectively. Similar gains were observed for the recognition of dynamic targets treated with fluorescent-activated retroreflective delineators (e.g., walking or jogging pedestrians). Unfortunately, neither of these two studies of fluorescent-activated delineators reported analyses based upon the age factor. However, there is no a priori reason to believe that fluorescent delineation treatments would not improve visibility performance for older adults; thus, providing an important new approach for providing improved guidance during nighttime driving. Unfortunately, the infrastructure investment required to implement this approach may be so great as to preclude its implementation in the foreseeable future.
4.2 Highway Signing
Highway signs represent one of the most important elements of the surface transportation infrastructure. They are critical for providing route guidance, regulatory information and critical operational and safety warnings and advisories. The perception of highway signs has traditionally been considered within a serial (i.e., sequential) information processing model consisting of four distinct stages: detection, recognition, comprehension and response. The detection process is associated with the conspicuity of a sign whereas the recognition process is usually measured in terms of legibility. Recent research involving these aspects of highway effectiveness are reviewed in the paragraphs which follow.
4.2.1 Sign Conspicuity
Since a sign must first be detected or “noticed” before it can be read and subsequently understood, it is not surprising that much research has been conducted to optimize highway sign conspicuity. Cole and Hughes (1984) have defined two classes of sign conspicuity: Attention conspicuity refers to the capacity of a sign to attract a driver’s attention when the driver is not actively searching for it. Search conspicuity, on the other hand, refers to the capacity of a sign to be quickly and reliably located during a purposive search. It is clear, however, that conspicuity is not merely a property of a sign, per se, but an emergent property of the interaction between a characteristic of a sign, the background upon which it is viewed and the state of the observer. The major factors influencing highway sign conspicuity include: size, brightness, color contrast, background scene complexity, location and design speed of the highway. However, the relative importance of these factors varies across daytime and nighttime viewing conditions (see Paniati & Mace (1993) for a concise review).
Mace, Garvey & Heckard (1994) examined age differences in the conspicuity distance afforded by incremental changes in sign brightness. Fifteen young (less than 40 years old) and 15 older (65+) observers looked for signs along the side of the road while seated in the front passenger seat of an automobile traveling at the posted speed limit of 35 mi/h (56 km/h). Conspicuity was defined as the distance at which the color of a sign was correctly identified. Target signs varied in brightness, size (24 or 36 in)(61 or 91 cm) and background color (orange; white; green). Brightness was varied by manipulating sign retroreflectivity (Type I, II, IV and VII corresponding to engineering grade, medium intensity, high intensity and super-high intensity sheeting materials, respectively). Although nighttime conspicuity increased with size, neither age of the observer nor sign brightness had a systematic effect upon visibility. This outcome is not entirely surprising since gains in sign conspicuity as a function of luminance have been shown to reach asymptote at relatively low luminance levels – suggesting that the luminance levels achieved by the Type I sheeting material had already exceeded the conspicuity asymptote of both the young and older observers in the study. Unfortunately, no photometric values for the stimuli were available in the report. The authors obtained similar results in a parallel study examining the daytime conspicuity of the same stimuli.
The findings reported above suggest that the additional luminance made available by high intensity (Type IV) and super-high intensity (Type VII) retroreflective materials may have little impact upon highway sign conspicuity. There is some evidence, however, that the optimal brightness level for sign conspicuity may be markedly elevated under conditions of high background scene complexity (Olson, 1988). This might be especially the case among older drivers who appear to be disproportionately affected by background visual “clutter” (Kaussler, 1990). In an attempt to examine this possibility, Schieber and Goodspeed (1997) used a nighttime scene simulator to assess age differences in the glance conspicuity of highway signs as a function of sign brightness and background visual complexity. Sign luminance was manipulated to simulate the levels afforded by engineering grade versus super-high intensity retroreflective sheeting materials. Young (mean age: 31.8; range: 22-44) and older (mean age: 71.5; range: 61-80) observers were required to detect and report the location of a highway sign embedded in a background road scene that was presented for 250 msec. At low levels of background scene complexity there were no age differences in the time needed to detect and localize the target sign regardless of its brightness. However, when the sign was embedded in a complex background driving scene significant increases in the time needed to detect the target were observed – especially in the case of the older observers. This change in sign conspicuity resulting from the introduction of the complex background was significantly reduced, however, when target sign luminance was increased to a level approximating the brightness achieved by super-high intensity sheeting materials. Hence, it would appear that age-related increases in the deleterious effects of background visual clutter (typically experience in densely populated urban areas) might be partially offset by the strategic application of highway signs constructed from super-high intensity retroreflective sheeting materials.
Recent experimental work has suggested that state-of-the-art developments in sheeting materials might also contribute to increasing the daytime conspicuity of highway signs.
Durable fluorescent sheeting materials have been introduced with the high levels of retroreflectivity needed for nighttime visibility as well as improved brightness and color contrast during daytime conditions. Durable fluorescent pigments (unlike the UV headlamp activated pigments discuss above) achieve their unique color appearance by converting the short, but visible, wavelength light from the sun (usually absorbed and converted to heat) into longer wavelength light. It is this recruitment of short wavelength light that fosters the “brighter than bright” appearance of fluorescent materials (see Burns & Pavelka, 1995). Preliminary reports suggest that the use of fluorescent orange signs at a North Carolina construction site resulted in significant improvements in driver behavior (e.g., merging ahead of a lane drop) relative to conventional non-fluorescent orange signs (Hummer & Scheffler, 1998). Jenssen, Moen, Brekke, Augdal & Sjohaug (1996) conducted a study to determine the distance at which volunteer observers could detect and recognize the shape, color and contents of fluorescent versus non-fluorescent signs. Young and older participants searched for signs while seated on an open-platform railway car which traveled at a speed of 15 km/h along a 4 km straight track. The color of the sign targets employed were yellow, yellow-green and orange (both fluorescent and non-fluorescent). The detection distance afforded by the fluorescent signs was significantly greater than that demonstrated for their non-fluorescent counterparts. Young observers demonstrated a 79 m increment in detection distance for fluorescent signs while older observers demonstrated a 92 m advantage. This age difference in the size of the fluorescent sign conspicuity benefit was statistically significant. Hence, there is clear evidence that the strategic deployment of fluorescent highway signs may be particularly helpful for older drivers. What remains to be discovered, however, is the mechanism by which fluorescent materials improve the behavioral salience of highway signs. That is: Do fluorescent signs improve driving behavior because they increase attentional conspicuity, search conspicuity or both?
4.2.2 Sign Legibility
Survey studies of driver visual capacities and complaints have consistently revealed that older adults experience considerable difficulties reading traffic and street-name signs (e.g., Kline, Kline, Fozard, Kosnik, Schieber & Sekuler, 1992). This is not surprising given the characteristic declines in acuity and contrast sensitivity that accompany normal adult aging (see Schieber, 1992). Recent studies have examined a number of design factors that can be manipulated to mitigate age differences in highway sign legibility distance, including: letter size/height, brightness and letter font characteristics.
4.2.2.1 Letter Height
In his detailed review of the literature in TRB (1988) Special Report 218, Mace (1988) demonstrated that the specification of letter size for use on highways signs in the U.S. clearly fails to meet the legibility requirements of a large percentage of older road users. This de facto standard for the specification of highway signs is the so-called 50:1 legibility index which traces its history back to the classic work of Forbes and his colleagues (Forbes & Holmes, 1939; Forbes, Moskowitz & Morgan, 1950). The 50:1 legibility index specifies that for every 50 ft of legibility distance required of a sign one needs to add 1 inch to the height of its letters. Hence , if a given situation demanded that a sign be readable from a distance of 400 ft then the 50 ft/in (6 m/cm) legibility index would specify a required letter height of 8 inches. Even under ideal observation conditions a driver must have a visual acuity of 20/23 or better to be capable of reading letter heights specified by the 50 ft/in rule. Epidemiological data indicates that approximately half of the persons 65-75 years of age would fail to achieve this level acuity, even when wearing their eyeglasses (Schieber, 1992). Mace (1988) suggests that the traditional 50 ft/in (6m/cm) legibility index be changed to a specification of 40 ft/in (4.8 m/cm) which presumes a visual acuity of approximately 20/30. This proposed modification would fall short of accommodating all drivers since the typical minimum acuity level used in the licensing of drivers in the U.S. is 20/40 (Bailey & Sheedy, 1988). However, the 40 ft/in (4.8 m/cm) design specification would accommodate the visual acuity levels of 85 percent of those 65-75 years of age (Schieber, 1992). Empirical data reviewed by Mace also supported this recommendation. For example, Olson & Bernstein (1979) found that older drivers in their sample would be accommodated by the 40 ft/in rule provided that signs had an internal contrast of at least 5:1 and a luminance of 10 cd/m2 or better. In support of this recommendation, federal rule making has been initiated to modify the legibility index from the 50 ft/in (6m/cm) specification to a more accommodating value of 40 ft/in (4.8 m/cm). In addition, the FHWA has adopted a change in the MUTCD that would increase the required height of letters used in street-name signs from the current 4 in (10 cm) to a more appropriate value of 6 in (15 cm) (Federal Highway Administration, 1997).
Recent work appears consistent with the recommended adoption of the 40 ft/in legibility index specification. Chrysler, Danielson and Kirby (1996) measured legibility distances for 8-inch (20 cm) Landolt ring stimuli mounted upon 24-inch (61 cm) signs. Young (mean age = 20.5) and older (mean age = 65.6) adults seated in the passenger seat of an automobile traversing a closed course demonstrated mean legibility distances of 467 ft (142 m) and 320 ft (97.5 m), respectively. The equivalent legibility index needed to accommodate the average older driver was 40 ft/in (i.e., 320 ft / 8 in = 40 ft/in). Another recent field study of legibility distance for highway signs also obtained results supporting the adoption of a less demanding legibility index specification. Hawkins, Picha, Wooldridge, Greene and Brinkmeyer (1999) reported 85th percentile daytime legibility distances of 55 ft/in (6.6 m/cm) for young (less than 40), 40 ft/in (4.0 m/cm) for middle-aged (55-64) and 32 ft/in (3.84 m/cm) for older (65+) drivers viewing white-on-green signs constructed using 16-in tall Series E letters.
However, the simple linear relationship between letter size and the legibility distance afforded by a highway sign implied by the legibility index has recently been challenged by Mace, Garvey & Heckard (1994). Daytime legibility distance thresholds were collected from 15 young (16-40 years of age) and 15 older (65+) drivers for Series C and D letters presented in black-on-white format as well as for Series E letters presented in white-on-green format. Legibility distances were determined for 5 different letter heights: 6, 8, 10, 12 and 16 inches (15.24, 20.32, 30.48 and 40.64 cm). Legibility distance increased as a function of letter height. However, the nature of this relationship was not the simple linear function implied by design specifications such as the legibility index. Their data for white-on-green Series E letters is depicted in Figure 5. The linear relationship implied by the legibility index appears to break down at approximately 600 ft (183 m) - corresponding to a letter of height of between 8-10 in for the young observers and a letter height of approximately 12 in for the older observers. Increases in letter heights beyond these critical values result in legibility gains significantly below those predicted by the legibility index formula. These findings are anomalous but potentially of critical importance with respect to highway design interventions aimed at accommodating the changing visual needs of aging drivers. The suboptimal performance of highway signs with very large letters in this particular case could have been due to “crowding effects” resulting from insufficient spacing between the letters used to construct the target stimuli.
[pic]
Figure 5. Sign Legibility Distance as a Function of Letter Height and Age.
Note the divergence from the ideal linear function (from Mace, et al., 1994)
4.2.2.2 Font Characteristics
As discussed briefly above, highway signs constructed from state-of-the-art high brightness materials can sometimes suffer from a visibility-reducing condition known as irradiation or halation. This problem is particularly noteworthy for white-on-green guide signs using the traditional Series E “bold” character font. Under some viewing geometries portions of the large, wide stroke width letters can appear to “glow” and “blur into one another”. Spaces within and between the letters appear to “fill in” due to this halation effect and legibility is subsequently reduced. Susceptibility to the effects of irradiation glare is believed to increase with advancing adult age (Schieber, 1994b). Mace, Garvey and Heckard (1994) modified the traditional highway sign text font so that the interstitial spaces within critical letters were increased in area. By increasing the size (area) of the critical gaps within certain letters they hoped to develop a text font that would be less susceptible to irradiation effects. Their initial work yielded signs that resisted irradiation effects at night but that suffered from reduced legibility distance during daytime viewing.
Given this promising starting point, Garvey, Pietrucha and Meeker (1997; 1998) conducted a more extensive follow-up investigation in which they used reiterative design procedures (i.e., the recursive blur technique) to develop an entire text font that was optimized for both nighttime and daytime viewing conditions. The new-generation highway sign alphabet they developed is known as the Clearview font (see Figure 6). Garvey and his associates then conducted a field study to determine the effectiveness of their new font. Legibility distance data were collected from a group of 48 older (ages 65-83) observers for signs constructed using either the Clearview or traditional Series E text font. Results of the study indicated that the nighttime legibility distances achieved by older drivers increased by an average of 17 m when Clearview letters were used to replace Series E letters (101 versus 118 m). Unlike the initial work of Mace, et al. (1994), no decline in daytime legibility distance was observed for the Clearview font. This font was specifically designed to improve the legibility of high-brightness highway signs at night. Under these conditions, Clearview resulted in a 16% increase in the recognition distance afforded to older drivers.
[pic]
Series E (Modified) Font
[pic]Clearview Font
Figure 6. Series E (Modified) versus Clearview Fonts used for
Large Freeway Signs. Note that the interstitial spaces in the Clearview characters
are much larger – minimizing the negative effects of “over glow” or “halation”
when viewed under nighttime driving conditions.
(from Hawkins, Picha, Wooldridge, Greene and Brinkmeyer, 1999)
Hawkins, Picha, Wooldridge, Greene and Brinkmeyer (1999) examined the relative legibility of white-on-green guide signs constructed using 16-in (40 cm) Series E, British Transport Medium and Clearview text fonts. A field study of young ( 157 |> 200 |
|Nighttime: |-- |25 < I 314 |> 200 |
|Nighttime: |-- |25 < I 726 |> 200 |
|Nighttime: |-- |25 < I 3000 cd/m2 |
Only about one-third of the light reaching the retina of a typical 20 year-old observer will be transmitted through the eye of the typical 75 year-old driver. Under some circumstances (dim illumination; green-blue colored targets) this age-related reduction in retinal illumination may be as great as a factor of 10 (Schieber, 1988). This clearly suggests that the emerging visual requirements of older drivers need to be considered in any effort to generate standards or guidelines for traffic signal lights. However, neither the ITE (1985) nor the CIE (1988) specifications thoroughly consider the potential difficulties with traffic signal visibility that may be encountered by older drivers. Recent research suggests that some aspects of these guidelines may fall short of accommodating the needs of older drivers. An extensive analytic review conducted by Freedman, et al. (1997) concluded that the minimum daytime brightness levels for red traffic signals specified by the ITE as well as for green and yellow signals by the CIE may be too low to accommodate the needs of older drivers. An examination of Table 10 reveals that the intensity values in question fall short of Freedman, et al.’s determination of 200, 265 and 600 cd, respectively, for red, green and yellow traffic signal lights. Although the ITE (1985) standard remains silent regarding the issue of nighttime intensity levels, the CIE (1988) guidelines recommend a minimum nighttime intensity of 25 cd and a maximum nighttime intensity of 200 cd. However, the CIE guidelines specify that a range of nighttime intensities between 50-100 cd should be maintained whenever possible. Performance data collected from a sample that included representative older drivers suggest that even the more conservative nighttime minimum intensity of 50 cd may be too low. Freedman, et al. (1985) recommended that the minimum nighttime intensity levels required to accommodate the needs of older drivers were 50, 95 and 220 cd for red, green and yellow traffic signal lights, respectively. Little experimental work has been done to determine maximum nighttime intensity levels, although there is some indirect evidence that the CIE (1988) recommendation of 200 cd may be associated with excessive levels of glare for older drivers stopped at a signalized intersection. Yet, despite the observations summarized in this paragraph, it is interesting to note that all of the survey studies of driver visual complaints cited above failed to reveal self-reported difficulties with traffic signal lights.
Currently, the National Cooperative Highway Research Program (NCHRP) is sponsoring a comprehensive study entitled Visibility Performance Requirements for Vehicular Traffic Signals (see Freedman, et. al., 1997). This investigation includes a series of laboratory and field studies to determine performance-based requirements for traffic signal intensity, intensity distribution, and related photometric parameters using a subject population that over-samples representative older drivers. The results of this study, when completed, will be translated into a model standard that hopefully will be adopted in future releases of the ITE and CIE specifications (both of which are currently being revised).
5.0 Highway Lighting
Although the number of vehicle miles driven at nighttime are much fewer than those during the daytime, over half of all traffic fatalities occur at night. On a per mile basis, the nighttime fatality rate is approximately three times greater than that experienced during daytime driving (National Safety Council, 1993). Numerous studies have shown that well-designed fixed-lighting installations on roadways can significantly reduce nighttime driving accidents, especially those not directly related to alcohol intoxication (Commission Internationale de l’Éclairage, 1992). Given that normal aging is associated with large declines in nighttime visual functioning, it would appear that older drivers would especially benefit from improvements in highway illumination (see Schieber, 1994b). However, despite the fact that numerous studies have identified nighttime illumination as a critical area for research and development (see Schieber, 1994a), little progress in this area has been made since the release of TRB (1988) Special Report 218.
Perhaps one of the reasons that little work has been conducted to study the nighttime illumination requirements of older drivers has been that, at present, they do relatively little driving at night. Data from the 1990 National Personal Transportation Study depicted in Figure 11 reveal that the percent of vehicle trips completed between dusk and dawn (7 PM - 6 AM) declines from 24.9 percent at age 20-24 all the way down to approximately 2 percent for those 85 years of age or older. Mortimer and Fell (1989) reported that drivers 65 and older accounted for only 1 percent of all driver fatalities between midnight and 6 AM. Numerous survey studies also report that older drivers intentionally reduce the amount of driving they do at night (e.g., Kline, et al., 1992). This voluntary reduction in nighttime driving exposure is often interpreted as an adaptive behavior that compensates for marked age-related reductions in the ability to see at night.
[pic]
Figure 11. Percentage of vehicle trips taken at night as a function of age – based upon the 1990 National Personal Transportation Study (Hu, et al., 1993)
The Federal Highway Administration recently initiated a contract-research project titled Night Driving and Highway Lighting Requirements for Older Drivers. The objectives of this initiative include an analytic study (based upon the Small Target Visibility model and related techniques) aimed at uncovering lighting practices that will optimize the safety and mobility of older drivers. Special emphasis will be placed upon improved illumination engineering designs that improve the contrast of objects on the roadway (i.e., potential hazards) while minimizing the opportunity for glare. The planned project includes a controlled field study designed to validate the design principles resulting from the analytic study. Results from the study should be available in the Spring of 2001.
Projects like the FHWA initiative noted above, however, may be somewhat premature as researchers still know very little about how the effects of age-related visual changes impact nighttime driving performance. It may be difficult to “fix” the nighttime visual problems of older drivers before identifying the specific nature of those problems. Instead, several preliminary lines of research are needed in the area of aging and nighttime illumination of highways: First, an analytic study is needed to predict the future mobility needs of older drivers during nighttime hours. It appears likely that the next generation of drivers to grow old may want (need) to drive more frequently at night. If this turns out to be true, considerable efforts will need to be made to improve the nation’s highway lighting infrastructure. Next, a comprehensive set of observational field studies is needed to see if the nighttime mobility of older drivers changes as a function of the quantitative and qualitative level of existing highway lighting installations. Such a study could provide the basis for determining the extent to which improved lighting might contribute to the mobility of mature drivers. Finally, we need to determine the effects of low-levels of roadway illumination upon basic driving behaviors such as speed, lane keeping, and turning efficiency as well as the subjective experience of workload and personal safety.
6.0 Summary Recommendations
As a way of summarizing the previous sections of this report, two list of recommendations for (1) immediate implementation and (2) future research needs are compiled below. These lists are presented in descending order of priority. Appearance upon one of the lists is justified by supporting evidence presented in the preceding review of the research literature. However, it should be recognized that the basis for the author’s prioritization scheme is highly biased by personal experience and his limited domains of expertise.
6.1 Recommendations for Immediate Implementation and Practice
1) The 85th percentile 75-year-old design driver should be adopted for highway research, design and decision-making purposes.
2) A 3 ft/sec pedestrian walking speed should replace the 4 ft/sec value implicit in the Manual of Uniform Traffic Control Devices. Such a proposal has, in fact, been approved for inclusion in the Millenium Edition of the MUTCD.
3) The 40:1 legibility index for letter height on highway signs should be fully incorporated into the Manual of Uniform Traffic Control Devices and related state and local guidelines. Such a proposal has been approved for the 2000 edition of the MUTCD.
4) A standing review panel should be setup and funded to continuously evaluate the theoretical as well as the practical merits of the recommendations contained in the Older Driver Highway Design Handbook. This panel could maintain the handbook as a “living” document as well as be highly instrumental in identifying areas where new research efforts are needed to support the mobility and safety of older road users.
5) There should be continued support and application of the ASHTO Intersection Sight Distance and Stopping Sight Distance models for highway geometric design as these guidelines appear to accommodate the needs of the 85th percentile 75-year-old driver.
6) State-of-the-art high brightness retroreflective materials should be recommended for highway signs in unlighted or highly cluttered urban areas.
7) The U.S. DOT should provide direct support for implementing the change over from 4-inch to 6-inch letter heights for street name signs as now specified in the Manual of Uniform Traffic Control Devices.
8) Encourage the increased use of durable fluorescent signs. Fluorescent orange signs should be considered for universal deployment in highway work zones. Fluorescent Yellow-Green signs should be widely deployed for Pedestrian and School Zone warning signs as now recommended in the Manual of Uniform Traffic Control Devices. Finally, the use of fluorescent yellow warning signs should be encouraged for deployment in high-density and/or high-speed traffic situations.
9) Encourage greater use of symbol versions of highway warning signs that have been shown to yield both high comprehension levels and superior legibility distances (relative to their text sign counterparts).
10) Encourage the voluntary application of the Clearview font for new and replacement street name signing applications. Support this effort through the dissemination of downloadable versions of the font that are compatible with commonly deployed computer graphics applications programs.
6.2 Research Recommendations
1) Develop performance-based requirements for highway sign legibility and use these parameters to establish and validate minimum retroreflectivity requirements for these devices. Document cost-benefit analyses and tradeoff studies used to justify any specifications that fail to fully meet the performance-based requirements of the design older driver.
2) Develop performance-based requirements for roadway delineation and use these parameters to establish and validate minimum retroreflectivity requirements for roadway markings. Document cost-benefit analyses used to justify specifications that fail to meet the performance-based requirements. The ability to accommodate the increased visual requirements of the design older driver must be explicitly addressed in any such rule making activities.
3) Carefully evaluate the conceptual and methodological basis of gap acceptance models for highway geometric design relative to the characteristics of the 85th percentile 75-year-old design driver.
4) Conduct studies to ascertain and model how older and younger drivers scan their visual environment (using state-of-the-art eye tracking techniques). Determine what visual information drivers use and when do they need it. This information will provide invaluable input for developing a system-level approach to providing information to drivers via traffic control devices as well as ITS in-vehicle technologies.
5) Older drivers appear to report as well as demonstrate problems merging into high-speed and/or high-density highway traffic. Research is needed to ascertain how geometric design, traffic control devices and/or ITS can be modified to accommodate these problems.
6) Evaluate the impact of Small Target Visibility (STV) models of highway lighting design upon performance, comfort and fatigue among older drivers.
7) Investigate the magnitude and extent to which driver fatigue may disproportionately influence performance in the older population.
8) Evaluate the performance and workload demands of proposed traffic calming techniques (such as lane narrowing and roundabouts) upon the oldest drivers prior to the development of new traffic calming regulations and guidelines.
9) Develop a system-level approach to the design, placement and maintenance of highway signs (Bigger, brighter may not represent the optimal solution; instead, we need to learn how to utilize principles of information redundancy, complementarity and consistency).
10) Model and evaluate the dramatic reluctance of older drivers to execute right-turn-on-red operations in urban areas upon traffic capacity.
11) Evaluate the impact of various levels of street lighting upon the current mobility of older drivers. Conduct cost/benefits analyses to determine if increased investments in highway lighting infrastructure will be needed to accommodate the rapidly increasing proportion of (older) drivers experiencing nighttime visibility problems.
12) Further study and quantify the role of durable fluorescent colors in improving the conspicuity and legibility of highway signs among older drivers. The safety and mobility impact of these signs should be especially robust in work zone areas as well as high-density urban traffic situations.
8.0 References
Allen, R.W., O’Hanlon, J.F. & McRuer, D.T. (1977). Driver visibility requirements for roadway delineation. Volume 1. Effects of contrast and configuration on driver performance and behavior. Report No. FHWA-RD-77-165. Washington, DC: Federal Highway Administration.
American Association of State Highway and Transportation Officials (AASHTO)(1994).
A policy on geometric design of highways and streets. Washington, DC: AASHTO.
Bailey, I.L. & Sheedy, J.E. (1988). Vision screening for driver licensure. In Transportation Research Board, Transportation in an aging society: Improving mobility and safety for older persons. Special Report 218. Volume 2. Washington, DC: National Research Council. pp. 294-324.
Bailey, S.S., Jones, S.A., Stout, R.J., Bailey, J.H., Kass, S. & Morgan, B.B. (1992). Issues of elderly pedestrians. Transportation Research Record, 1375, 68-73.
Benekohal, R.F., Resende, P.T.V., Shim, E., Michaels, R.M. & Weeks, B. (1992). Operations problems of elderly drivers in Illinois. Springfield, IL: Illinois Department of Transportation. 134 pp.
Blaauw, G.J. & Padmos, P. (1982). Nighttime visibility of various types of pavement markings: A study on durability, including conditions of rain, fog and dew. SAE Technical Report No. 820412. Warrendale, PA: Society of Automotive Engineers.
Blomberg, R.D. & Cleven, A.M. (1998). Development, implementation and evaluation of a pedestrian safety zone for elderly pedestrians. DOT-HS-808-692. National Highway Traffic Safety Administration, U.S. Department of Transportation, Washington, DC
Bonneson, J.A. & McCoy, P.T. (1994). Driver understanding of protected and permitted left-turn signal displays. Transportation Research Record, 1464, 42-50.
Burns, D.M. & Pavelka, L.A. (1995). Visibility of durable fluorescent materials for signing applications. Color Research and Application, 20, 108-116.
Chiu, S.A., Mourant, R. & Bond, R.A. (1997). A study of driver behavior in highway construction zones with respect to age and illumination conditions. Paper No. 970847. 76th Annual Meeting of the Transportation Research Board, Washington, DC.
Chrysler, S.T., Danielson, S.M. & Kirby, V.M. (1996). Age differences in visual abilities in nighttime driving field conditions. Proceedings of the Human Factors and Ergonomics Society 40th Annual Meeting, Philadelphia, PA. pp. 923-927.
Coffin, A, & Morrall, J. (1995). Walking speeds of elderly pedestrians at crosswalks. Transportation Research Record, 1487, 63-67.
Cole, B.L. & Hughes, P.K. (1984). A field trial of attention and search conspicuity. Human Factors, 26, 299-313.
Comité Européen de Normalisation (1997). Pavement marking materials: Pavement marking performance for road users. Report No. prEN 1436. Brussels, Belgium: Comité Européen de Normalisation.
Commission Internationale de l’Éclairage (1988). A guide for the design of road traffic lights. CIE Publication No. 79. Vienna, Austria: CIE Central Bureau.
Commission Internationale de l’Éclairage (1992). Road lighting as an accident countermeasure. CIE Publication No. 93. Vienna, Austria: CIE Central Bureau.
Deacon, J.A. (1988). Potential improvements in roadway delineation for older drivers. In Transportation Research Board, Transportation in an aging society: Improving mobility and safety for older persons. Special Report 218. Volume 2. Washington, DC: National Research Council. pp. 253-269.
Dewar, R.E. (1992). Driver and pedestrian characteristics. In J.L. Pline (Ed.), ITE Traffic engineering handbook. (4th edition). Englewood Cliffs, NJ: Prentice-Hall. pp. 1-28.
Dewar, R., Kline, D., Schieber, F. & Swanson, A. (1994). Symbol signing design for older drivers [Report No. FHWA-RD-94-069]. McLean, VA: Federal Highway Administration.
249 pp.
Dewar, R. & Olson, P. (in press). Human Factors in Traffic Safety. Tucson, AZ: Lawyers & Judges Publishing Company.
Ethen, J.L. & Woltman, H.L. (1986). Minimum retroreflectance for nighttime visibility of pavement markings. Transportation Research Record, 1093, 43-47.
Fambro, D.B., Fitzpatrick, K. & Koppa, R.J. (1997). Determination of stopping sight distance. NCHRP Report 400. Washington, DC: National Academy Press. 83 pp.
Federal Highway Administration (1997). National standards for traffic control devices; Revision of the Manual of Uniform Traffic Control Devices. Federal Register, 62(6), 1364-1373.
Forbes, T.W. & Holmes, R.S. (1939). Legibility distances of highway destination signs in relation to letter height, width and reflectorization. Highway Research Board Bulletin, 19, 321-326.
Forbes, T.W., Moskowitz, K. & Morgan, G. (1950). A comparison of lowercase and capital letters for highway signs. Highway Research Board Bulletin, 30, 355-373.
Freedman, M., Davit, P.S., Staplin, L.K. & Breton, M.E. (1985). Traffic signal brightness: An investigation of nighttime dimming. Final report. FHWA-RD-88-005. Washington, DC: U.S. Department of Transporation.
Freedman, M., Staplin, L.K., Gilfillan, D.P., & Byrnes, A.M. (1988). Noticeability requirements for delineation on nonilluminated highways. Report No. FHWA-RD-88-028. Washington, DC: Federal Highway Administration.
Freedman, M., Flicker, L., Janoff, M.S., Schwab, R.N. & Staplin, L.K. (1997). Visibility performance requirements for vehicular traffic signals. Interim report (NCHRP Project No. 5-15). Rockville, MD: Westat, Inc.
Garber, N.J. & Srinivasan, R. (1991). Risk assessment of elderly drivers at intersections: Statistical modeling. Transportation Research Record, 1325, 8-16.
Garvey, P.M., Pietrucha, M.T. & Meeker, D.T. (1997). Effects of font and capitalization on legibility of guide signs. Transportation Research Record, No. 1605, 73-79.
Garvey, P.M., Pietrucha, M.T. & Meeker, D.T. (1998). Clearer road signs ahead. Ergonomics in Design, 6, 7-11.
Graham, J.R., Fazal, A. & King, L.E. (1996). Highway sign minimum luminance requirements for older drivers. Transportation Research Board Symposium on Night Visibility and Driver Behavior, Iowa City, IA. April 28-30.
Harwood, D.W., Mason, J.M., Byrdia, R.E., Pietrucha, M.T. & Gittings, G.L. (1996). Intersection sight distance. NCHRP Report 383. Washington, DC: National Academy Press. 107 pp.
Hasher, L. & Zacks, R.T. (1979). Automatic and effortful processes in memory. Journal of Experimental Psychology: General, 108, 356-388.
Hawkins, H.G., Picha, D.L., Wooldridge, M.D., Greene, F.K. & Brinkmeyer, G. (1999). Performance comparison of three freeway guide sign alphabets. Paper presented at the 78th Annual Meeting of the Transportation Research Board, Washington, DC.
Hofstetter, R., McGee, H., Crowley, K., Sequin, E. & Dauber, G. (1986). Improved perception-reaction time information for intersection sight distance. FHWA-RD-87-015. McLean, VA: Federal Highway Administration.
Hoxie, R.E. and Rubenstein, L.Z. (1994). Are older pedestrians allowed enought time to cross intersections safely? Journal of the American Geriatrics Society, 42, 241-244.
Hu, P.S., Young, J.R. & Lu, A. (1993). Highway crash rates and age-related driver limitations: Literature review and evaluation of data bases. ORNL/TM-12456. Oak Ridge, TN: Oak Ridge National Laboratory. 114 pp.
Hummer, J.E. & Scheffler, C.R. (1998). Driver performance comparison of fluorescent orange to standard orange work zone traffic signs. 78th Annual Meeting of the Transportation Research Board, Washington, DC.
Institute of Transportation Engineers (1985). Vehicle control traffic signal heads: A standard of the Institute of Transportation Engineers. Washington, DC: Institute of Transportation Engineers.
Institute of Transportation Engineers (2000). Roundabouts: An informational guide. ITE Journal, 70, 2.
Jacobs, R.J., Johnston, A.W. & Cole, B.L. (1975). The visibility of alphabetic and symbolic traffic signs. Australian Road Research, 5, 68-86.
Jenssen, G.D., Moen, T., Brekke, B., Augdal, A. & Sjohaug, K. (1996). Visual performance of fluorescent retroreflective traffic control devices. Part 1: Human factors visibility study. Report No. STF22-A96606. Trondheim, Norway: SINTEF Transport Engineering [ISBN 82-595-9760-8].
Johansson, G. & Rumar, K. (1971). Drivers’ brake reaction time. Human Factors, 13, 23-27.
Josua, S.C. & Saka, A.A. (1992). Mitigation of sight-distance problem for unprotected left-turning traffic at intersections. Transportation Research Record, 1356, 73-79.
Kline, T.J.B., Ghali, L.M., Kline, D.K. & Brown, S. (1990).. Visibility distance of highway signs among young, middle-aged and older observers: Icons are better than text. Human Factors, 32, 609-619.
Kline, D.W., Kline, T.J.B., Fozard, J.L., Kosnik, W., Schieber, F. 7 Sekuler, R. (1992). Vision, aging and driving: The problems of older drivers. Journal of Gerontology, 47, P27-P34.
Kline, D.W. & Fuchs, P. (1993). The visibility of symbolic highway signs can be increased among drivers of all ages. Human Factors, 35, 25-34.
Kloeppel, E., Peters, R.D., James, C., Fox, J.E. & Alicandri, E. (1995). Comparison of older and younger driver responses to emergency driving events. FHWA-RD-95-056. Office of Safety and Traffic Operations R&D, Federal Highway Administration, McLean, VA. 16 pp.
Knoblauch, R., Nitzburg, M., Dewar, R., Templer, J. & Pietrucha, M. (1995a). Older pedestrian characteristics for use in highway design. FHWA-RD-93-177. Office of Safety and Traffic Operations R & D, Federal Highway Administration, McLean, VA.
Knoblauch, R., Nitzburg, M., Reinfurt, D., Council, F., Zeeger, C. & Popkin, C. (1995b). Traffic operations control for older drivers. FHWA-RD-94-119. Office of Safety and Traffic Operations R & D, Federal Highway Administration, McLean, VA 101 pp.
Knoblauch, R., Nitzburg, M. & Seifert, M. (1997). An investigation of older driver freeway needs and capabilities. FHWA-RD-95-194. McLean, VA: Federal Highway Administration. 214 pp.
Lerner, N.D., Huey, R.W., McGee, H.W. & Sullivan, A. (1995). Older driver perception-reaction time for intersection sight distance and object detection. Final Report. Volume 1. FHWA-RD-03-168. McLean, VA: Federal Highway Administration. 103 pp.
Lerner, N.D., Benel, D.C.R., Huey, R.W. & Steinberg, G.V. (1997). Delineation of hazards for older drivers. Volume I. FHWA-RD-96-161. Office of Safety and Traffic Opertions R & D, Federal Highway Administration, McLean, VA. 152 pp.
Lockwood, I.M. (1997). ITE traffic calming definition. ITE Journal, 67, 22-24.
Mace, D.J. (1988). Sign legibility and conspicuity. In Transportation Research Board, Transportation in an aging society: Improving mobility and safety for older persons. Special Report 218. Volume 2. Washington, DC: National Research Council. pp. 270-293.
Mace, D.J., Garvey, P.M. & Heckard, R.F. (1994). Relative visibility of increased legend size vs. brighter materials for traffic signs. FHWA-RD-94-035. McLean, VA: Federal highway Administration. 49 pp.
Mahach, K.R., Knoblauch, R.L., Simmons, C.L., Nitzburg, M., Arens, J.B. & Tignor, S.C. (1997). Safety evaluation of ultraviolet-activated fluorescent roadway delineation. FHWA-RD-97-033. McLean, VA: Federal Highway Administrtaion. 21 pp.
McGee, H.W. & Taori, S. (1998). Impacts on state and local agencies for maintaining traffic signs within minimum retroreflectivity guidelines. Report No. FHWA-RD-97-053. McLean, VA: Federal Highway Administration. 44 pp.
Morales, J.M. (1987). Retroreflective requirements for traffic signs: A stop sign case study. ITE Journal, 57.
Mortimer, R.G. & Fell, J.C. (1989). Older drivers: Their night fatal crash involvement and risk. Accident Prevention and Analysis, 21, 273-282.
National Safety Council (1993). Accident Facts. Chicago, IL: National Safety Council.
Olson, P.L. (1988). Minimum requirements for adequate nighttime conspicuity of highway signs. (UMTRI-88-8). Ann Arbor, MI: University of Michigan, Transportation Research Institute.
Olson, P.L. & Bernstein, A. (1979). The nighttime legibility of highway signs as a function of their luminance characteristics. Human Factors, 21, 145-160.
Paniati, J.F. & Mace, D.J. (1993). Minimum retroreflectivity requirements for traffic signs. (Report No. FHWA-93-077). McLean, VA: Federal Highway Administration. 67 pp.
Pietrucha, M.T., Hostetter, R.S., Staplin, L. & Obermeyer, M. (1996). Pavement markings and delineation for older drivers. Volume I. Final Report. FHWA-RD-94-145. McLean, VA: Office of Safety and Traffic Operations R&D, Federal Highway Administration. 148 pp.
Reilly, W.R., Pfefer, R.C., Michaels, R.M., Polus, A. & Schoen, J.M. (1989). Speed change lanes. Arlington, VA: JHK & Associates. 84 pp.
Rumar, K.. & Marsh II, D.K. (1998). Lane markings in night driving: A review of past research and of the present situation. Report No. UMTRI-98-50. Ann Arbor, MI: University of Michigan, Transportation Research Institute. 83 pp.
Schieber, F. (1987). Optimizing the visual environment to the older eye: 2D-Fourier models of spatial vision in action. The Gerontologist, 27, p. 166A.
Schieber, F. (1988). Vision assessment technology and screening older drivers: Past practices and emerging techniques. In Transportation Research Board, Transportation in an aging society: Improving mobility and safety for older persons. Special Report 218. Volume 2. Washington, DC: National Research Council. pp. 325-378.
Schieber, F. (1992). Aging and the senses. In J.E. Birren, R.B. Sloan and G. Cohen (Eds.), Handbook of mental health and aging. New York: Academic Press. pp. 251-306.
Schieber, F., Kline, D.W. & Dewar, R. (1994). Optimizing symbol highway signs for
older drivers. In Proceedings of the 12th Triennial Congress of the International Ergonomics Association. Volume 6. pp. 199-202.
Schieber, F. (1994a) High priority research and development needs for maintaining the safety and mobility of older drivers. Experimental Aging Research, 20, 35-43.
Schieber, F. (1994b). Recent developments in vision, aging and driving: 1988-1994. Report No. UMTRI-94-26. Ann Arbor, MI: Transportation Research Institute, University of Michigan. 69 pp.
Schieber, F. & Goodspeed, C.H. (1997). Nighttime conspicuity of highway signs as a function of sign brightness, background complexity and age of observer. Proceedings of the Human Factors and Ergonomics Society, Santa Monica, CA: Human Factors and Ergonomics Society. pp. 1362-1366.
Schieber, F. (1998). Optimizing the legibility of symbol highway signs. In A. Gale (Ed.), Vision in vehicles. VI. Amsterdam: Elsevier Science Publishers. pp. 163-170.
Schieber, F. (2000). What do driving accident patterns reveal about age-related changes in visual information processing? In K.W. Schaie & M. Pietrucha (Eds.), Mobility and transportation in the elderly. New York: Springer. pp. 207-211.
Schlabbach, K. (1997). Traffic calming in Europe. ITE Journal, 67, 38-40.
Simmons, C.J. & Paniati, J.F. (1995). Developing minimum retroreflectance values for in-service traffic signs. Compendium of Technical Papers, pp. 597-600. Washington, DC: Institute of Transportation Engineers.
Sivak, M. & Olson, P.L. (1983). Optimal and replacement luminance of traffic signs: A review of applied legibility research. Report No. UMTRI-83-37. Ann Arbor, MI: University of Michigan, Transportation Research Institute.
Smiley, A. (1999). Contribution of human factors to highway safety. TR News, 201, 11-14.
Staplin, L., Lococo, K. & Byington, S. (1998). Older driver highway design handbook. FHWA-RD-97-135. McLean, VA: Federal Highway Administration. 261 pp.
Staplin, L., Lococo, K., Sim, J. & Drapcho, M. (1989). Age differences in a vision information processing capability underlying traffic control device usage. Transportation Research Record, 1244, 63-72.
Staplin, L., Lococo, K. & Sim, J. (1990). Traffic maneuver problems of older drivers. FHWA-RD-92-092. McLean, VA: Federal Highway Administration. 213 pp.
Staplin, L., Harkey, D.L., Lococo, K.H. & Tarawneh, M.S. (1997). Intersection geometric design and operational guidelines for older drivers and pedestrians. Volume 1. Final Report. FHWA-RD-96-132. McLean, VA: Federal highway Administration. 234 pp.
Sterns, H.L. & Gray, J.H. (1999). Work, leisure and retirement. In J.C. Cavanaugh and S.K. Whitbourne (Eds.), Gerontology: An interdisciplinary perspective. New York: Oxford University Press. pp. 355-390.
Transportation Research Board (1988). Transportation in an aging society: Improving mobility and safety for older persons. Special Report 218. Volumes 1 and 2. Washington, DC: National Research Council.
Turner, J.D., Nitzburg, M. & Knoblauch, R.L. (1998). Ultraviolet headlamp technology for nighttime enhancement of roadway markings and pedestrians. Transportation Research Record, 1636, 124-131.
U.S. Department of Transportation (1988). Manual of uniform traffic control devices. Washington, DC: U.S. Government Printing Office.
Williams, J.C., Ardekani, S.A. and Asante, S. (1992). Motorist understanding of left-turn signal indications and auxiliary signs. Transportation Research Record, 1376, 57-63.
Zwahlen, H.T. & Schnell, T. (1998). The visibility of road markings as a function of age and retroreflectivity under low-beam and high-beam illumination at night. Paper presented at the 77th Annual Meeting of the Transportation Research Board, Washington, DC.
Zwahlen, H.T. & Schnell, T. (1998). Driver eye scanning behavior when reading symbolic warning signs. In A. Gale (Ed.), Vision in vehicles. VI. Amsterdam: Elsevier Science Publishers. pp. 3-11.
Zwahlen, H.T. & Schnell, T. (in press). Minimum in-service retroreflectivity of pavement markings. Transportation Research Record.
................
................
In order to avoid copyright disputes, this page is only a partial summary.
To fulfill the demand for quickly locating and searching documents.
It is intelligent file search solution for home and business.