1. sports
2. walking

Sciknow Publications Ltd.
RJMS 2015, 2(2):34-41
DOI: 10.12966/rjms.05.01.2015
Research Journal of Modeling and Simulation
©Attribution 3.0 Unported (CC BY 3.0)
Modeling Pedestrian Walking and Crossing Characteristics at
Baghdad CBD
Saad Issa Sarsam*, and Marwa Wahab Abdulameer
Department of Civil Engineering, College of Engineering, University of Baghdad, Baghdad, Iraq
*Corresponding author (Email: [email protected])
Abstract –To enable and encourage walking, understanding of the characteristics of pedestrian movements is vital. This work
examines the factors, which influence walking and crossing speed. The variations in walking speed of pedestrians are modeled
and related to pedestrian characteristics such as gender, age group, and clothing traditions. On the other hand, the crossing speed
of pedestrians was detected, and modeled. The counts of pedestrians were performed using video counting. The case study was
performed in two streets located in a crowded commercial zone at the city center of Baghdad. Data were subjected to statistical
analysis using IBM SPSS Statistics 19 software and final models for walking and crossing speed were selected. It has been found
that Iraqi pedestrians walk slower than other pedestrians do in the developed countries or in the region with minimum walking
speed of 29.85 m/min, and crossing speed of 1.17 m/sec. The free-flow speed of pedestrians was observed to be 63.5m/min. Age,
gender, and clothing traditions were found to significantly contribute to pedestrian speed. Pedestrians in the age range from
18-50 years old were the fastest group of pedestrians and pedestrians over 50 years old were the slowest. Male pedestrians had
significantly faster walking speeds than female pedestrians did. Pedestrians wearing western style (trousers) were found to be
faster than those wearing Arabic style.
Keywords –Age group, Clothing traditions, Crossing speed, Gender, Modeling, Pedestrian characteristics, Walking speed
1. Introduction
Proper planning of pedestrian facilities is crucial to allow optimum performance. This is especially so when considering that
pedestrian facilities are an integral part of the overall transportation network. To reduce environmental pollution,
pedestrianization has become an integral part of sustainable modern urban design. Thus, the design, arrangement and
development of support infrastructures should be in favor of pedestrian movements to popularize walking. To achieve so,
pedestrian facilities should be planned based on the concrete information on user characteristics, travelling patterns and
objectives of pedestrians flow. One of the objectives of the pedestrian studies is to evaluate the effects of a proposed policy on the
pedestrian facilities before its implementation. The implementation of a policy without pedestrian studies might lead to a very
costly trial and error due to the implementation cost (i.e. user cost, construction, marking etc.). On the other hand, using good
analysis tools, the trial and error of policy could be done at the analysis level. (Fruin, 1971), (Hoogendoorn et al. 2007) state that
two people walking side by side or passing each other while travelling in opposite directions take up an average space of 1.4 m
with adequate buffer areas on either side as demonstrated in Fig. 1. The minimum width that best serves two pedestrians walking
together or passing each other is 1.8 m. For design purposes, the (HCM, 2000) sets out a simplified body ellipse of 50 cm x 60 cm
for standing areas, with a total area of 0.3 m2 as shown in fig.2.
Fig. 1. Spatial Dimensions for Pedestrians, Pedestrian
Fig. 2. Pedestrian Body Ellipse for Standing Areas,
Research Journal of Modeling and Simulation (2015) 34-41
35
(Hoogendoorn et al., 2007)
(HCM, 2000)
(Daamen and Hoogendoorn, 2003) found that the walking speeds of individual appear to follow a normal distribution, with an
estimated mean of 1.34m/s and standard deviation of 0.37. Besides, (Finns and Walton, 2006) have conducted survey on mean
walking speed of different countries/cities. Different from Daamen’s findings in literature, the results show that the mean
walking speed of different countries can range from 0.7m/s to 1.8m/s (Schneider et al., 2008).
Federal Highway Administration of United States (HCM, 2000) uses 1.22m/s as the “normal” walking speed for designing
crosswalks. However, in New Zealand Government’s Pedestrian Planning and Design Guide (NZ Transport Agency, 2009),
adopted 1.5m/s as the mean walking speed of a “fit, healthy” adult, around 25% faster than that of United States. For aged and
those with mobility impairments, their mean walking speed is specified as 1.2m/s.
When determining pedestrian density, location and size of measurement area shall be carefully selected. (Hoogendoorn et al.
2007) demonstrated that high density of pedestrian can occur very locally such as waiting in front of stairs. The average
measured density therefore depends directly on how the measurement area is defined (Hoel, 1968; Koushki, 1988).
Pedestrian crossing speed is an important traffic engineering design parameter. The level of exposure of pedestrians to
vehicle traffic depends on their walking speeds. The literature suggests certain pedestrian speed values that may be used as a
guide in the design and operation of traffic signals. The Federal Highway Administration (HCM, 2000) recommends a pedestrian
crossing speed of 1.22 m/s for traffic signal timing. However, this value was questioned in a number of occasions. (McGee et al.
1983) indicated that many pedestrians, especially older pedestrians, do not walk that fast. (HCM, 2000) indicated that nearly
one-third of pedestrians walk slower than 1.22 m/s and nearly 15% of pedestrians walk at or below 1.06 m/s. (Dewar, 1992), as
cited in (Abdulameer, 2014) , (Sarsam and Abdulameer, 2014) recommends that 0.91 to 0.99 m/s be used for traffic signal timing.
(Sarsam, 2002) modeled the pedestrian crossing behavior at Mosul, he stated that a series of summary logistic regressions were
used to reduce the large body of results to a compact form which could be used in planning and design for safer pedestrian
facilities. Safety assessments can profitably be linked with the model measures to detect the degree of exposure to risk in this
vulnerable travel mode. He concluded that the crossing speed of 0.83 m/sec is almost lower than that found by other researchers.
Pedestrian were unaware of the danger of lengthy exposure to traffic movement, the tendency of the pedestrian is to take the
shortest course even if it involves a certain amount of risk when crossing. Such behavior may also indicate lower traffic
educational level. (Sarsam, 2013) stated that the contributing factors of the walking speed for male and female pedestrian are
clothing tradition, gender, and age group. He concluded that Male pedestrians move faster than female pedestrians do.
Pedestrians in the age group of 15–30 years had the highest speed of the range 77.7 – 66.1 m/min. on the other hand; female
pedestrian had the range of 74.5 – 69.9 m/min. for the same age group.
The literature also suggests that different locations have different effects on pedestrian movements. For example,
(Al-Masaeid et al., 1993) developed pedestrian speed-flow relationships for CBD areas in developing countries, and compared
them to the average values quoted in the 1985 edition of the HCM. One of the earliest researchers of pedestrian behavior was
(Polus et al., 1983) who used slow motion video surveys to collect pedestrian data. Since then, the method has been widely used
and described including (Benz and Fruin, 1984; Turvey et al., 1987; Khisty and Jotin, 1982). The purpose of this study is to
investigate how urban characteristics and land use affects pedestrian mobility; in terms of walking and crossing speed, and to
model the effect of pedestrian characteristics such as clothing tradition, age group, and gender on walking and crossing speed.
2. Methodology
Fig. 3. Al- Karada Dakhil and Al-Sina’a Street in Baghdad City, (Google Earth, 2013)
36
Research Journal of Modeling and Simulation (2015) 34-41
The pedestrian speed data were collected at two selected locations in Baghdad CBD area. The first site is located in a recreational
and shopping zone (AL- karada Dakhil); the second site is a commercial and educational zone holding colleges, (Al- Sina’a
Street) as shown in fig.3. It was expected that sites with different land use could show different pedestrian characteristics.
The collection of the field data was made for sample lengths of 1 hour and during good weather conditions, i.e. a sunny or
cloudy day without rain. The hours in which the counts were performed, were the ones where the peak hour was expected to take
place. It must be stated that some counts were performed for more than an hour, e.g. 2 hours. These hours were selected
considering the background information of the place. Specifically, the ranges selected were 13:00-14:00 and 17:00-18:00. The
workdays were used as the main sample days for this study. In this respect, random days among this group were chosen of March
and April 2013. The data that was gathered which include the walking and crossing time; the crossing direction of pedestrians;
approximate age; this was based on a subjective judgment. Young were considered to be below 18 years, adults between 18 and
50 years and the rest as elderly; Gender; and clothing Tradition style: Arab style and western style as shown in fig.4.
3. Counts Method
3.1. Collection of Geometric Data
The required data of each survey site in Baghdad are recorded. The effective width of sidewalks, the number of lanes and the lane
width were recorded and measured for each survey site. The lane width was measured directly on the sites by using measuring
tape. It was observed to be in the range of 3.3 -3.75 m for the nearest lane to the curb side. Video recording was performed; the
video provided more details that could be observed in a repetitive manner and with awareness. The video camera used was a
Canon HG10 and the sampling period was for 1 hour. The studied segment of sidewalks has dimensions shown in table1; where
the arcades widths are measured as the available space for pedestrian to walk as demonstrated in fig.5.
Fig. 4. Typical Arab Clothing Style
Fig. 5. Al- Karada Dakhil and al sina’a Street in Baghdad City
Table 1. Dimension of Sidewalk test section and crossing width for each Street
Street location
Section length (m)
Section width (m)
Street crossing width (m)
Al-Karada Dakhil
10
2.5
7.5
Al-Sina’a
10
3
10
3.2. Pedestrian walking speed
The technique adopted in the field work is by marking a longitudinal section of known length and width on the pedestrian facility
and continuously recording the movement of pedestrians within this section using video camera. Pedestrians were manually
timed over a measured test length, speeds were then calculated. Random pedestrian about to enter the section was selected and
tracked through the study area. The time taken by a pedestrian of various age, gender, and clothing style groups to traverse the
test length was measured using a digital stop watch, the entry and exit times in and out of the test area were recorded. Walking
speed is then derived by dividing the known length of the section by the walking time. Data were subjected to statistical analysis
using IBM SPSS Statistics 19 software. The speed was calculated using the mathematical model (Khisty & Jotin, 1994). From
this data, regression models have been constructed and the predictive performances of these models were assessed.
3.3. Pedestrian Crossing Speed and Crossing Direction
At pedestrian crossings, each pedestrian was monitored from the time he arrived at the crossing until he had successfully crossed
the street. The time taken by various pedestrian age, gender, and clothing style groups to cross the carriageway was measured and
Research Journal of Modeling and Simulation (2015) 34-41
37
the crossing speed was determined by dividing the carriageway width by the crossing time. The direction taken by the pedestrian
in the crossing process (perpendicular or at an inclined positions to the traffic flow stream) was also determined for each group.
Similar procedure was implemented by (Coffin & Morall, 1995; Edlmann et al., 2000)
4. Modeling of Pedestrian Walking and Crossing Speed
A statistical model is a formalization of relationships between variables in the form of mathematical equations. It describes how
one or more variables are related to one or more other variables. Regression analysis is a statistical process for estimating the
relationships among variables; it includes many techniques for modeling and analyzing several variables. The relation between
pedestrian’s speed (walking and crossing) and pedestrians’ characteristics (age group, gender, and clothing traditions); are
derived by using Stepwise regression analysis which is a statistical method that uses the relationships between two or more
quantitative variables to generate a model that may predict one variable from the other(s). A confidence level of 95 percent, thus
a significant level of 0.05 was chosen. The significance level (p-value) is the probability of obtaining results as extreme as the
one observed.
4.1. Definition of Variables
Several variables are used to construct the models of walking and crossing speed for Baghdad. These variables can be listed as
follows:
W= Pedestrian’s Walking speed (m/min)
C= Pedestrian’s crossing speed (m/sec)
g= Gender (male=2, female=1)
a= Age group (<18= 2, 18-50= 3, >50= 1)
x= Clothing style (Arab style= 1, western style= 2)
L= Land use (educational= 2, recreational= 1)
4.2. Checking Sample Size
Sample size was checked by using equation (1), (Kennedy and Neville, 1986), Table 2 shows the calculations of sample size for
each model.
Sample Size = (Z-score) ²* SD*(1-SD) / (significant level) ²
(1)
Where:
Z-score: constant value corresponding to the confidence level (for confidence level of 95%, Z-score= 1.96)
SD: standard deviation
Significant level: 0.05
Table 2. Sample Size Calculations
Standard deviation
N
Required N
Crossing speed (m/sec)
0.232
300
273
Walking speed (m/sec)
0.747
300
290
4.3. Checking for Outliers
Outliers are values that lie far away from the main group of data; the cause of a faulty observation can be a mistake. The outliers
and influential observations are checked by using Chauvinist’s criterion to examine outliers of data used to ensure accuracy,
(Kennedy & Neville, 1986). Table 3 shows the results of these tests; it can be seen that all results are less than the tabulated
values. Therefore, there is no outlier.
Table 3. Results of Chauvinist Test for Outliers
Dependent
variables
N
Minimum
(X min)
Maximum
(X max)
Walking
speed
300
19.17
51.02
34.33
7.47
2.03
2.23
3.14
Crossing
speed
300
0.73
1.82
1.202
0.232
2.03
2.66
3.14
Standard
deviation (s)
38
Research Journal of Modeling and Simulation (2015) 34-41
4.4. Testing of Normality
This step was implemented using Kolmogorov-Smirnov (K-S test). The K-S statistics D are based upon the maximum distance
between F(y) and Fn (y), that is shown in equations 2; 3; 4; 5:
D=max. [F(y) - Fn(y)]
(2)
Where:
F(y) = Normal cumulative probabilities (From normal distribution table)
Fn (y) = Sample cumulative distribution function.
i
D+= Max. [ - F(yi)]
n
D - = Max. [F(yi) -
(3)
i−1
n
]
(4)
Since:
D = Max (D+, D-)
Tables 4 show the results of normality tests for each model.
(5)
Table 4. One-Sample Kolmogorov-Smirnov Test for Baghdad Walking and Crossing Speed
N
Normal parameters Mean
Standard deviation
Most extreme differences Absolute
Positive
Negative
Kolmogorov – Simirnov Z
Asymp. Sig. (2 – tailed)
Walking speed
Crossing speed
295
34.330
0.7470
0.086
0.086
- 0.064
1.469
0.027
297
1.201
0.4227
0.080
0.080
- 0.057
1.373
0.046
4.5. Multicollinearty
Multicollinearty is a problem in multiple regressions that develops when one or more of the independent variables is highly
correlated with one or more of the other independent variables. Based on the intercorrelation analysis, the independent variables
are eliminated one-by-one depending on significance. The process is repeated until significant predictor variable remained, at
that point interactions among the variables are considered. A correlation matrix is produced to determine the correlation
coefficients for the variables. The decision to add or delete a variable is made based on weather that variable improves the model
or not. By using SPSS software, the correlation coefficients between all of the variables are calculated and the correlation matrix
is setup. Table 5 shows the bivariate correlation coefficients to identify the underlying form of the relationship between the
dependent variable and each of the predictor variables.
Table 5. Partial Correlation Matrix for Walking and crossing Speed in Baghdad
Dimension
1
2
3
4
5
Variance proportions for walking speed
Land
Age
Cloth
Constant
Gender
group
style
use
0.02
0.02
0.02
0.03
0.02
0.02
0.05
0.24
0.00
0.07
0.00
0.01
0.07
0.85
0.05
0.39
0.21
0.26
0.10
0.57
0.57
0.71
0.42
0.03
0.30
Variance proportions for crossing speed
Land
Age
Cloth
Constant
Gender
group
style
use
0.01
0.03
0.01
0.03
0.01
0.00
0.04
0.11
0.00
0.06
0.01
0.08
0.01
0.96
0.02
0.01
0.77
0.00
0.00
0.21
0.97
0.09
0.88
0.00
0.70
4.6. Stepwise Regression Models
The best and commonly method used to determine parameter of prediction model is stepwise method (Kennedy & Neville, 1986).
This method computes the simple regression model for each independent variable. The independent variable is with the largest
F-statistic, in other words, the smallest p-value is chosen as the first entering variable. SPSS software uses the F-statistics and the
standard is usually set at F=3.8, which is chosen because the significant level is about 5 %. The standard is called the F-to-enter.
If at least one variable exceeds the standard, the procedure continues. It then considers whether the model would be improved by
adding a second independent variable. It examines all such models to determine which is best and whether the F-statistic of the
Research Journal of Modeling and Simulation (2015) 34-41
39
second variable (with the first variable already in the equation) is greater than F-to-enter. If two independent variables are highly
correlated, only one of them will enter the equation. Once the first variable is included, the added explanatory power of the
second variable will be minimal and its F-statistic will not be large enough to enter the model.
Tables 6 show the model summary, while table 7 presents the Coefficients and summary of stepwise regression for
pedestrian’s walking and crossing speed for Baghdad.
Table 6. Model Summary
Model
Walking speed model
Crossing speed model
R
R square
Adjusted R2
SEE
R
R square
Adjusted R2
SEE
1
0.792
0.627
0.625
4.572
0.917
0.841
0.841
0.1762
2
0.853
0.728
0.726
3.908
0.926
0.858
0.857
0.1671
3
0.881
0.776
0.774
3.549
0.930
0.865
0.864
0.1630
4
0.906
0.820
0.818
3.187
0.933
0.871
0.870
0.1593
Table 7. Coefficients Summary for Walking and crossing speed Models
Model
Unstandardized
coefficients
Walking speed model
Standardized
coefficient
t
Sig.
Beta
Lower
bound
Upper
bound
0.0
30.266
0.0
5.029
-12.5
0.0
0.219
8.572
0.307
8.418
B
SEE
Constant
31.06
0.400
77.66
Land use
Age
group
6.14
0.567
0.411
10.83
- 2.88
0.231
- 0.331
3.319
0.387
4.696
0.558
Gender
Cloth
style
95 % confidence
interval for B
Unstandardized
coefficients
Crossing speed model
Standardized
coefficient
t
Sig.
Beta
95% confidence
interval for B
Lower
bound
Upper
bound
B
SEE
31.840
1.540
0.029
52.24
0.0
1.482
1.598
7.261
-0.424
0.016
- 0.836
-26.49
0.0
-0.455
-0.392
- 3.33
- 2.42
0.107
0.017
0.125
6.122
0.0
0.073
0.141
0.0
2.557
4.081
0.117
0.026
0.139
4.501
0.0
0.066
0.169
0.0
3.598
5.794
-0.061
0.019
- 0.072
-3.263
0.0
-0.098
-0.024
4.7. Analysis of Error
This analysis is to test the goodness of linear models and the errors that have a constant variance σ2 (homoscedasticity
hypothesis). This can be achieved by scatter plot for standardized residuals (which is the difference between an observed value
(yi) and the predicted value) on Y-axis and the estimated value of the dependent variable Ŷ on X-axis. The points should be
equally distributed about the zero line. Fig.6 and fig.7 show the scatter plot of standardized residuals and dependent variable Ŷ
for each model.
Fig. 6. Scatter plot of Residual and Walking Speed
Fig.7. Scatter plot of Residual and Crossing Speed
Fig.6 shows that points are equally distributed about the zero line, whereas fig.7 is suffering from the non-homogenous error
variance problem.
4.8. Analysis of Results
The analysis of results and calculation of standard error, regression, coefficient of variation for models are presented in Table 8.
40
Research Journal of Modeling and Simulation (2015) 34-41
Table 8. Summary Results of Final Models
Obtained models
Crossing speed W = 1.54 − 0.424a + 0.107g + 0.117x − 0.061L
Walking speed C = 31.053 + 6.145L − 2.882a + 3.319g + 4.696x
Where:
W= pedestrian’s Walking speed for Baghdad (m/min)
C= pedestrian’s crossing speed for Baghdad (m/sec)
g= gender (male=2, female=1)
a= age group (<18= 2, 18-50= 3, >50= 1)
x= clothing style (Arab style= 1, western style= 2)
L= land use (educational= 2, recreational= 1)
R2
Adjusted R2
SEE
0.886
0.820
0.884
0.818
0.143
3.187
4.9. Checking of R-Critical
A high correlation coefficient R-value does not guarantee that the model fits the data well. The correlation between x and y is
considered significant at the given probability level when the calculated R exceeds the tabulated R-value. Table 9 shows the
tabulated R-values for the predicted models. Therefore, there is strong correlation between predicted variable and independent
variables in this model.
Table 9. The tabulated R-values for the predicted models
Variable
WB
CB
N
300
300
R- Calculated
0.905
0.941
R- Tabulated
0.113
0.113
4.10. Models Limitation
The limitation of the data used to establish the models are presented in Table 10. The intention of the limitation is not to suggest
that the Modeling effort has not been successful. It merely serves to alert the limitations of the data.
Table 10. Summary of Models Limitation
Model
Walking speed
Crossing speed
Maximum
51.02
2.03
Minimum
19.17
0.3
Mean
34.33
.2
4.11. Validation of the Developed Models
The graphic plotting of observed and estimated data is a most useful method of evaluation the overall performance of a
regression equation. If the point which result from the plot of estimated with observed data tend to stand nearby the line drawn at
45o, then the result model is considered satisfactory. This can be done by data splitting in two sets. About 70% of data are used to
build the models and 30% of it is used for the validation process. Figures 8 and 9 show the resulting plots.
Fig. 8. Estimated Value of Pedestrian’s Walking Speed
Fig. 9. Estimated Value of Pedestrian’s Crossing Speed
Research Journal of Modeling and Simulation (2015) 34-41
41
5. Conclusions
Within the limitations of field investigation procedure and assumptions, the following conclusions may be drawn:
1. Male pedestrians have significantly faster walking speeds than female pedestrians do by about 5% with mean walking
speed of 35.9m/min.
2. Pedestrians of 18–50 years age group are the fastest group of pedestrians with an average speed of 43.092 m/min.
Pedestrians over 50 years old were found to be the slowest group with an average walking speed of nearly 20 m/min.
3. Males wearing western style are walking faster than males with Arabic style by an average of 3.9 m/min and such variation
was not significant for female pedestrians.
4. The mean free flow walking speed of Baghdad pedestrians that has been observed is comparatively slower than that of
others countries, and was found to be 63.586 m/min.
5. Male pedestrians have significantly faster crossing speeds than female pedestrians do with an average crossing speed of
1.285 m/sec.
6. Adult pedestrian of age group (18-50) are faster than elder while crossing for both genders and for all sites. Their crossing
speed was in the range of 1.44 – 1.2 m/sec.
7. The developed models could be used in planning and design for safer pedestrian facilities. Safety assessments can
profitably be linked with these models to detect the degree of exposure to risk in this vulnerable travel mode.
References
Abdulameer, M. W. (2014). Modeling Pedestrians Speeds in Baghdad and Erbil cities, MSc. Thesis, Department of Civil Engineering, College of Engineering,
University of Baghdad, Iraq.
Al-Masaeid, H. R., Al-Suleiman, T. I., & Nelson, D. C. (1993). Pedestrian speed-flow relationship for central business district areas in developing countries.
Transportation Research Record, (1396), 69–74.
Benz, G., & Fruin, J. (1984). Pedestrian Time-Space Concept for Analyzing Corners and Crosswalks, Transportation Research Record (959), Transportation
Research Board, USA
Coffin, A., & Morall, J. (1995). Walking Speeds of Elderly Pedestrians at Crosswalks, Transportation Research Record (1487), Transportation Research Board.
Daamen, W., & Hoogendoorn, S. P. (2003), Experimental Research on Pedestrian Walking Behavior. 82th Annual Meeting at the Transportation Research Board,
USA
Pitcairn, T. K., & Edlmann, T. (2000). Individual differences in road crossing ability in young children and adults. British Journal of Psychology, 91(3), 391-410.
Finnis, K. K., & Walton, D. (2006). Field observations of factors influencing walking speeds. Ergonomics.
Fruin, J. (1971). Pedestrian planning and design, Metropolitan Association of Urban Designer and Environmental Planners, Inc., New York.
Hoel, L. (1968). Pedestrian travel rates in Central Business Districts, Traffic Engineering and Control, January, UK.
Hoogendoorn, S. P., Daamen, W., Boer, A. D., & Vaatstra, I. (2007). Assessing passenger comfort and capacity bottlenecks in Dutch train stations.
Transportation Research Record: Journal of the Transportation Research Board, 2002(1), 107-116.
Kennedy, j., & Neville (1986). Basic statistical methods for engineers and scientists, third edition.
Khisty, C. J. (1982). Pedestrian Cross Flows in Corridors, Transportation Research Record 847, Transportation Research Board, USA.
Khisty, C. J. (1994). Evaluation of Pedestrian Facilities: Beyond the Level of Service Concept, Transportation Research Record 1438, TRB, National Research
Council, Washington, DC, USA.
Koushki, P. A. (1988). Walking Characteristics in Central Riydah, Saudi Arabia. Journal of transportation engineering, 114(6), 735-744
McGee, H. W., Hooper, K. G., Hughes, W. E., & Benson,W. (1983). Highway design and operational standards affected by driver characteristics: final technical
report. Report FHWA/RD 83/015. Washington. DC, FHWA, US Department of Transportation.
NZ Transport Agency (2009). Pedestrian Planning and Design Guide. New Zealand Government.
Polus, A., Schofer, J. L., & Ushpiz, A. (1983). Pedestrian flow and level of service. Journal of Transportation Engineering, 109(1), 46-56.
Sarsam, S. (2002). Modeling pedestrian crossing and walking behavior at Mosul CBD, Proceedings, Safety on roads: 2nd international conference 21- 23
October-2002 Bahrain SORIC` 02.
Sarsam, S. (2013). Assessing Pedestrian flow characteristics at Baghdad CBD area, 2nd Scientific Engineering Conference, University of Mosul, 19-21
November, Mosul, Iraq.
Sarsam, S., & Abdulameer, M. (2014). Evaluation of Pedestrians Walking Speeds in Baghdad City, Journal of Engineering, Vol.20, No. 9, September, Iraq.
Schneider, S., Bauer, D., Brändle, N., & Ray, M. (2008). Estimating Pedestrian Movement Characteristics for Crowd Control at Public Transport Facilities.
Proceedings of the 11th International IEEE Conference on Intelligent Transportation Systems, Beijing, China, October, 12-152.
Turvey, I. G., May, A. D., & Hopkinson, P. G. (1987). Counting methods and sampling strategies for determining pedestrian numbers, Institute for Transport
Studies, ITS Working Paper, 242, University of Leeds, UK.
Transportation Research Board (2000). HCM, Highway Capacity Manual, Special Report 210, National Research Council, Washington, USA.