See discussions, stats, and author profiles for this publication at: [Link]

net/publication/312298511

Use of Multinomial Logit Model in Evaluation of Service Levels of Pedestrian

Facilities

Conference Paper · December 2016

CITATIONS READS

3 288

3 authors:

Rima Sahani Dudi Praveena

National Institute of Science & Technology 1 PUBLICATION 3 CITATIONS
19 PUBLICATIONS 52 CITATIONS
SEE PROFILE
SEE PROFILE

Prasanta Bhuyan
National Institute of Technology Rourkela
87 PUBLICATIONS 246 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Modified INAFOGA method for critical gap estimation at u-turn median openings View project

Entry Capacity Modelling of Roundabouts under Heterogeneous Traffic Flow Conditions View project

All content following this page was uploaded by Prasanta Bhuyan on 13 January 2017.

The user has requested enhancement of the downloaded file.

 Abstract Number: 85

Use of Multinomial Logit Model in Evaluation of Service Levels

of Pedestrian Facilities
Rima Sahani1, Dudi Praveena2, P. K. Bhuyan3
1
Ph. D Research Scholar, Department of Civil Engineering, NIT Rourkela, India,
Email: luckyrima44@[Link]
2
Former M. Tech Student, Department of Civil Engineering, NIT Rourkela, India,
Email: dudipraveena@[Link]
3
Assistant Professor, Department of Civil Engineering, NIT Rourkela, India,
Email: pkbtrans@[Link]

Abstract
Main aim of this study is to provide service measures for walking facilities such as sidewalk, signalized intersection
and un-signalized intersection using pedestrian perception data sets. Based on factors which are mainly affecting the
pedestrian’s satisfaction level like traffic, safety, comfort, maintenance and aesthetics proper questionnaire has been
prepared for these three facilities. Multinomial logit modeling concept has been applied to estimate probability and
category of satisfaction levels of different facilities. Analysis shows that 38.4% of total pedestrians were having
problems while walking due to illegal on-street parking and vending encroachment as they covers maximum space
of the sidewalks.

Keywords: Pedestrian Perception, Walking Facilities, Service levels

1. Introduction
In a developing country like India walking is an important way of mobility with a high percentage of mode share.
However transportation planners and designers are getting more and more inclined towards motorized mode, which
leads to discourage people to walk. Undoubtedly, in sustainable transportation system walking is an indispensable
mode but unfortunately not enough attention is paid in development of safe and supportive infrastructure for
pedestrian facilities. To improve an environment as pedestrian friendly i.e. comfortable and safe the primary job is
to assess the satisfaction levels of that facility; which can be directly evaluated from the perception of the road users
as it is a qualitative aspect. In this paper an attempt has been made to provide a service measure for walking
facilities such as sidewalk, signalized intersection and un-signalized intersection, which will help the road planners,
designers and field engineers to assess the quality of the facility for the improvement.

Pedestrian Level of Service (PLOS) is a complex term which represents the operating condition of pedestrian
facilities and level of comfort pedestrian experience in using these facilities as Level of Service (LOS) indirectly
represents the actual feeling of the people. Although the concept of LOS is meant to reflect road user perception but
most of the researchers focused on operational characteristics and road geometric parameters to measure service
parameters of facilities meant for pedestrians. The information of which factor affecting the PLOS and to what
extent will be known from pedestrians based on their daily experiences about any pedestrian facility. So, it is
important to measure the qualitative service level rather focusing only on quantitative way of measurement.
Many researchers have done experiments in order to assess service levels of pedestrian facilities. Although LOS
concept was introduced in Highway Capacity Manual (HCM) (1965), after several versions of publications, the
version of HCM (2010) improvised the definition as LOS is a quantitative satisfaction of performance measure
which represents the quality of service. Recent approach on pedestrian service model developed by Gallin (2001)
has not considered the pedestrian perception and in Trafitech model developed by Jensen (2007) has not addressed
issues on perception biasness of pedestrians. Muraleetharan et al. (2005) have developed a method for the
determination of PLOS at intersections using a stepwise multi-variable regression analysis. This study concludes
1
that ‘turning vehicles’ have more influence on PLOS than other factors and the safety of pedestrians decreases with
increase in turning vehicles. Lee et al. (2005) presented a new set of LOS standards for pedestrians’ crosswalks at
signalized intersections, where authors experimented that pedestrian LOS is negatively influenced by bi-directional
pedestrian flow. Dandan et al. (2007) model did not consider the parameters of actual dimensions and quantity such
as width of the facility. Araujo and Braga (2008) have proposed a different methodology for the evaluation of PLOS
of crossing pedestrians at signalized intersections by application of qualitative variables. For assessing the
perceptions of the subjects based on Paired Comparison and Constant Sum, psychometric methods were used by the
authors.
Rahaman et al. (2012) have developed a model using the Analytical Hierarchy Process (AHP) in the order to
prioritize pedestrians walking environment through shopkeeper’s expectations measures. Kang et al. (2013)
proposed a model for the assessment of sidewalk level of service in the presence of bicyclists employing a statistical
analysis technique such as Ordered Probit method. It was concluded that pedestrian perceptions on LOS are strongly
influenced by the pedestrian flow rate. Archana and Reshma (2013) have developed a pedestrian LOS model for
signalized intersections using a stepwise regression analysis. Authors concluded that pedestrian flow rate, crossing
time and surface condition of the crosswalk influences PLOS more than any other parameters and suggested that
with reduction in crossing time of pedestrians, cross pedestrian flow rate increases and average pedestrian delay
decreases by which high levels of service can be attained. Bian et al. (2013) have developed a PLOS model for un-
signalized intersections based on user perception and field quantitative parameters with the help of step-wise
regression analysis. In this study authors have shown that PLOS is highly influenced by conflicting motorized and
non-motorized vehicular traffic. Left, right and through bicycles movement from side streets, presence of median,
right and left turning bicycles from the street parallel to crosswalk, through bicycles approaches from opposite
direction of the crosswalk were also observed to influence PLOS. Ling et al. (2014) have proposed a modified
method of pedestrian video simulation (PCVS) used to estimate the pedestrian level of service and found that
increase in delay, turning traffic, and mixed two-wheeler volume have negative impact on LOS but pedestrian
volume and refuge island presence and two-step crossing improves LOS. Kong and Yang (2010) investigated the
association between the impact speed and risk of pedestrian casualties in passenger vehicle collisions with the
application of multinomial logistic regression based on real-world accident cases in China. Guo et al. (2014)
examined the preferences for pedestrians of crossing locations and the influencing factors in choosing a street to
cross. Authors concluded that diversion distance is the most influencing factor in decision making and pedestrians
prefers under or over pass for safety purpose.
Although many researchers have considered several factors contributing quantitative estimation of LOS, but user’s
perception in defining PLOS has not been given due considerations particularly in a developing country like India.
The main objective of this study is to develop a set of models for pedestrian facilities (sidewalk, signalized
interception, un-signalized intersection) offering a hierarchy of service levels (A-F). The focus of this study is as
follows: i) to identify factors affecting pedestrian service levels, ii) to design an innovative questionnaire for each
facility in order to estimate PLOS, iii) to develop PLOS models of three pedestrian facilities by using multinomial
Logit analysis over a set of perception data collected from nine mid-sized cities.
2. Methodology
In order to asses quality of service of walking facilities such as sidewalk, signalized intersection, un-signalized
intersection, analysis has been carries out by taking the on-site perception of pedestrians. Questionnaires used for
perception data collection have been prepared by considering some primary variables. The collected perception data
are then investigated to get service levels of pedestrian facilities by developing probability PLOS models of each
facility.
2.1 Variable Selection and Questionnaire development
For perception survey questionnaires includes five primary variables with some common socio-demographic
variables. These demographic variables are asked to the participants of each facilities to observe the pedestrian
characteristics such as age, gender, education, occupation, purpose of trips, house hold size, number of vehicles
owned per house hold, platoon size, main mode of travel etc. Other than these questions five primary variables such
as traffic, safety, comfort and convenience, road maintenance and aesthetics effects on pedestrian movement are
taken into consideration to estimate service levels of the facilities. For the three facilities considered in this study
these five variables contain several attributes (questions) under each. The final form of questionnaires has been
decided after a thorough pilot survey. Pedestrians are asked to rate each attributes in a five rating scale with respect
to their satisfaction level perceived while using the facility.
2
To assess the service quality under ‘traffic’ variable for sidewalk facilities, pedestrians were asked about whether
they faced problems due to illegal turns of vehicles, high speed of traffic, vehicles swing towards pedestrians, illegal
parking or presence of heavy vehicles and non-motorized vehicles. In case of intersection, participants were asked to
rate their satisfaction level upon the effect of multiple factors such as traffic speed, waiting time, crossing time, red
phase timing in a cycle, left turning and right turning vehicles and problem due to continuous flow of traffic etc. In
order to assess effect of ‘safety’ variable on the overall perceived satisfaction following questions were included:
discomfort caused due to aggressive driving on the main carriageway, undesirable movement intrusion of cars on
sidewalk and positive effect of barriers between sidewalk and main carriageway if provided. Under the primary
variable ‘intersections’, attributes like apparent danger anticipated during crossing, casual approach in attaining a
mobile call while crossing, readiness of drivers to give space for pedestrian crossing, dishonor the traffic rule by
vehicular movement even in red phase etc. are asked. ‘Comfort’ variable contains attributes like comfort provided
by sidewalk, pleasure felling during walking, step out of the way to avoid other pedestrian, heavy crowd and
development of streetscapes. To assess comfort at intersections pedestrians are asked concerning street light facility
during night, adequacy of space and time to cross, availability of zebra crossing and need of grade separated
crossing etc. Under ‘maintenance’ pedestrians are asked about signage and signage of roadside features, operational
quality, maintenance quality, visual cluster of lights and pavement markings of the road. Whereas under the variable
‘aesthetics’ questions were asked about presence of median, designated roadside vending zone, effect of adverse
weather, and need of any improvement in near future. This study converted the perceived scores of each variable
from discrete to continuous in a normalized 10 point scale as continuous data work much better in model fitting. In
the questionnaires, one question was set aside to ask about ‘overall satisfaction’ in a 6 point rating scale. Different
notions used are: 1, 2, 3, 4, 5 and 6 for extremely satisfied, very satisfied, somewhat satisfied, somewhat dissatisfied,
very dissatisfied and extremely dissatisfied. This ‘overall satisfaction’ rating has been used as the dependent variable
and five variable scores such as traffic score (T), safety score (S), comfort score (C), maintenance score (M) and
aesthetic score (A) are used as independent variables for the model development.
2.2 Model development for pedestrian service levels
In order to determine the probability categories of PLOS (A-F) statistical approach is employed. Here multinomial
logit regression analysis has been used as this technique is most useful for understanding the effect of several
independent variables on a single dependent variable. Here the dependent variable ‘y’ having k outcomes as 1, 2, 3,
4, 5 and 6 then,

p( y = 1)
ln( ) = b1 xi Þ p( y = 1) = p( y = 6) exp(b1 xi )
p ( y = 6)
p ( y = 2)
ln( ) = b 2 xi Þ p( y = 2) = p( y = 6) exp(b 2 xi )
p ( y = 6)
. .
. .
. .
p( y = 5)
ln( ) = b 5 xi Þ p ( y = 5) = p( y = 6) exp(b5 xi ) (1)
p ( y = 6)

Where x is a vector of variable such as traffic score, safety score etc. which are used for determining perceived
satisfaction levels and β is estimable coefficient. As it is known that sum of probabilities of all the outcomes should
be equal to one then,
P(Y=1) +P(Y=2) +…………….P(Y=k-1) +P(Y=k) = 1,
and by adding all the left and right hand side terms of equation (1)

p( y = 1) + p( y = 2) + ........ + p( y = 5) = p( y = 6)[exp(b1 xi ) + exp(b 2 xi ) + .... + exp(b 5 xi )]

é 5 ù
å
1 - p( y = 6) = p( y = 6)ê exp(b 6 xi )ú
êë k =1 úû

3
 é 5 ù
å
1 = p( y = 6) ê exp(b 6 xi )ú + p( y = 6)
êë k =1 úû

é 5 ù
1 = p( y = 6) ê1 +
êë
å exp(b x )úú 6 i
k =1 û

1
p( y = 6) =
é1 +
å exp(b 6 xi )ù
5
êë k =6 úû
Substitute P(y=6) in equation (1)

exp( b1 xi )
p( y = 1) =
é1 +
å exp( b 6 xi )ù
5
êë k =1 úû
exp( b 2 xi )
p ( y = 2) =
é1 +
å exp( b 6 xi )ù
5

ëê k =1 úû
.
.
.
exp( b 5 xi )
p( y = 5) = (2)
é1 +
å exp( b 6 xi )ù
5
êë k =1 úû

Except the reference category, remaining all the categories values are taken as average, so that there will be only
single value for each score for a category. Likewise, for the other categories averages are calculated and summations
are found out for the logit values of all categories. For each response, the probability of satisfaction level is
calculated for all categories. The maximum probability value among all categories is the predicted probability of
satisfaction level for that respondent and the corresponding category are taken as the predicted category. The basic
multinomial logit model used in this study is be represented as:
log it ( p) = a i + bi X i (3)

Where αi is the intercept parameter of the response level of satisfaction, βi is the coefficient of parameter and Xi is
the parameter score.
3. Study Are and Data Collection
In this research, perception data collection scheme was implemented in order to measure the typical service factors
of pedestrian facilities. Responses of pedestrians have been collected from nine mid-sized cities of India, i.e.
Bhubaneswar, Cuttack, Ranchi, Raipur, Jamshedpur, Vijayawada, Visakhapatnam, Vizianagaram and Tirupati
having population in the range of 0.3-1.2 millions. These nine cities from where data sets were collected are shown
in Fig.1. These cities are favorably selected as diverse activities have been observed from residential, commercial,
official, educational, recreational areas etc. to get data sets from varied class and community. Data has been
collected from three important pedestrian facilities such as sidewalk, crosswalk at signalized and un-signalized
intersections having diverse approach road characteristics operating under the influence of heterogeneous traffic
flow. Roadways are characterized by having access facilities, on-street parking provisions, nearby vending zones
along 2 to 4 lane roads with grade separated footpaths in some cases. Intersections are connected with 3-5 approach
legs, which are either major roads, secondary arterial roads or minor roads. Some intersections are well marked with
zebra crossings and provided with advanced traffic signals and some are with very poor condition of signals and
markings. Data have been collected in five phases between July-December of 2015 during 9-11 AM in the morning
and 4-6 PM in the afternoon from different cities.

4
Three different (for 3 pedestrian facilities) convenient and proper questionnaires have been prepared and conformed
through pilot survey using which participants satisfaction levels are collected in terms of ranking. At first
participants are convinced about the purpose of data collection and then are asked to fill the questionnaires to assess
their real time experience after using the facilities. Around 1425, 630 and 675 effective pedestrian’s perception were
collected at all sidewalks, un-signalized and signalized intersections respectively. As view on a single item changes
from person to person, perception data had been collected from different gender, age group and platoon size. For
2730 pedestrians responded at all the 3 facilities demographic values of the collected data are shown in the Table 1.
From the total data sets 56.2% are male participants, 38% are of young age and 65.8% of pedestrian walk single
during data collection.

Fig. 1. Study locations of selected cities of India

Table 1. Distribution statistics of demographic variable of surveyed respondents

Total data
Variable Description
Frequency Percentage (%)
Male 1535 56.2
Gender
Female 1195 43.8
Young (15-35) 1039 38.1
Age Middle(35-55) 1011 37
Old(>55) 680 24.9
1 1797 65.8
Platoon
2 474 17.4
size
more than 2 459 16.8

The percentage of each category for different demographic variables such as gender, age and platoon size are
represented using bar charts as shown in Fig. 2.

a) Overall satisfaction responses with respect to gender (b) Overall satisfaction responses with respect to age

5
(c) Overall satisfaction responses with respect to platoon size
Fig.2. Overall satisfaction responses with respect to demographic variables

4. Result and Analysis

Responses were collected using questionnaire which was given on six point scale ranging from extreme satisfaction
to extreme dissatisfaction. Ratings given to all the questions under a particular parameter were calculated as score of
that particular parameter. As, the outcome variable having more than two categories, the multinomial logit model
(MLM) is used in this study. In multinomial logit model, the odds ratios of independent parameters are not same
across the response category. For each category of outcome variable the model developed is different.

The perception data collected at the three facilities from all the cities are analyzed to develop pedestrian’s service
level model. Using likelihood ratio, the data fitness to model was known. Two values for -2loglikelihood is
computed i.e. one is for intercept only (without using any parameters) and the other one is for final model (with all
independent variables). The values from each model are subtracted from one another. The resulting value is Chi-
square, which resembles good model fit with higher Chi-square value with significance <0.05. Values for model
fitting criteria and likelihood ratio test for three facilities are shown in Table 2.
Table 2. Fitness of user perception data to the model for sidewalks using MLM

Model Fitting Criteria Likelihood Ratio Tests

Facilities Model
Degrees of
-2 Log Likelihood Chi-Square Sig.
freedom
Intercept Only 3929.859
Sidewalk
Final 3424.341 505.518 85 0
Signalized Intercept Only 2202.89
Intersection Final 1999.316 203.574 70 0
Un-signalized Intercept Only 3929.859
intersection Final 3424.341 505.518 85 0

In order to get the probabilities of different service categories of sidewalk facilities 70% of the pedestrian perception
data has been used for model fit purpose. Applying MLM technique equation (1) gives the following results;
For
p(1)
Satisfaction level 1: S1= = e 0.686 -1.587 *T - 0.809 * S - 0.59 *C - 0.27 * M - 0.725 * A (4)
p(1)
p ( 2)
Satisfaction level 2: S2= = e11.038 - 0.829 *T - 0.295 * S - 0.142 *C - 0.272 * M - 0.641* A (5)
p ( 6)
p(3)
Satisfaction level 3: S3= = e 8.631- 0.321*T - 0.041* S - 0.432 *C - 0.103 * M - 0.374 * A (6)
p (6 )
p ( 4)
Satisfaction level 4: S4= = e 6.402 - 0.232 *T - 0.181* S - 0.494 *C - 0.132 * M - 0.006 * A (7)
p ( 6)
p(5)
Satisfaction level 5: S5= = e 6.535 - 0.363 *T - 0.156 * S - 0.249 *C - 0.084 * M - 0.099 * A (8)
p (6 )
To measure the probability of each category,

6
 S1
Probability of satisfaction level 1: p (1) = (9)
1+ S

S2
Probability of satisfaction level 2 p ( 2) = (10)
1+ S
S3
Probability of satisfaction level 3 p (3) = (11)
1+ S

S4
Probability of satisfaction level 4 p ( 4) = (12)
1+ S
S5
Probability of satisfaction level 5 p (5) = (13)
1+ S
1
Probability of satisfaction level 6 p (6 ) = (14)
1+ S
Where S= S1+S2+S3+S4+S5

Before getting the value of S, satisfaction value S1 has been calculated by taking the averages of pedestrian perusing
the option extremely satisfied with the facility. Similarly the averages of satisfaction values S2, S3, S4 and S5 has
been calculated. Then S is calculated by the summation of all satisfaction values. Probability value of satisfaction
level 6 is same for all responses as it is the reference category. For each pedestrians response calculate all
probabilities of satisfaction level. The maximum probability among them is the predicted probability of satisfaction
level and the corresponding category is the predicted category for that response. Pedestrian perception values show
that the mid value of PLOS is 3.5. For the categorization the ranges of PLOS are taken are shown in Table 3.
Table 3. Ranges of PLOS categories for pedestrian facilities

PLOS Facility
Ranges
Category Condition
A <1.5 Excellent
B 1.5-2.5 Good
C 2.5-3.5 Average
D 3.5-4.5 Inferior
E 4.5-5.5 Poor
F >5.5 Terrible

For validation purpose taking the data of Vijayawada city, analysis shows that the average service level of the city is
3, which represents the PLOS category ‘C’. Also from the results it has been found that the average probability of
PLOC ‘C’ is 0.362. Detail values of the probabilities and categories for sidewalks of Vijayawada city is shown in
Table 4.
Table 4: Predicted category of each sidewalk for Vijayawada city

average probability average service Predicted

Sidewalk
of satisfaction level level of facility category
Benz circle road 0.3473 2.93 C
DV Manor road 0.3307 3.27 C
Govt. hospital road 0.314 3.27 C
Nirmala convent road 0.3827 2.73 C
Prasad V Potluri road 0.3927 2.8 C
Ramesh hospital road 0.3627 2.93 C
Indira Gandhi stadium road 0.4033 2.53 C

Similarly analysis has been carried out for all the three facilities applying the perception data of nine cities. The
coefficients of each parameter for each category of satisfaction level, for all the 3 facilities (Sidewalks, Signalized
intersections and Un-signalized intersections) are shown in Table 5. These models give different intercepts of
independent variables for each category, which implies the odds ratio is different across the response category.

7
Table 5. Coefficients of each parameter for each satisfaction level using Multinomial logit model

sidewalks signalized intersections Un-signalized intersections

Satisfaction Variables Coefficients Sig. Coefficients Sig. Coefficients Sig.
Intercept
traffic -1.587 0 -1.605 0 -0.228 0
Excellent safety -0.809 0 -0.328 0.034 0.622 0
(PLOS A) comfort -0.59 0.049 -0.07 0.021 -0.633 0.049
maintenance -0.27 0.009 0.651 0.039 0.405 0.044
aesthetics -0.725 0 -0.315 0.002 -0.155 0
Intercept 11.038 0 9.945 0 4.743 0
traffic -0.829 0 -1.073 0 -0.407 0
Good safety -0.295 0.024 -0.436 0.007 0.319 0.024
(PLOS B) comfort -0.142 0.004 0.272 0.044 -0.046 0.034
maintenance 0.272 0.013 0.894 0 0.182 0.029
aesthetics -0.641 0 -0.513 0.002 -0.078 0
Intercept 8.631 0 1.638 0 2.32 0
traffic -0.321 0.034 -0.508 0.014 -0.368 0.034
Average safety -0.041 0.003 -0.142 0.049 0.163 0.04
(PLOS C) comfort -0.432 0.002 0.13 0.004 -0.074 0.002
maintenance 0.103 0.004 0.523 0.004 0.421 0.037
aesthetics -0.374 0.001 -0.239 0.002 0.028 0.001
Intercept 6.402 0.001 3.571 0 2.118 0.001
traffic -0.232 0.004 -0.289 0.002 -0.342 0.04
Inferior safety -0.181 0 -0.29 0.001 0.246 0.019
(PLOS D) comfort -0.494 0.001 -0.006 0.007 0.318 0.001
maintenance -0.132 0.045 0.218 0.022 -0.044 0.045
aesthetics 0.006 0.004 0.003 0.008 -0.074 0.03
Intercept 6.535 0.001 2.82 0 6.623 0.001
traffic -0.363 0.033 -0.473 0.03 -0.182 0.033
Poor safety -0.156 0.002 -0.267 0.022 -0.005 0.048
(PLOS E) comfort -0.249 0.001 0.132 0.047 0.102 0.014
maintenance 0.084 0.043 0.327 0.002 0.184 0.013
aesthetics -0.099 0.044 -0.283 0.001 0.01 0.041

The predicted service categories of different cities using the proposed probability models for three facilities are
shown in Fig.3. Sidewalk analysis for Cuttack city shows that 22% of pedestrians rate the facility as of good service
level where as 75% have given average rating and 3% of poor rating. The poor service quality is due to the presence
of heavy traffic with no specified sidewalks in the commercial areas. In Ranchi city 13% of participants have given
the rating of LOS E (poor facility) and 2% given as LOS E (terrible condition). This occurs because, vending
activities covers the most part of the sidewalks so that walkers has to use main carriage way while moving and have
to face conflict with the vehicles. In other cities nearly about 15% of participants have given rating as good and 70%
given rating as average service levels.

While considering about signalized intersection in Raipur city pedestrians got highest 23% of good service quality
than other cities. In Jamshedpur city 30% of pedestrian feels that they were getting inferior service quality at
signalized intersections and 65% were having average kind of service. Un-signalized intersection data analysis
shows that the cities are neither having excellent service nor poor or terrible service quality. But most of the cities
were offering average kind of service. In Cuttack and Ranchi cities according to 60% of pedestrian, un-signalized
intersection are providing inferior service quality. For Vishakhapatnam the 50% of participants have given average
rating whereas for Vizianagaram the percentages of ratings were 8.9%, 51.1% and 40% for good, average and
inferior service quality. It has been found from the analysis that any of the pedestrian facilities are not providing
excellent service levels; which explains that pedestrians expect better facilities for walking purpose.

8
(a) Percentage of pedestrian satisfaction for each category of sidewalks facilities

(b) Percentage of pedestrian satisfaction for each category of signalized intersections

(c) Percentage of pedestrian satisfaction for each category of un-signalized intersections

Fig. 3. Percentage of pedestrian satisfaction using Multinomial logit model

4. Conclusi on
Analysis in the current study shows that variables like traffic influence, safety, comfort, maintenance and aesthetic
are having significance effect on satisfaction level determination as p-value are around 0.000 for the three facilities
(sidewalk, signalized intersection, un-signalized intersection). During perception survey, from the total 2730
participants 1535 are male, 1039 are of young age (15-35 age), and 1797 pedestrians were walking single.
Demographic data analysis shows that male pedestrians were 7% more satisfied than female where as younger
pedestrian are 2% more satisfied than middle and old age. Also it has been found that 45% of the participants have
given average ratings to the different pedestrian facilities. It is noted that satisfaction levels of single pedestrian
always is in higher range than the pedestrian moving in group. This was may be due to 66% of the total participants
are isolate walkers.
From the analysis it has been observed that very less number of pedestrian facilities were offering very high
satisfaction level or extreme dissatisfaction among users. This is occurring may be due to the reason that study areas
taken in this study are of mid-sized cities having population size less than a million. In these cities there are not

9
 severe problems as perceived by the pedestrians as of now. Under do nothing situation over a period of time will
reduce service level further and will also add to pedestrian displeasure on road facilities. 43% of total participants
have given opinion that due to heavy vehicular flow and illegal turns of vehicles; pedestrian faces more difficulties.
38.4% pedestrians perceived that illegal on-street parking and vending encroachment covers the maximum space of
the sidewalks so that reduced space is left for walker. According to views of 43.2% of participants, activities of
some aggressive drivers creates dangerous environment for other road users. Pedestrian service measures evaluated
using the model developed in this study can provide an easy understanding about the condition of the walking
facilities. Roadway designers can use the service measure to improve the quality or to provide any alternate design
to achieve better facility.
References
Araujo, G.P., Braga, M.G.C. (2008). “Methodology for the qualitative evaluation of pedestrian crossings at road
junctions with traffic lights”, Transportation, pp. 539–557.
Archana, G., Reshma, E.K. (2013). “Analysis of Pedestrian Level of Service for crosswalk at intersections for urban
condition”, International Journal of Students Research in Technology and Management, Vol.6, pp.604-609.
Bian, Y., Lu, J., Zhao, L. (2013). “Method to Determine Pedestrians Level of Service for Un-signalized
Intersections”, Applied Mechanics and Materials, Vol.253-255, pp.1936-1943.
Dandan, T., Wei, W., Jian, L., Yang, B. (2007). “Research on Methods of Assessing Pedestrian Level of Service for
Sidewalk”, Journal of Transportation Systems Engineering and Information Technology, Science Direct, Vol. 7(5),
pp.74-79.
Gallin, N. (2001). “Pedestrian friendliness -guidelines for assessing pedestrian level of service”, Paper presented at
the Australia: Walking the 21st century conference, Perth, Australia.
Guo, H., Zhao, F., Wang, W., Zhou, Y., Zhang, Y., Wets, G. (2014) “Modeling the Perceptions and Preferences of
Pedestrians on Crossing Facilities”, Discrete Dynamics in Nature and Society, Hindawi Publishing Corporation,
Vol.2014, pp.1-8.
Highway Capacity Manual, (1965), Transportation Research Board, Washington, D.C.
Highway Capacity Manual, (2010), Transportation Research Board, Washington, D.C.
Jensen, S., U. (2007). “Pedestrian and Bicyclist Level of Service on Roadway Segments”, Transportation Research
Record. Journal of the Transportation Research Board, 2031, pp 43-51.
Kang, L., Xiong, Y., Mannering, F.L. (2013). “Statistical analysis of pedestrian perceptions of sidewalk level of
service in the presence of bicycles”, Transportation Research Part A, Vol.53, pp.10–21.
Kong, C., Yang, J. (2010), “Logistic Regression Analysis of Pedestrian Casualty Risk in Passenger Vehicle
Collision in China”, Accident Analysis and Prevention, Vol.42, pp. 987-993.
Lee, J.Y.S., Goh, P. K., Lam, W.H. K. (2005). “New Level-of-Service Standard for Signalized Crosswalks with Bi-
Directional Pedestrian Flows”, Journal of Transportation Engineering, ASCE, Vol. 131(12), pp. 957–960.
Ling, Z., Ni, Y., Cherry, C.R, Li, K. (2014). “Pedestrian Level of Service at Signalized Intersections in China Using
Contingent Field Survey and Pedestrian Crossing Video Simulation”, TRB Annual meeting.
Muraleetharan T., Adachi, T., Hagiwara, T., Kagaya, S. (2005). “Method to determine Pedestrian Level of Service
for Crosswalks at Urban Intersection”, Journal of the Eastern Asia Society for Transportation Studies, Vol.6,
pp.127-136.
Rahaman, K.R., Lourenco, J. M., Viegas, J.M. (2012). “Perceptions of Pedestrians and Shopkeepers in European
Medium-sized Cities: Study of Guimaraes, Portugal”, Journal of Urban Planning and Development, ASCE, Vol.
138(1), pp.26-34.

10

View publication stats