International Journal of Recent Technology and Engineering (IJRTE)

ISSN: 2277-3878, Volume-8 Issue-5, January 2020

Determination of Pedestrian Level of Service at

Signalized Midblock Locations for Mixed Traffic
Conditions
Teja Tallam, K. M. Lakshmana Rao

Abstract: walking is considered as one of the most important Pedestrian crossings are broadly classified as, at grade and
modes of transportation in India. But it is observed that the Grade separated pedestrian crossings.
facilities for the pedestrians are ignored during design, planning 1. At grade pedestrian crossing, is defined as the place where
and maintenance stage. But these days due to increase in pedestrians cross the carriage way at the same level as that
population in urban areas, traffic congestion has become a major
of the vehicular movement. This includes crossings at
problem for safe pedestrian crossing. It is necessary to objectively
quantify how well roadways accommodate pedestrian travel. intersections and midblocks.
Estimation of pedestrian level of service (PLOS) is the most 2. Grade separated pedestrian crossing, is defined as the
common approach to assess quality of operations of pedestrian place where pedestrians and vehicles cross the carriage
facilities. Due to more urbanisation and also large distance way at different level. Foot over bridges, underpass and
between the successive intersections people are forced to cross at overpass comes under this type of pedestrian crossing.
their respective midblock. Pedestrians and motorcyclists make up to half of those killed
This paper aims in understanding pedestrian characteristics or on the roads, highlighting the necessity for these road users
pedestrian behaviour which is a fundamental in pedestrian give more consideration in road safety programs.
planning process and finding the level of service for the
Unsurprisingly, thickly populated countries have higher
pedestrians (PLOS) at selected signalised midblock. Pedestrian
data required was collected using video graphic technique during numbers of pedestrian deaths (as per year 2009) which
two peak hours in a day at Kukatpally and Nizampet signalised includes China, India, and the Russian Federation. As
midblocks in Hyderabad city. The factors considered for the indicated by The Times of India report, "pedestrians
calculation of PLOS are their delay, crossing time, speed, density represent 33% street fatalities". In the previous 5 years, 666
and volume of pedestrians. people on foot have lost their lives in deadly incidents in
Greenshields’s macroscopic model was used to resolve Hyderabad police Commissioner-ate restrains alone.
important parameters like free speed (vf) and jam density (kj) by This paper includes the study of pedestrian behaviour at
plotting their respective graphs. Finally, regression analysis is signalised midblocks. Safety, comfort, convenience and
carried in R software to calculate pedestrian LOS using the above
minimum delay are the important parameters considered for
factors considered. Clustering technique is used to obtain the
LOS scores for the collected pedestrian data. LOS calculated the pedestrians at midblocks. Therefore it is necessary to
from model outputs is compared with the values in Indo HCM evaluate the quality of the crosswalk with these qualitative
2017. parameters. Level of service (LOS) is generally used to
measure the quality of traffic service, in this paper this is
Key words: Pedestrian LOS, Signalised Midblock,
used to measure the quality of pedestrian service at selected
Greenshield’s Model, Indo HCM 2017, R studio.
signalised midblocks based on their crossing speed and
I. INTRODUCTION
delay. The parameters delay (D), speed (V), density (K) and
flow (Q) are retrieved manually from the videos taken, by
Walking is considered as as one of the most significant mounting the camera at certain elevation. Greenshield’s
modes to commute in Indian situation. It is seen that a large model is used for the calibration of pedestrian data. The
portion of the pedestrian issues are disregarded during retrieved data is put in the Greenshield’s formulae to
design and planning stage. Until the mid-seventies, calculate jam density (Kj), maximum flow (Qmax), free
pedestrian traffic has scarcely been the subject of research flow speed (Vmax), volume capacity ratio and their
where just pedestrian behaviour was seen by watching and respective graphs are plotted.
inferring simple theories and by inferring models to decide The main aim is to determine LOS of the pedestrians using
the behaviour of pedestrians. Recently, where more clustering technique based on delay and v/c ratio. Clustering
advancements that are recently increased the technology is a statistical technique of partitioning a sample into
became available and pedestrian models have been easily homogeneous groups to produce an operational
developed to understand their behaviours and to measure the classification. LOS scores are determined by developing the
quality of services. In spite of these advancements, no number of clusters using K mean clustering analysis. The
legitimate consideration is given to pedestrian facilities. algorithm is run in the R program using sapply() function
Since no appropriate consideration is given to these over the K range, where k is the number of clusters chosen.
facilities, pedestrian mishaps have transcendently expanded.

Revised Manuscript Received on January 15, 2020.

* Correspondence Author
Teja Tallam*, Research Scholar, Jawaharlal Nehru Technological
University Hyderabad, India. Email: [email protected]
K. M. Lakshmana Rao, Professor, Jawaharlal Nehru Technological
University Hyderabad, India. Email:[email protected]

Published By:
Retrieval Number: E6160018520/2020©BEIESP Blue Eyes Intelligence Engineering
DOI:10.35940/ijrte.E6160.018520 2751 & Sciences Publication
Determination of Pedestrian Level of Service at Signalized Midblock Locations for Mixed Traffic Conditions

II. LITERATURE SURVEY use condition, number of vehicles encountered, median

width and number of lanes have significant effect on
Several studies have been conducted by the various
pedestrian LOS at unsignalised midblock crosswalks.
researchers on pedestrian behaviour and pedestrian level of
Asaithambi et.al. (2016) investigated pedestrian road
service in urban areas. Some of them are listed as below:
crossing behaviour of an intersection before and after signal
Axler (1984) suggested more warrants and other
installation. Parameters like crossing patterns, waiting time,
considerations for grade separated pedestrian crossings. This
crossing time, and pedestrian crossing speed, pedestrian
included pedestrian hourly volume, vehicle volume, any
gender, age, types of vehicles and flow were retrieved from
physical barrier like underpass, overpass and the elevation
the video camera survey. From this study, after installing the
difference etc.
signal pedestrians chose one-step crossing, as the
Lyons et.al. (2000) explained that signalised midblock
pedestrians got space to wait to travel across the street at
pedestrian crossing is a general method of resolving the
signal red of vehicular traffic.
conflicts between pedestrian and vehicular traffic. This
paper focuses on developing an alternatives for signalised
III. OBJECTIVES
midblock pedestrian crossing that are more responsive to the
needs of pedestrians without causing excess delay to the The objectives of the study include:
vehicular traffic. ANN models are developed to enhance the  To study macroscopic stream model for pedestrian
operation of signalised midblock pedestrian crossings. midblock crosswalks.
Xuehaochu et.al. (2002) developed a methodology for  To evaluate pedestrian Level of Service (PLOS) at
pedestrian LOS at midblock locations. The objective was to selected signalised midblocks.
determine what variables are correlated with pedestrians’
perceived quality of service for midblock crossings. Results IV. METHODOLOGY
showed that the levels of crossing difficulty tend to increase
with the width of painted medians, signal spacing, and The methodology adopted for the study is shown below in
turning movements. The results also indicated that presence figure 1:
of pedestrian signals lowers the crossing difficulty.
Noyce et.al. (2009) used a micro simulation approach for
determining the common alternatives for signalising a
typical midblock crosswalk with diverged geometrics, aimed
to explore how changing geometrics and signalisation
schemes affects various measures of effectiveness from both
pedestrian and vehicle perspectives.
Kadali et.al. (2014) examined the gap acceptance of a
pedestrian using ANN model. Video graphic data was
captured with the help of three cameras placed at Worli in
Mumbai; one is used for knowing the pedestrian behaviour
and the other two for recording vehicular characteristics.
From the results it has been found that the pedestrian rolling
gap, frequency of attempt, vehicular gap size, pedestrian
speed change condition and vehicle speed play an important
role in pedestrian gap acceptance.
Lin et.al. (2014) developed an LOS for pedestrians at an
unsignalised midblock in China. Data was collected from
the pedestrian crossing by questionnaire survey. Pedestrians
perception were noted from two respects separately i.e., the
convenience and the safety. Regression and Pearson
correlation analysis were carried out to develop LOS model
and results revealed that Los model developed was reliable
with R2 as 0.8.
Vedagiri et.al (2014) modelled pedestrian midblock
crosswalk LOS under mix traffic conditions. The objective Fig. 1. Flow chart showing the methodology adopted in
of the study is to find out the factors affecting pedestrian the study.
LOS at mid-block locations based on pedestrian perceived
conditions. A complete 2154 pedestrians were asked to give V. STUDY LOCATION AND DATA
evaluations of mid-block after they cross the road in the city COLLECTION
of Mumbai. From the study it can be concluded that the
variables such as age, frequency of using crosswalk, purpose Three signalised midblock locations Kukatpally,
of the trip, safety condition, crossing difficulty and median Dilshuknagar and Nizampet are considered for the study as
width have a significant effect on pedestrian LOS at mid- shown in the figures below. A two hours pedestrian data is
block location under mixed traffic scenario. collected for two peak hours in a day using videography
Kadali et.al. (2015) explained pedestrian LOS with respect technique.
to land use such as shopping, residential and business areas.
NLOGIT software package and the ordered probit (OP)
model were used for the analysis. Results conclude that land

Published By:
Retrieval Number: E6160018520/2020©BEIESP Blue Eyes Intelligence Engineering
DOI:10.35940/ijrte.E6160.018520 2752 & Sciences Publication
 International Journal of Recent Technology and Engineering (IJRTE)
ISSN: 2277-3878, Volume-8 Issue-5, January 2020

The data required for the study (flow, speed and delay) are VI. RESULTS AND DISCUSSION
extracted manually from the video for every 5min interval.
A. Calibration of Greenshields macroscopic model
The other parameters max speed (Vmax), jam density (Kj)
and max flow (Qmax) is determined using Greenshield’s Macroscopic models are used to determine the relation
formulae and their respective graphs are plotted. between traffic parameters (flow, speed and density). The
most important relation among them is speed and density.
The simplest relation between them is first proposed by
Greenshield. A linear relationship is developed between
speed (v) and density (k) using the Equation (1):

(1)

The relation between flow (q) and density (k), shown in

equation (2):
(2)
Fig. 2. Kukatpally signalised midblock.
The relation between speed (v) and flow (q) is shown in
equation (3):
(3)

Using the pedestrian data, a graph is plotted as shown in

figure3, between speed and density from which the basic
equation is obtained as shown in equation (4). Here X is
considered as the pedestrian density and Y is crossing speed
of the pedestrian.

(4)
Fig. 3. Nizampet signalised midblock
2.5
Pedestrian Crossing Speed

y = -0.3966x + 1.4014
2.0

1.5
(m/s)

1.0

0.5

0.0
0.0 0.5 1.0 1.5 2.0
Pedestrian Density K (ped/m)

Fig. 4. Dilshuknagar signalised midblock Fig. 4. Pedestrian Crossing Speed Vs Pedestrian Density.

The green time provided for the pedestrians at two locations According to calibration of Greenshield’s model, jam
is 30 sec and the green time provided for the vehicles is 125 density is obtained as 3.53ped/m and free flow speed as
sec. The flow recorded at the locations is as follows: 1.4m/s. The respective Greenshield’s graphs are plotted as
shown below in the figure (5):
 Kukatpally morning (9AM to 11AM) – 3671
pedestrians/hour
 Kukatpally evening (4PM to 6PM) - 4868
pedestrians/hour
 Nizampet morning (9AM to 11AM) - 2263
pedestrians/hour
 Nizampet evening (4PM to 6PM) - 2768
pedestrians/hour
 Dilshuknagar morning (9AM to 11AM) - 3183
pedestrians/hour
 Dilshuknagar evening ( 4PM to 6PM) - 3104
pedestrians/hour

Published By:
Retrieval Number: E6160018520/2020©BEIESP Blue Eyes Intelligence Engineering
DOI:10.35940/ijrte.E6160.018520 2753 & Sciences Publication
Determination of Pedestrian Level of Service at Signalized Midblock Locations for Mixed Traffic Conditions

Three popular methods are used to determine the optimal

1.5 number of clusters. They are: 1. Elbow Method
Crossing Speed(v) in

2. Silhouette Method
3. Gap Statistic
1.0 In this study Elbow method is used to find the number of
m/s

clusters using R software. It is a method of interpretation

and validation of total within clusters designed to find the
0.5 appropriate number of clusters in a dataset.

0.0
0 1 2 3 4
Density(k) in ped/m
Flow(q) in ped/s

0.5 Fig. 6. Validation Measures for Optimal Number of

Clusters using Elbow Method.
From the figure 6, it is clear that the number of clusters
0 obtained is six (k=6). Once the optimal k is resolved, the
0 1 2 3 4 program is re-run with k equivalents to 6 to assess the
Density(k) in ped/m clusters. Average crossing speed and volume to capacity
ratio (v/c ratio) were used as an input to the clustering
algorithm in R studio. Using the free flow speeds data the
validation parameters are classified into some number of
1.5
classes to classify the values for signalized midblocks. The
Crossing Speed(v) in m/s

validation parameters for the obtained values are shown in

below Figure 5.10.
1.0

0.5

0.0
0 0.5 1 1.5
Flow(q) in ped/s

Fig. 5. Greenshield’s graphs showing relation between

basic parameters for the collected pedestrian data.
sAccording to IRC 103 guidelines max free flow speed is
1.2m/sec. From the study it is revealed that free flow speed Fig. 7. Pedestrian level of service of signalized midblocks
for the midblocks is 1.4 m/sec which is slightly higher than using elbow method.
IRC guidelines. Clusters formed are used to determine the total time taken
B. Development of LOS scores: by the pedestrian to cross the midblock (waiting time or
delay + crossing time) for each LOS. The time ranges are
To develop Level of Service score clustering technique is
determined from the six clusters formed as shown in table:
adopted using K-mean clustering analysis. Clustering is a
technique where the dataset is divided into broad sets for
finding subgroups in one set of observation. K-means
clustering is the simplest form of clustering technique where
the observations are divided into k number of clusters. In
order to know how many number of clusters an observation
can be divided optimal number of clusters is determined.

Published By:
Retrieval Number: E6160018520/2020©BEIESP Blue Eyes Intelligence Engineering
DOI:10.35940/ijrte.E6160.018520 2754 & Sciences Publication
 International Journal of Recent Technology and Engineering (IJRTE)
ISSN: 2277-3878, Volume-8 Issue-5, January 2020

Autometa can be applied for microscopic analysis of the

data. Further analysis can be considered using pedestrian
Table-I: Proposed Time Ranges for Different PLOS
age, gender, with or without baggage, purpose of the
Proposed in Indian Conditions for Signalized Midblocks
walking trip etc. Further research can be extended to
PLOS Total time taken by the pedestrian develop criteria for PLOS at signalized midblocks for
different types of roads.
to cross midblock (s)
A <52 REFERENCES
1. Asaithambi, G., Kuttan, M.O. and Chandra, S., 2016. Pedestrian road
B 52-64 crossing behavior under mixed traffic conditions: A comparative
study of an intersection before and after implementing control
C 64-82 measures. Transportation in Developing Economies, 2(2), p.14.
D 82-90 2. Axler, E.A., 1984. Warrants for pedestrian over and
underpasses (No. FHWA-RD-84-082). United States. Federal
E 90-98 Highway Administration.
3. Baltes, M.R. and Chu, X., 2002. Pedestrian level of service for
F >98 midblock street crossings. Transportation research record, 1818(1),
pp.125-133.
4. Kadali, B.R. and Vedagiri, P., 2014. Modeling Pedestrian Crosswalk
Level of Service (LOS) Under Mixed Traffic Condition at Mid-Block
Based on the time ranges LOS is categorised from A to F. In Locations (No. 14-2641).
LOS-A sufficient area is provided for pedestrians to select 5. Kadali, B.R. and Vedagiri, P., 2015. Evaluation of pedestrian
freely their own walking speed and there will be no crosswalk level of service (LOS) in perspective of type of land-
use. Transportation research part A: policy and practice, 73, pp.113-
congestion. In LOS-F pedestrians have extremely restricted 124.
walking speed indicating difficulties such as pedestrian 6. Kadali, B.R. and Vedagiri, P., 2016. Review of pedestrian level of
vehicle interaction, improper signal timing, and maximum service: Perspective in developing countries. Transportation Research
pedestrian volume. Record, 2581(1), pp.37-47.
7. Kadali, B.R., Rathi, N. and Perumal, V., 2014. Evaluation of
Table - II: Obtained PLOS for the Selected Midblock pedestrian mid-block road crossing behaviour using artificial neural
Locations network. Journal of traffic and transportation engineering (English
edition), 1(2), pp.111-119.
Average 8. Lu, G. and Noyce, D.A., 2009. Pedestrian crosswalks at midblock
locations: Fuzzy logic solution to existing signal
time taken Average PLOS operations. Transportation research record, 2140(1), pp.63-78.
PLOS 9. Zhao, L., Bian, Y., Lu, J. and Rong, J., 2014. Method to determine
by the Delay for according pedestrian level of service for the overall unsignalized midblock
Midblock based on
pedestrians the to Indo crossings of road segments. Advances in Mechanical Engineering, 6,
Location proposed p.652986.
to cross pedestrians HCM
scores
the road (s) 2017
(s)
Kukatpally 89.04 D 71.4 E
Nizam-
87.82 D 68.73 E
pet
Dilshuk
52.44 B 38.45 D
nagar

This study provides the measure of midblocks performance.

Utilizing PLOS value at midblocks their plan can be
comprehended based on the pattern of pedestrian travel
which helps in characterizing its condition and furthermore
understands the necessities of pedestrians. This study helps
in improving pedestrian facilities ensuring safety and
comfort. As per the study, LOS for the Kukatpally and
Nizampet signalised midblock is D and for Dilshuknagar it
is obtained as B which indicates the occurrence of minor
pedestrian conflicts.

VII. CONCLUSION AND FUTURE SCOPE

LOS scores are developed using K-mean clustering analysis
for the obtained speeds and are compared with the Indo
HCM 2017 methodology. As the present PLOS the LOS
scores mostly lie in between B and D, as the demand
increases facilities has to be improved to cater to the future
pedestrian volumes. Different techniques like Cellular

Published By:
Retrieval Number: E6160018520/2020©BEIESP Blue Eyes Intelligence Engineering
DOI:10.35940/ijrte.E6160.018520 2755 & Sciences Publication