TRANSPORTATION RESEARCH RECORD 1538
27
Field Studies of Pedestrian Walking Speed and Start-Up Time
RICHARD L. KNOBLAUCH, MARTIN T. PIETRUCHA, AND MARSHA NITZBURG
Todays traffic environment is not well adapted to the needs of the older pedestrian. Unfortunately, except in the case of children, little is known about the characteristics and behavior of pedestrians. Although the simple fact that older pedestrians walk more slowly than younger ones is easily supported by eld data, existing data on walking speeds and start-up time (i.e., the time from the onset of a Walk signal until the pedestrian steps off the curb) have many shortcomings. A series of eld studies was conducted to quantify the walking speed and start-up time of pedestrians of various ages under different conditions. Sixteen crosswalks in four urban areas were studied. Data were collected on walking speeds and start-up times relative to site and environmental factors, including street width, posted speed, curb height, grade, number of vehicle travel lanes, signal cycle length, pedestrian-signal type, street functional classification, crosswalk type, and channelization. Data on a subject group of pedestrians who appeared to be 65 years of age or older and a control group of pedestrians under age 65 were collected. Results indicate a broad range of walking speeds among pedestrians. The 15th-percentile walking speed for younger pedestrians (ages 14 to 64) was 1.25 m/sec (4.09 ft/sec); for older pedestrians (ages 65 and over) it was 0.97 m/sec (3.19 ft/sec). For design purposes values of 1.22 m/sec (4 ft/sec) for younger pedestrians and 0.91 m/sec (3 ft/sec) for older pedestrians are appropriate. Results also indicated that walking rates are inuenced by a variety of factors, including the functional classication and vehicle volumes on the street being crossed, the street width, weather conditions, the number of pedestrians crossing in a group, the signal cycle length, the timing of the various pedestrian-signal phases, whether right turn on red is allowed, pedestrian signals, medians, curb cuts, crosswalk markings, stop lines, and on-street parking. However, for each of these factors, the effect on crossing speeds, although statistically signicant, is not meaningful for design.
The older road user has received much attention during the past decade, and with good reason. The proportion of those over age 65 in the North American population is increasing and will continue to increase dramatically. Research related to older road users has also increased. In 1992 the Federal Highway Administration was sponsoring eight major research projects on older road users. An examination of these projects indicates that older drivers, not older pedestrians, receive the majority of attention in research related to older road users. Todays traffic environment is not well adapted to the needs of the older pedestrian. Unfortunately, except in the case of children, very little is known about the characteristics and behavior of pedestrians. The simple fact that older pedestrians walk more slowly than younger ones is easily supported by eld data; however, the existing data on walking speeds and start-up time (i.e., the time from the
R.L. Knoblauch and M. Nitzburg, Center for Applied Research, 9308 Georgetown Pike, Great Falls, Va. 22066. M.T. Pietrucha, The Pennsylvania State University, The Pennsylvania Transportation Institute, Research Office Building, University Park, Pa. 16802.
onset of a Walk signal until the pedestrian steps off the curb) have many shortcomings. For example, Dahlstedt (1) instructed a group of people aged 70 or older to cross an intersection at fast, very fast, or normal speed. Fast for about 60 percent of the group was less than 1.22 m/sec (4 ft/sec); normal for 90 percent of the group was also less than 1.22 m/sec (4 ft/sec), and the 85th-percentile speed was about 0.67 m/sec (2.2 ft/sec). The Manual on Uniform Traffic Control Devices (MUTCD) (2) suggests a walking speed of 1.22 m/sec (4 ft/sec) for traffic signal timing. A literature review by McGee, et al. ( 3) indicated that many pedestriansperhaps 30 percent of the population, many of whom are olderdo not normally walk that quickly. In fact, the Traffic Control Devices Handbook (TCDH) (4) also notes that one-third of all pedestrians cross more slowly, with 15 percent at or below 1.06 m/sec (3.5 ft/sec). TCDH states that those having slower walking speeds have the moral and legal right to complete the crossing once they have entered the intersection. The Traffic Engineering Handbook (5) suggests that 0.91 to 0.99 m/sec (3 to 3.25 ft/sec) would be a more appropriate value to use for traffic signal timing. An Institute of Transportation Engineers committee concerned with pedestrian issues (6 ) conducted a survey at a Florida location with a large population of elderly pedestrians. The committee recommended 0.76 m/sec (2.5 ft/sec) as an appropriate walk speed (for locations with a high volume of elderly pedestrians), and found this speed to be adequate for 87 percent of pedestrians observed. However, walking speeds are influenced by environmental, traffic, and pedestrian characteristics. The effects of terrain on walking speeds are unknown, although it can be expected that the elderly would be affected more when walking up or down a grade than the young. Similarly, it can be expected that the elderly would react more strongly to higher vehicular densities and traffic speeds, out of a fear of traffic. Moore ( 7 ) noted that the closer the approaching vehicle, the faster the mean crossing time1.52 m/sec (5 ft/sec) when the approaching vehicle was 3 sec away, 1.22 m/sec (4 ft/sec) when the approaching vehicle was not too close. Finally, pedestrian speed on sidewalks and crosswalks is strongly related to the number of pedestrians in the flow. The relationship between speed, flow, and space occupied (i.e., density) for a representative population group has been examined by Fruin ( 8) and others. However, the abilities of the elderly in crowds have not yet been documented. Because of the shortcomings in data on the walking speeds of older pedestrians a series of eld studies was conducted to quantify the walking speed and start-up time of pedestrians of various ages under different environmental conditions.
28
TRANSPORTATION RESEARCH RECORD 1538
METHOD Site Selection Sixteen crosswalks at signal-controlled intersections in four urban areas (Richmond, Virginia; Washington, D.C.; Baltimore, Maryland; and Buffalo, New York) were selected. Sites were selected to allow for a minimum of 26 to 30 pedestrians over 65 years of age to be observed during an 8-hr data-collection period. During pilot testing, estimates of population variance were computed and a sample size of 26 to 30 was determined to be sufficient to quantify the effects of site-specific factors. To quantify walking speeds and start-up times relative to different site and environmental factors, the following types of site-specic information were recorded: Street width, Posted speed, Curb height, Grade, Number of travel lanes, Signal cycle length, Pedestrian signal type, Street classication, Crosswalk type, and Channelization.
People in wheelchairs; and Pedestrians walking bikes or dogs. To accurately quantify normal walking speeds of the various subject groups, pedestrians who exhibited any of the following behaviors were also not observed: Crossing diagonally, Stopping or resting at the median, Entering the roadway running (anything faster than a fast walk), Stepping into the roadway (leaving the curb) before starting to cross or waiting for traffic during crossing, and Entering or exiting the roadway more than 1.2 m (4 ft) outside the crosswalk. The gender of each subject was recorded, as well as whether he or she was walking alone or in a group. The group size was also noted when applicable. A group was dened as two or more pedestrians crossing the roadway at about the same time, regardless of whether or not they were apparently friends or associates. In addition, subjects paths were monitored to determine whether they started and ended their crossings inside or outside the crosswalk. Inside the crosswalk was established as within or on the painted crosswalk lines. Compliance with the pedestrian signal (or traffic signal for sites without pedestrian signals) was recorded. Observers specically recorded whether the Walk, Dont Walk, or ashing Dont Walk signal was on at the start of the subjects crossing. When the signal changed during the crossing the information was recorded along with the subjects locationwhether the rst or second half of the crossingat the time the change occurred. Nine other pedestrian behaviors were recorded when they occurred: Confusion (hesitation, change in direction of travel) exhibited before crossing; Confusion exhibited after entering the roadway; Cane use; Following the lead of other pedestrians; Inattention when pedestrian signal changed to Walk; Stopping in the crosswalk during the crossing; Difficulty going down the curb; Difficulty going up the curb; and Running during part of the crossing (anything faster than a fast walk). To verify the accuracy and reliability of the age-estimation abilities of the observers, several eld verications were done. First, the age-estimation accuracy of several observers was measured, then correlations between the estimates of all of the observers were determined. The results of these verication procedures are discussed after the next section.
Weather Conditions Data were collected during three different types of weather conditions: Dry: clear (no precipitation), with dry roads and dry sidewalk; Rain: any type of rain from drizzle and mist to moderate rain, with wet roads and wet curbs. Data were not collected during heavy rain because pedestrians tended not to be outside. Data were also collected immediately after the precipitation stopped when the road and sidewalk were still wet; and Snow: when there was snow or ice in the atmosphere, on the road or sidewalk, or both. Data were collected on weekdays during daylight conditions. Surface conditions at the curb, in the crosswalk, weather conditions, and the estimated wind intensity were recorded for each observation.
Subject Selection Data were collected on a subject group of pedestrians who appeared to be 65 years of age or older. Data on a control group of pedestrians under age 65 were also collected. The following individuals were specically not observed: Children under 13 years of age; Pedestrians carrying children, heavy bags, or suitcases; Pedestrians pushing strollers or grocery carts; Pedestrians holding hands or assisting others across the roadway; Pedestrians using a quadpod cane, walker, two canes, or crutches;
Procedure Pedestrian crossing times were measured with a hand-held, digital, electronic stopwatch. The watch was started as the subject stepped off the curb and stopped when the subject stepped up on the opposite curb after crossing. At sites with a pedestrian signal, pedestrian start-up times were also measured.
Knoblauch et al.
29
Verication of Observer Age Estimates and Start-up Time Measurement To determine the ability of the eld observers to properly identify older pedestrians a simple verication procedure was conducted. Five eld observers estimated the ages of a randomly selected sample of nine pedestrians who ranged in age from 54 to 85 years. With one exception, a 62-year-old male, each observer correctly identied as over 65 each pedestrian who subsequently reported his or her age as over 65. All ve observers thought that the 62-year-old male was over 65; estimates ranged from 68 to 75 years old. Interrater reliability for age was assessed using intraclass correlations and Pearson r correlations. The intraclass correlation was .78 for the ve raters. Pearson r correlations between individual raters ranged from .71 to .93; between subjects actual age and each rater it ranged from .70 to .82. This indicates that the observers, as a group, were quite good at identifying pedestrians over the age of 65 and that there was a more than acceptable level of agreement among observers. A similar procedure was used to verify the reliability of the stopwatch measurements of pedestrian crossing times. The crossing times of the same nine pedestrians were measured by the ve eld observers. Again, intraclass correlations and Pearson r correlations were used to determine interrater reliabilities. The intraclass correlation was .998 for the ve raters, and all Pearson r correlations between individual raters were greater than .99. This indicates that the observers were each following the timing procedure in a similar manner.
by notation of an age-by-site interaction (A S) in the signicance column. Although many of the mean differences shown are statistically signicant, this is in part because of the relatively large number of observations made. When examining these tables it is important to consider the relative magnitude of the differences and whether or not the differences are meaningful. Many were found to be statistically signicant, but do not make any practical difference to facilities design practices. The following discussion highlights some of the walking-speed differences noted for the entire sample of pedestrians observed. The mean walking speed for younger pedestrians was 1.51 m/sec (4.95 ft/sec); it was 1.25 m/sec (4.11 ft/sec) for older pedestrians. The 15th-percentile speeds were 1.25 and 0.97 m/sec (4.09 and 3.19 ft/sec) for younger and older pedestrians, respectively. These mean differences are signicant at the 0.05 level. Other highlights include the following observations. Young male pedestrians had the fastest walking speeds [1.56 m/sec (5.11 ft/sec)]; older females had the slowest [1.19 m/sec (3.89 ft/sec)]. The differences between younger men and younger women [0.10 m/sec (0.32 ft/sec)] and between older men and older women [0.13 m/sec (0.42 ft/sec)] are approximately the same; Pedestrians who start on the Walk signal walk more slowly than those who cross on either the ashing Dont Walk or the steady Dont Walk. The differences between the Walk crossers and the ashing Dont Walk crossers suggest that some pedestrians may understand the concept of the clearance phase, specically, that although it is dangerous to cross on the steady Dont Walk (and if doing so one should walk as fast as possible), it is slightly less dangerous to cross on the ashing Dont Walk. [This is contrary to information from some other studies (e.g., Zegeer [9]), which indicate a general lack of understanding of the ashing Dont Walk.] At sites without pedestrian signals pedestrians who start to cross legally (on the green) also tend to walk more slowly than those crossing against the signal indication (on the red). These differences between compliers and noncompliers have important implications to the design process. It is believed that the walking speeds of compliers provide a more appropriate basis for design than do the walking speeds of those who are crossing illegally; Locations with longer signal cycles had faster crossing speeds, probably because such locations tend to be wider roadways; Shorter pedestrian-signal times (Walk and ashing Dont Walk) also tend to be associated with faster walking speeds. It is not known if this relationship is causally related (i.e., the pedestrians walk faster because they know the crossing times are short) or if the shorter crossing times are typically displayed where pedestrians tend to walk more quickly (i.e., local streets). As might be expected, the walking speeds associated with the various steady Dont Walk cycle lengths are similar to those associated with the overall cycle length. Table 2 presents the mean and 15th-percentile of walking speeds for younger and older pedestrians who were observed crossing with the signal, that is, compliers. This subset included 4,460 pedestrians, approximately 62 percent of those included in Table 1. As a group, they tend to walk more slowly than the pedestrians who cross illegally. The mean crossing speed for the younger compliers was 1.46 m/sec (4.79 ft/sec) versus 1.51 m/sec (4.95 ft/sec) for all younger pedestrians observed. The older compliers averaged 1.20 m/sec (3.94 ft/sec), whereas all the older pedestrians observed averaged 1.25 m/sec (4.11 ft/sec). Both of these differences are signi-
RESULTS The overall objectives of the research dictated the orientation of data analysis, the purpose of which was to gather descriptive information on the walking abilities and street-crossing behavior of older pedestrians. Therefore, the data analysis is descriptive, as opposed to analytical. The walking speeds and start-up times of young and older pedestrians across a variety of situational factors are described. Although many of the differences shown are statistically signicant, it is important to consider the absolute magnitude or meaningfulness of these differences.
Pedestrian Walking Speeds Table 1 presents the mean and 15th-percentile walking speeds in meters per second for young and older pedestrians at different types of locations and under several different environmental conditions. The 15th-percentile walking speed represents the speed that 15 percent of the pedestrians do not exceed; conversely, it is the speed that 85 percent of the pedestrians do exceed. A total of 7,123 pedestrians were observed. Included were 3,458 pedestrians under 65 years of age and 3,665 pedestrians 65 and over. These data describe all the pedestrians observed: those crossing in compliance with the signal and those crossing against the signal. As is subsequently described, those who cross against the signal tend to walk more quickly. The table shows that each of the site and environmental factors collected indicate a signicant effect caused by age and by each of the site and environmental characteristics. For approximately onehalf of the site factors examined there was also a signicant interaction between pedestrian age and the site factor. This is indicated
TABLE 1 Mean and 15th-Percentile Walking Speeds in Meters per Second for Young and Older Pedestrians: All Pedestrians
(continued on next page)
TABLE 1 (continued)
(continued on next page)
32
TRANSPORTATION RESEARCH RECORD 1538
TABLE 1 (continued)
cant at the 0.05 level (t-test). The 15th-percentile crossing speed for the younger compliers was 1.21 m/sec (3.97 ft/sec), whereas older compliers crossed at 0.94 m/sec (3.08 ft/sec). This difference was compared using the test statistic: x85 x85 1 1 1.539 pooled + N N This produced a Z-ratio of 23.26, indicating that the 15th percentile values are signicantly different below the 0.001 level. Highlights of the effect of site-related and environmental factors on the walking speeds of the compliers are as follows: Younger females walk 0.1 m/sec (0.32 ft/sec) more slowly than younger males, whereas older females are 0.12 m/sec (0.4 ft/sec) slower than older males; Those who start or end their crossing outside the crosswalk tend to walk more quickly; Compliers crossing at locations with pedestrian signals are not walking more quickly than compliers crossing at locations with only a traffic signal; Single compliers also tend to walk more quickly than compliers walking in a group; Weather conditions have a signicant effect on walking speed. Older pedestrians, especially, walk more slowly when it is snowing. One of the slowest 15th-percentile values observed [0.9 m/sec (2.94 ft/sec)] was for older pedestrians crossing snow-covered roadways; Longer signal cycle lengths appear to be associated with faster walking times; Compliers also tend to walk more quickly at locations with short steady Walk and short ashing Walk intervals; and Longer steady Dont Walk intervals, like longer total signal cycle length, appear to be associated with slower walking times. Discussion of Pedestrian Walking Speed Results Not surprisingly, the data set for all pedestrians observed contains walking speeds that are signicantly faster than the data set for only those who crossed with the signal (compliers). For design purposes it is appropriate to use the data based on compliers. The walking speeds for this subset show statistically signicant variations across a variety of site and environmental conditions. However, both the mean and 15th-percentile data are tightly clustered for both younger and older pedestrians. The 15th-percentile value represents the walking speed that is exceeded by all but the slowest-walking 15 percent of the older pedestrian population. The means for the younger pedestrians range from 1.38 to 1.56 m/sec (4.51 to 5.12
ft/sec) across all conditions, with an overall mean speed of 1.46 m/sec (4.79 ft/sec). The means for the older pedestrians range from 1.14 m/sec to 1.29 m/sec (3.73 ft/sec to 4.24 ft/sec), with an overall mean speed of 1.21 m/sec (3.98 ft/sec). For design purposes a mean speed of 1.22 m/sec (4.00 ft/sec) would appear appropriate. The 15th-percentile scores are also tightly clustered. For younger pedestrians they range from 1.18 to 1.27 m/sec (3.86 to 4.18 ft/sec) with an average 15th-percentile speed across all sites of 1.19 (3.97). For older pedestrians, they range from 0.88 to 1.01 m/sec (2.90 to 3.31 ft/sec), with an average 15th-percentile speed across all sites of 0.94 m/sec (3.08 ft/sec). For design purposes for which a 15thpercentile value is appropriate it would appear that 0.91 m/sec (3 ft/sec) would be reasonable to use for older pedestrians.
Pedestrian Start-up Times Table 3 presents the mean and 15th-percentile start-up times for young and older pedestrians. Because start-up times could be measured for only those pedestrians who waited for the signal to change before starting their crossing, these pedestrians are by denition compliers. Start-up times were measured only at locations with a pedestrian signal and were dened, as has been stated as the elapsed time from the onset of a Walk signal to the moment when a pedestrian steps off the curb and starts to cross. Start-up times were measured only for those pedestrians who stopped at the curb and waited for the signal to change before starting to cross. It is interesting that the same site and environmental factors that almost always showed signicant interactions with walking speeds (as shown in Tables 1 and 2) did not (with the exception of roadway width) have a signicant interaction with start-up time. However, with the single exception of day of week, the start-up times viewed against every factor were signicantly longer for older pedestrians than for younger pedestrians. It could be hypothesized that being in a group would affect startup time. The presence of others could conceivably delay start-up or, alternatively, it might be expected that the group would start crossing as soon as the quickest pedestrian in the group started. The data indicate that there is no such effect. Younger pedestrians had identical mean start-up times (1.93 sec) whether alone or in a group. Older pedestrians had nearly identical start-up times of 2.5 sec when alone and 2.43 sec in a group. Some of the site and environmental factors included in Tables 1 and 2 are not included in Table 3. This is because the distribution of the sample across some of the categories did not occur with sufficient frequency to allow meaningful comparisons. For example, all but 11 of the 355 younger-pedestrian observations occurred on major arterials, so there were insufficient cases on either collector-
TABLE 2 Mean and 15th-Percentile Walking Speeds in Meters per Second for Young and Older Pedestrians: All Compliers
(continued on next page)
TABLE 2 (continued)
(continued on next page)
Knoblauch et al.
35
TABLE 2 (continued)
distributor or local streets to conduct meaningful analyses. This was because pedestrians rarely wait for the pedestrian signal unless they are forced to do so by oncoming traffic. Discussion of Pedestrian Start-Up Time Results Start-up times for the pedestrians observed did not show the same variability across site and environmental conditions that was noted for the walking-speed data. This was largely because of the limited variability between the sites at which start-up time could be measured. Start-up time could be measured only when the pedestrians chose to wait for the proper signal indication before starting to cross, which typically happens only when pedestrians are forced to wait because of oncoming traffic. The mean start-up times for younger pedestrians varied from 1.83 sec for males to 2.01 sec for females, with an overall mean value of 1.93 sec. For older pedestrians the mean values ranged from 2.39 sec for males to 2.57 sec for females, with an overall mean value of 2.48 sec. For design purposes a mean value of 2.5 sec would appears to be appropriate for older pedestrians, and a mean value of 2.0 sec would be suitable for younger pedestrians. The 85th-percentile values for younger pedestrians ranged from 2.76 sec (males) to 3.31 sec (females), with an overall value of 3.06 sec. For older pedestrians start-up times varied from 3.66 to 3.95 sec, with an overall value of 3.76 sec. For design purposes an 85thpercentile value of 3.75 sec would be appropriate for older pedestrians, and an 85th-percentile value of 3 sec would be suitable for younger pedestrians.
SUMMARY Based on the abundance of literature on the subject it would appear that pedestrian walking speeds have been studied by many different researchers in great detail. However, examination of some of these studies reveals many differing results for older-pedestrian walking speeds. There is also little or no information on the ranges or distributions of these older-pedestrian walking speeds. Furthermore, when specic studies are examined in detail, looking specically at number of observations, age of subjects observed, and range of conditions observed, it is clear that before the described effort a denitive walking speed study (especially for older pedestrians) had yet to be done. Information on pedestrian start-up time is extremely important in many aspects of highway design. As far as can be determined, there is no existing information on this subject. Therefore, the results presented give a rm basis for the selection of reliable pedestrian walking speeds and pedestrian start-up times.
The walking speed and start-up time data sets include information on both compliers and noncompliers. Compliers are pedestrians who cross during an appropriate pedestrian signal or traffic signal indication, and noncompliers are pedestrians who violate the pedestrian signal or traffic-signal indication. Pedestrians who violate signal indications likely do so knowingly. Therefore, they are aware of the increased (relative) risk in which they may be placing themselves and because of this they choose to leave the sidewalk area or walk at a speed that is different from the speed they might choose if they were crossing with the signal. For this reason the only information appropriate for design purposes are the data on pedestrian compliers, which are shown in (Table 2). In most cases a designers interest in pedestrian start-up times and walking speeds is twofold. First, a designer wants to have a true representation of how much time is needed by pedestrians of all ages and abilities to cross specic sections of roadway so a geometric layout and operational situation that is safe can be designed. Second, a designer is usually interested in having a design operate as efficiently as possible. This frequently translates into a desire to minimize the time necessary to accommodate the pedestrian and maximize the time available for the movement of vehicle traffic. With better information on start-up times and walking speeds relative to the site and environmental factors discussed, a designer can consider the effect of many different elements relative to these two major design objectives. The data indicate statistically signicant differences between start-up times and walking speeds by age category for most of the site and environmental factors discussed. It is also noted that the intuitive directionality of the start-up times and walking speeds (i.e., the times and speeds change in the direction expected, given an intuitively based supposition of the inuence of the factor) follow general expectations. However, in every case the differences between times and speeds, although statistically signicant, are not meaningful for a designer. Given the practically insignicant changes between start-up times and walking speeds by age category for most of the site and environmental factors, it is recommended that designers focus on the aggregated times and speeds for all complying walkers. Past studies have shown that there are no defendable criteria for acceptable minimum levels of human performance. Many researchers have suggested that performance be arbitrarily set at the 15th-85th percentile, nominally the break points in the cumulative frequency (ogive) plot of normally distributed data. It is believed that by using these values the bulk of the subject population is accommodated and only the true outliers are excluded from consideration. Therefore, it is recommended that the 15th-percentile values for older pedestrians for both start-up times and walking times (shown in Tables 2 and 3) be adopted as the standard for each of these parameters.
TABLE 3 Mean and 85th-Percentile Start-Up Times in Seconds for Young and Older Pedestrians: All Pedestrians
(continued on next page)
TABLE 3 (continued)
(continued on next page)
38
TRANSPORTATION RESEARCH RECORD 1538
TABLE 3 (continued)
ACKNOWLEDGMENT The research reported in this paper was sponsored by FHWA, U.S. Department of Transportation.
REFERENCES
1. Dahlstedt, S. Walking Speeds and Walking Habits of Elderly People. Swedish Road and Transport Research Institute, Stockholm, undated. 2. Manual on Uniform Traffic Control Devices for Streets and Highways. FHWA, U.S. Department of Transportation, 1988. 3. McGee, H. W., K. G. Hooper, W. E. Hughes, and W. Benson. Highway Design and Operational Standards Affected by Driver Characteristics: Final Technical Report. Report FHWA/RD 83/015. FHWA, U.S. Department of Transportation, 1983. 4. Traffic Control Devices Handbook. FHWA, U.S. Department of Transportation, 1983.
5. Dewar, R. E. Driver and Pedestrian Characteristics. In Traffic Engineering Handbook (4th edition, J. Pline, ed.), Prentice Hall, Englewood Cliffs, N.J., 1992. 6. ITE Committee 4A-6, Pedestrians. Traffic Control Devices for Elderly and Handicapped Pedestrians . Institute of Transportation Engineers, Washington, D.C., undated. 7. Moore, R. L. Psychological Factors of Importance in Traffic Engineering. Presented at International Study Week in Traffic Engineering, Stresa, Italy, 1956. 8. Fruin, J. J. Pedestrian Planning and Design. Elevator World, 1987. 9. Zegeer, C. V. Synthesis of Safety ResearchPedestrians. FHWA, U.S. Department of Transportation, 1991.
The contents of this paper reect the views of the authors and do not necessarily reect the official views or policies of the U.S. Department of Transportation. Publication of this paper sponsored by Committee on Pedestrians.
