1st International Conference on Transportation Infrastructure and Materials (ICTIM 2016) ISBN: 978-1-60595-367-0
Pedestrian Walking Speed Characteristics at Signalized Intersections in New Urban District Communities Jiang-ling Wu1, Sheng-rui Zhang2, Amit Kumar Singh3, Qiu-ping Wang4 1
PhD Candidate, Highway School, Chang’an University, Xi’an Shaanxi 710054, China;
[email protected] 2 Professor, Highway School, Chang’an University, Xi’an Shaanxi 710054, China;
[email protected] 3 Engineer, Atkins, 6504 Bridge Point Parkway, Suite 200,| Austin, Texas USA 78730;
[email protected] 4 Professor, School of Civil Engineering, Xi'an University of Architecture and Technology, Xi’an Shaanxi 710055 China;
[email protected]
ABSTRACT: Walking speed is an important parameter for signal timing and safety analysis at signalized intersections. Walking speed varies significantly at different location and over time, and need to be re-calibrated for safer and efficient intersection designs. This paper documents a study performed to analyze pedestrian’s speed characteristics at signalized intersections in new urban district communities. Field data were obtained through video camera at four typical 4-leg intersections in the high-tech metropolitan area of Xi’an, China. Statistical descriptions and the distribution fittings were conducted by Statistical Product and Service Solutions （SPSS）. Results shows that the pedestrian crossing speed characteristics are influenced by both the pedestrian’s own characteristics and an intersection's geometric and operational characteristics. Male and female walking speed distributions are all in line with the normal distribution, and the youth, middle-aged, and elderly pedestrians speed distributions follow log-normal distribution. The average walking speed varies significantly with the variance of the crosswalk lengths at the same intersection. INTRODUCTION With the rapid development of big cities in China, pedestrian-vehicle conflicts are happening more than ever before. Pedestrians have received more attention because of the lack of safety during the past decades. Pedestrian walking speed at an intersection depends on several factors such as physical characteristics of pedestrian, age, gender, crosswalk length, and intersection location. People from different geographical location walk at different speed. With unpreceded change in people’s lifestyle, and people’s walking characteristics have changes significantly over past decades. With better transportation infrastructure and ease of access to transportation, propensity of people for walking in decreasing. Walking speed is a crucial input for intersection design and signal timings assignment at an intersection. As dynamics of pedestrian walking behavior changes with location and time, it is important to calibrate pedestrian walking speed to current and local pedestrian speed to design a safer and efficient intersection. As pedestrian walking speed and their behavior changes significantly from one location in the world to another location, extensive studies has been done to calibrate pedestrian walking speed at local level. Coffin et al. (1995)[1] researched walking speeds of elderly pedestrians at various types of crosswalks and found that people over the age of 60 are not a homogeneous group. The study suggested design walking speeds at signalized intersections 453
near seniors and nursing homes (1.0 meter per second [m/sec]), midblock crosswalks (1.0 m/sec), and normal signalized intersections (1.2 m/sec). A series of field studies was conducted by Knoblauch et al. (1996)[2] at 16 crosswalks in four urban areas to quantify the walking speed and start-up time of pedestrians. The 15th percentile walking speed and design purpose values for younger pedestrians (ages 14 to 64) and older pedestrians (ages 65 years or older) were obtained. Guerrier et al. (1998)[3] studied five intersections in the city of Miami Beach to identify the problems encountered by young, middle-aged, and older pedestrians. The average walking speed for young and old pedestrians was concluded. The 15th percentile walking speed for pedestrians and different age pedestrians was determined. Tarawneh (2001)[4] recommended the 15th percentile speed of 1.11 m/sec as the design speed for traffic signals design in Jordan. Specifically, in areas where older pedestrians, 65 years or older, are frequently encountered, a design speed of 0.97 m/sec was recommended by Tarawneh to accommodate at least 85% of this slower group. Bennett et al. (2001)[5] investigated pedestrian movement characteristics at four intersection signalized in Australia. John et al. (2007)[6] gave the history of the 4.0 feet per second (ft/sec) walking speed, going back to the 1961 Manual of Uniform Traffic Control Devices (MUTCD). Alhajyaseen et al. (2011)[7] analyzed the effects of pedestrian age group and the interaction between bi-directional pedestrian flows on the capacity of signalized crosswalks. The study shows that the maximum reduction in capacity occurs at approximately equal pedestrian flows from both sides of the crosswalk, and elderly pedestrians might cause a significant reduction in capacity by up to 30%. Jing (2007)[8] studied pedestrian walking speeds based on gender, age, individual and pedestrian group at an unsignalized intersection in Changchun. The distributions in different lanes were studied. Xu et al. (2015)[9] summarized factors that influenced pedestrian walking speeds. Pedestrian walking characteristics also vary based on their location and physical characteristics such as height, weight, lifestyle, etc. Feng et al. (2004)[10] claimed that the average pedestrian crossing speed is 1.47 m/sec in Harbin, China, and the 85th percentile speed and the 15th percentile speed are 1.74 m/sec, 1.19 m/sec, respectively. They recommended the 15th percentile speed of 1.19 m/sec as the design speed for traffic signals design. Specifically, in areas where older pedestrians are frequently encountered, a design speed of 1.07 m/sec, which is 15% of the speed of the elderly, was recommended by Feng to accommodate at least 85% of the elderly pedestrians. Kong et al. (2010)[11] provided a design pedestrian speed of 0.93 m/sec, ~1.1 m/sec at intersections in Shanghai. Pedestrian speed and their variation has been studied extensively in academia. However, little is known about the characteristics and behavior of pedestrians’ movements at intersections in China, except in the case of children. This paper aims to calibrate the pedestrian’s walking speed within a high-tech metropolitan area of Xi’an, China, analyzes pedestrians’ speed characteristics at signalized intersections under influences of the pedestrians own characteristics and the intersections’ geometric and operational characteristics, and mainly focuses on the pedestrian walking speed distribution characteristics. Based on the results, the design speeds are recommended. METHODOLOGY Four typical 4-leg signalized intersections in the Xi’an urban area were selected. The 454
sites are located at the intersections of Gaoxin Road and Keji Road, Kechuang Road and Bowen Road, Gaoxin Road and Second Keji Road, Second Keji Road and Bowen Road in High-Tech Zone of Xi’an (Table 1). Table 1. Signalized Intersection Facilities. Serial No. a b
Signal timing(s) Green Red Yellow 45 56 3 56 45 3
Intersections Gaoxin Road Keji Road
Road Gaoxin Road Keji Road
Crosswalk length(m) 36 36
Kechuang Road Bowen Road
Kechuang Road
20
32
29
3
Bowen Road Gaoxin Road Second Keji Road Second Keji Road Bowen Road
10 36
29 48
32 33
3 3
Road Hierarchy Arterial Arterial Minor Arterial Branch Arterial
30
36
45
3
Arterial
30
40
33
3
Arterial
10
30
43
3
Branch
c
Gaoxin Road – Second Keji Road
d
Second Keji Road – Bowen Road
Data were collected with a video camera on Monday, January 9, 2012 from 12:00 am to 02:00 pm and 04:30 pm to 06:30 pm. Surface conditions at the curb and in the crosswalk, weather conditions, and wind intensity were recorded for each observation. To collect consistent and reliable data more easily, the position of the camera and the observation sample size have to be controlled properly. Cameras were mounted above the observation area. The random sampling is adopted to get limited samples. Some pedestrian’s ages were casual inspected by the observer, so that all sample pedestrians’ age can be estimated accurately. The purpose of controlling the sample size is to quantify accurately normal walking speeds. The minimum number of observed pedestrians can be decided by the following formula. K n E
2
（1）
Where, E – The permissible error of observations, which depends on the required precision of pedestrian walking speeds. E 0.07m / s , in general. K – Confidence coefficient. K equals to 1.96 at the 95% confidence level, in general. – Standard deviation, which equals to 0.37 m/sec [12]. According to Equation (1), the minimum number of observed pedestrians is obtained, i.e., n=107 . Many factors, such as the gender and age of pedestrians, trip purpose, alone or walking in a group, the length of crosswalks, traffic volume, and signal timing, can influence the pedestrian’s normal walking speed[3,8,13]. This paper focuses on three main factors: gender, age, and the length of crosswalks. Some distribution types were taken to fit the distribution of the male and female pedestrian walking speeds, and goodness-of-fit tests were conducted with Kolmogorov-Smirnov. K-S goodness-of-fit test results indicate both the male and the female walking speeds obey the normal distribution very well. Figures 1 and 2 show the histogram 455
and the normal Q-Q plot of the male walking speed, respectively. Similarly, Figures 3 and 4 show the histogram and normal Q-Q plot of the female walking speed, respectively.
Figure 1. Histogram of the male walking speed.
Figure 2. Normal Q-Q plot of the male walking speed.
Figure 3. Histogram of the female walking speed.
Figure 4. Normal Q-Q plot of the female walking speed.
As shown on Figures 1 and 3, the distributions of the male and female walking speeds are slightly skewed. The mean, standard deviation and maximum of the male walking speed are higher than the female. The observed values, as shown on Figures 2 and 4, are consistent with the expected normal values. The results of K-S goodness-of-fit tests for the normal distribution are shown in Table 2, from which we can see significance levels of the male and female are all more than 0.05, which suggests that both the male and the female walking speeds follows the normal distribution. Table 2. Distribution Test of Pedestrian Walking Speed. Kolmogorov-Smirnov Z Gender Degree of Freedom Significance Level Female 0.794 621 0.554 Male 0.996 927 0.275
456
Table 3 presents statistical parameters. The male walking speed varies from 0.50 m/sec to 2.25 m/sec, and the mean speed is about 1.23 m/sec. While the female walking speed varies from 0.63 m/sec to 1.82 m/sec, and the mean speed is about 1.13 m/sec, which is 0.1 m/sec less than the males. The 15th percentile crossing speed for the male compliers was 0.90 m/sec, whereas the female compliers crossed at 0.89 m/sec. Table 3. Walking Speed distributions of Male and Female Pedestrians. Percentile(m/s) Min Max Gender N Mean Std. Dev. th (m/sec) (m/sec) 15 50th 85th Female 621 1.12892 0.215971 0.62 1.82 0.886 1.125 1.333 Male 927 1.23401 0.317665 0.50 2.25 0.900 1.237 1.564 Age groups include the youth (16–40 years old), the middle-aged (41–65 years old), and the elderly (above 65). Walking speed boxplots of different age groups, as shown on Figure 5, illustrate that the youths crossed fastest and that the elderly walked slowest among all the groups. Kolmogorov-Smirnov goodness-of-fit tests were conducted to fit the distribution of the three age groups. K-S goodness-of-fit test results indicate each group’s walking speeds all obey the Logarithmic normal distribution.
Figure 5. Walking Speed Boxplots of Different Age Groups. The results of the K-S goodness-of-fit tests for the Logarithmic normal distribution are shown in Table 4, from which we can see significance levels of the three groups are all more than 0.05, which means walking speeds of the youth, the middle-aged, and the elderly obey the Logarithmic normal distribution well. Statistical parameters are presented in Table 5.
Age Youth Middle-aged Elderly
Table 4. Distribution Test of Pedestrian Walking Speed. Degree of Kolmogorov-Smirnov Z Significance Level Freedom 0.339 691 0.055 0.316 622 0.063 0.794 135 0.554 457
According to Table 5, the range of the youth walking speed is 0.62 m/sec, ～2.25 m/sec, and the mean speed is about 1.25 m/sec. The range of the middle-aged walking speed is 0.52 m/sec,～2.06 m/sec, and the mean speed is about 1.13 m/sec. The range of the elderly walking speed is 0.50 m/sec, ～2.00 m/sec, and the mean speed is about 1.02 m/sec. The mean walking speed of middle-aged pedestrians is less than the youth by 0.12 m/sec, and more than the elderly by 0.11 m/sec. Table 5. Walking Speed Distributions of Pedestrians with Different Age Groups. Mean（m/sec） Sample Size Age Std. Dev. Min (m/sec) Max (m/sec) 1.25 691 0.31 0.62 2.25 Youth 1.13 622 0.27 0.52 2.06 Middle-aged 1.02 135 0.25 0.50 2.00 Elderly What happens when an arterial and a minor arterial meet, or an arterial meets with a branch? In other words, how could the hierarchical inequality of roads at a signalized intersection impact the characteristics of pedestrian walking speeds? In order to answer that, walking speeds at intersections of major streets and minor streets are analyzed. Zhang[13] analyzed pedestrian walking speed characteristics on the length of 14-m and 40-m crosswalks at a four-way signal intersection and concluded that the mean walking speed on the 14-m and 40-m crosswalks are 1.31 m/sec and 1.35 m/sec, respectively. The limitations of this report located on these two crosswalks studied were from different intersections. The influences of the crosswalk length at a signal intersection(LP) on the pedestrian walking speed should be considered. Table 6 shows the statistics of the male walking speed at different scale intersections. Table 6. Statistics of the Male Walking Speed at Different Scale Intersections. Intersections LP(m) Mean (m/sec) Std. Dev. Min (m/sec) Max (m/sec) 36 1.22 0.285 0.69 2.12 Gaoxin Road Keji Road 36 1.20 0.29 0.63 2.25 20 1.32 0.25 0.87 1.82 Kechuang Road Bowen Road 10 1.01 0.34 0.50 2.00 This preliminary research indicates that the difference degree of the length of crosswalks in an intersection has effects on pedestrian walking speeds. At the same intersection, the more differences degree of the length of crosswalks is, the greater differences of pedestrian walking speed are. The 15th percentile is usually used as the design walking speed for intersection designs[15]. The existing research results of design walking speed at signalized intersections in different cities of China[10][11] are shown in Table 7. Table 7. The Recommended Value of Design Walking Speed at Signalized Intersections in other Cities (m/sec). City Mean 15th Percentile 85th Percentile Recommendation Shanghai 1.40 1.09 1.70 0.93~1.10 Harbin 1.47 1.19 1.74 1.19 458
The recommendations of design walking speeds for different gender and age pedestrians in Xi’an are shown in Table 8. The range of the 15th percentile walking speed by gender and age category is 0.786, ~0.933 m/sec, as shown in Table 8, and overall pedestrians’ 15th percentile walking speed is 0.868 m/sec. The principles of determining the design walking speed are not only to ensure the intersections operate effectively and reduce traffic delays, but also ensure that most pedestrians, especially the elderly and disabled, could cross the intersection safely and smoothly. A design pedestrian speed of 0.85 m/sec, ~0.93 m/sec at intersections in Xi’an is recommended. Compared to other research results, recommended design walking speeds are smaller. This might be due to differences in pedestrian spatially characteristics, the socioeconomic development, and the urban life rhythm, since error factors were ruled out. Shanghai is a well-developed city, compared to Xi’an, and the tempo of life there is faster, which results in people walking faster. The economic development level of Xi’an is not significantly different from Harbin, but the height of pedestrians in Harbin is higher than that in Xi’an. Table 8. The Recommended Value of Design Walking Speed at Signalized Intersections in Xi’an (m/sec). Sample
Male
Female
Youth
Middle-aged
Elderly
Overall
Mean(m/sec)
1.23
1.13
1.25
1.13
1.02
1.18
th
0.900
0.886
0.933
0.857
0.786
0.868
th
85 Percentile
1.564
1.333
1.560
1.395
1.248
1.472
Recommendation
0.90
0.89
0.93
0.86
0.79
0.87
15 Percentile
CONCLUSION Based on the abundant literatures regarding walking speed, it would appear that pedestrian walking speeds have been studied by many different researchers in great detail. However, the examination of some of these studies reveals many differing results for pedestrian walking speeds. There is limited information on distributions of pedestrian walking speeds. This paper mainly focuses on statistical distributions of pedestrian walking speeds. The design speed at intersections in Xi’an was obtained to revise the current intersection design standard and transportation policy. 1) Male and female walking speed distributions are all in line with the normal distribution, and the youth, middle-aged, and elderly pedestrians’ speed distributions follow the log-normal distribution. 2) The difference degree of the length of crosswalks at an intersection has effects on pedestrian walking speeds. The greater the differences degree of the length of the crosswalks is, the greater the differences of pedestrian walking speed are. The significance of effects is one of works in the future. 3) A design pedestrian speed of 0.85 m/sec, ~0.93 m/sec at four-way intersections in Xi’an is recommended. 4) Pedestrian walking speeds are lower than the other cities that have been reported. This research can provide theoretical and data supports for the current intersection design standard and transportation policy. For speed, fitting with normal distribution has a big drawback is that negative 459
values are theoretically possible in normal distribution. However, it is not true in practice. More discussions may be worth given in the future. Today’s traffic environment in China is not well adapted to the needs of the older pedestrian and the disabled. The future research should focus more on the vulnerable groups. ACKNOWLEDGMENTS The research reported in this paper was sponsored by the National Science Foundation of China (General Program) (Grant No. 51278396). The authors would like to thank Dong Guan for giving support to collect field data and reviewers who provided constructive feedback on an earlier version of this paper. REFERENCES [1] Coffin A., and J. Morrall. (1995). “Walking speeds of elderly pedestrians at crosswalks.” Transportation Research Record, 1487:63-67. [2] Knoblauch, R.L., M.T. Pietrucha, and M. Nitzburg. (1996). “Field studies of pedestrian walking speed and start-up time”. Transportation Research Record: Journal of the Transportation Research Board, 1538(1):27-38. [3] Guerrier, J.H., and S.C. Jolibois. (1998). “The safety of elderly pedestrians at five urban intersections in Miami.” Proceedings of the Human Factors and Ergonomics Society Annual Meeting. SAGE Publications 42(2):171-175. [4] Tarawneh, M.S. (2001). “Evaluation of pedestrian speed in Jordan with investigation of some contributing factors.” Journal of Safety Research, 32(2):229-236. [5] [6] LaPlante J., and T.P. Kaeser. (2007). “A history of pedestrian signal walking speed assumptions.” 3rd Urban Street Symposium: Uptown, Downtown, or Small Town: Designing Urban Streets That Work. [7] Alhajyaseen, W.K.M., H. Nakamura, and M. Asano. (2011). “Effects of bi-directional pedestrian flow characteristics upon the capacity of signalized crosswalks.” Procedia-Social and Behavioral Sciences, 16:526-535. [8] Chao Jing. (2007). “Study on Pedestrian Crossing Characteristics.” Changchun: Jilin University. [9] Jiqian Xu, and Xuewu Chen. (2015). “Traffic Engineering, 4th edition.” China Communications Press Co., Ltd. pp. 23-25, Beijing. [10] Feng Shu-min, and Wu Yue-xin. (2004). “Crossing speed analysis for pedestrian at signalized intersection.” Journal of Harbin Institute of Technology, 36(1):76-78. [11] Kong Ling-zheng ， Xu Ruo-ying ， Xu Yi-lun ， and Jiang Xiao. (2010). “Investigation and Study on Pedestrian Crossing Speed at Signalized Intersection.” Transport Standardization, 024:86-89. [12] Deweu Li, and Baoming Han. (2011). “Pedestrian Traffic.” Beijing: China Communication Press, pp.43. [13] Jian Lu，Huiqiong Ye, and Donglei Yao. (2002). “Reasonable distance of pedestrian crossing facilities.” Journal of Traffic and Transportation Engineering, 2(4):63-67.
460
[14] Zhuang Zhang. (2010). “The traffic characteristics and violation analysis of pedestrians at four-way signals intersection.” Beijing: Beijing University of Technology. [15] Yulong Pei, and Shumin Feng. (2006). “Research on Design Speed of Urban Pedestrian Crossing.” Journal of Highway and Transportation Research and Development, 23(9):104-107.
461
