Proceedings of IOE Graduate Conference, 2017

Volume: 5 ISSN: 2350-8914 (Online), 2350-8906 (Print)

Performance Improvement of a Signalized Intersection (A Case

Study of New Baneshwor Intersection)
Sakar Shrestha a , Anil Marsani b
a, b
Transportation Engineering Program, Department of Civil Engineering, Pulchowk Campus, IOE, TU, Nepal
Corresponding Email: a [email protected]

Abstract
New-Baneshwor is one of the major junctions and also known as the heart of the Kathmandu Valley. This
place consists of all the major educational institution, hospitals, a wide range of market, and most important
“constitutional assembly” of the country. So it is one of the busiest places of Kathmandu Valley with high density
traffic. The main cause of the problems that occurred in the area of New-Baneshwor is due to heavy traffic volumes
which exceed the capacity of the intersection. In view of the above, it was necessary to analyze the existing
conditions and proffers some alternative solutions. Before carrying out an analysis, a field surveys at peak hours,
for example, in the morning (09:20 to 11:05 am) and evening (05:20 to 06:50 pm) was conducted. The number of
vehicles was counted manually with “video graphic survey” according to types of vehicles. From the analysis study,
it was found that the Level of Service (LOS) was F during morning as well as evening peak hours. This means that
the capacity of the existing road is no longer able to accommodate the traffic flow. One of the alternative solutions
for the congestion that occurs at the intersection of New-Baneshwor is to apply the efficiency of the intersection
that is not provided with the Four phases signal planning, Three phases signal planning by providing U-Turn, Fly
over with existing scenario, Four phases signal planning with flyover, Three phases signal planning with flyover
by providing U-Turn verifying queue pocket area for U-Turn. Based on the travel time and delay reduction with 5
comparative modeling it shows that the three phases signal planning with flyover by providing U-Turn effectively
decreases delay and travel time by 81.92% and 80.1% in morning and evening peak time respectively maintaining
Level of Service C, in addition, Maitighar, Tinkune, Old Baneshwor as well as Sankhamul lane is found to have
decreased by minimum 60% in the morning and evening which was found to be the most technically efficient to be
applied.
Keywords
Intersection, VISSIM, Level of Service

1. Introduction either meeting or crossing at same or different level and

is the main reason for the congestion. An intersection
Traffic congestion is a condition on transport networks has always played its role by controlling the conflict and
that occurs as use increases, and is characterized by merging stream of traffic so that delay that happened
slower speeds, longer travel times, and increased can be reduced or minimized. Normally geometric
vehicular queuing. The most common example is the parameter is the one that control and regulate the
physical use of roads by vehicles. When traffic demand vehicle path through intersection [1]. Not only that
is great enough than the interaction between vehicles assessment process can be done as to determine the
slows the speed of the traffic stream, this results in some traffic flow and turning movement at the selected
congestion. Most of the transportation engineers believe intersection. This can lead to determining whether the
that congestion will stay and we have to live with it. traffic flow exceeds the capacity of the intersection can
Limited financial resources and restricted right of way sustain or not. If there is no overcapacity experienced
makes it truer for developing countries. then the intersection cab be considered operate well
An Intersection is the junction of two or more roads below its capacity and smooth traffic going on. On the

Pages: 389 – 396

Performance Improvement of a Signalized Intersection (A Case Study of New Baneshwor Intersection)

other hand if there is overcapacity that happened then delays were reduced by 30.41% over the same period;
there should be inclusion of travel time and delay for the saving in travel time and vehicle operating costs
the assessment of the intersection. Delay that happens at amounted to 421.65 Million Baht [5].
intersection has been a major problem in the analysis of
It was found out that the degree of saturation (DS) was
congestion [2]. The intersection delay study is valuable
1.61 between Teuku Umar and Setia Budi road during
in evaluating the efficiency of effectiveness of a traffic
the morning peak hours and 1.56 during the afternoon
control method.
peak hours which means that the capacity of the
VISSIM is a microscopic, discrete traffic simulation existing road is no longer able to accommodate the
system modeling motorway traffic as well as urban traffic traffic flow and compared alternative solutions for the
operations. Based on several mathematical models, the congestion that occurs at the intersection of Jatingaleh
position of each vehicle is recalculated every 0.1–1 s. by applying a Fly over, Underpass and the combination
The system can be used to investigate private and public of Fly Over-Underpass. Based on the flow reduction
transport. In order to simulate multi-modal traffic flows, calculation with 3 comparative modeling it show that
technical features of pedestrians, bicyclists, motorcycles, the Fly Over is the most technically efficient to be
cars, trucks, buses, trams, light (LRT), and heavy rail are applied in his research following the flow of traffic can
provided with options of customization [3]. be reduced by 93.42% (morning)and 93.59%(afternoon)
while after the underpass in the form only able to reduce
25.24% (morning)and 14.28%(afternoon) [6].
2. Literature Review Studies of flyover for improvement of intersection were
done suggesting in order to reduce traffic congestion
Study on ”Estimation of delay at signalized at-grade intersection near a big city, one method is
intersections for mixed traffic conditions of a construction of flyover bridge at the old junction in two
developing country” was conducted from which various directions on one of the main highways. The flyover
problems associated with delay estimation under mixed facilitates the traffic flow in the directions of the bridge,
traffic conditions in a developing country (India) and the but the infrastructure cannot fully solve all of the
methods to overcome them were discussed and an problems especially on the secondary road. Under the
attempt was made to improve the accuracy estimating bridge, although it relieves the traffic congestion at the
the same. Five isolated signalized intersections from a intersection; the traffic signal still uses the same control
fast developing industrial city located in Tamil Nadu, as the “before” situation, that is the fixed time control
India were chosen for the study. Site specific PCU 13 plan. With the flyover bridge in place, it was found
values were developed considering the static and that about 30-35% of all traffic volumes diverted to the
dynamic characteristics of vehicles. Saturation flow was bridges, and time delay reduced by 30% over the same
also directly measured in the field for the prevailing period. This paper provides suggestions to increase the
roadway, traffic and signalized conditions and expressed benefits of the flyover such as creating a new cycle and
in PCU/h. Control delay was also measured following phase times and improving the physical area under the
HCM (2000) guidelines. Later, this was compared with bridge [7].
that estimated from the theoretical delay model. Even
after taking several measures, good correlation between Evaluation of indirect right turn treatment to reduce
observed and predicted delay could not be obtained. conflicts and congestion at signalized intersections in
Therefore, in the present scenario field measured control urban areas was conducted. Volume and travel time
delay was taken into account to define LOS [4]. studies were conducted at three signalized intersections
in Islamabad where indirect right turn treatments were
Study about the performance of the flyover upgraded applied. Travel time study was done using GPS device.
intersection was conducted and points out the remaining To monitor the traffic flow and driver behavior video
problems including long delays under the flyover and recordings were also done on each intersection during
found that about 35-40% of the total traffic volume the peak hours. Microscopic simulation model VISSIM
diverted to the flyover, and despite an increase in traffic and field travel time runs were used to evaluate before
volume of +29.46%, at the intersection, the vehicle

390
 Proceedings of IOE Graduate Conference, 2017

and after scenarios. It is found that at some locations, in urban areas evidencing exciting roads or the scope of
application of indirect right turn reduce travel time and improvement of intersection at-grade are very limited.
increase the vehicle output. However, at one So, due to the unavailability of land and to achieve fast
intersection the implementation of indirect right turn and safe movement of vehicle various alternatives must
treatment resulted in increased travel times and reduced be analyzed at New-Baneshwor intersection.
vehicle output. This is because of the number of
vehicles affected by the closure of intersection are 3.2 The Objective of the Study
higher and also the movement of vehicles at the U-turns
was effected by on-street parking [8]. The overall objective of this study is to make
comparison of probable congestion reduction
It was proposed right turn lane design with sufficient approaches at New Baneshwor intersection with the
lengths to improve the safety and operation of help of micro-simulation model via VISSIM. Specific
intersections by providing space for deceleration and objective:-
storage of right-turn vehicles. On the other side,
insufficient length of right-turn lane will result in the • to identify the current level of service at New-
right turn lane over flow and the blockage of right-turn Baneshwor intersection
lane entrance by through traffics, which were referred
toas right-turn overflow and blockage problems. [9] • to calibrate and validate micro-simulation model
to evaluate the travel time and delay of vehicles at
Study on comparative study of VISSIM and SIDRA on
existing scenario
signalized intersection was performed. Due to
increasing volume of traffic intersection jam happens • to establish probable alternatives to reduce the
constantly. An effective way to solve the problem is to congestion problem at New-Baneshwor
seek appropriate control mechanisms with the help of intersection
traffic simulation software. Based on the real world
traffic data takes from the West Wenhua Road and • to compare the travel time and the delay after
Changchun Road of Xian yang City, their features and application of different probable approaches
evaluation results of signalized intersections have been
analyzed from the perspective of operation simplicity 4. Methodology
and the output error. Results show that SIDRA
operation is easier while VISSIM output is more 4.1 Field Data Collection
accurate. This study can provide references for
simulation software selection[10]. Video cameras were used as a field data collection tool to
record the footage of traffic flow at the New-Baneshwor
intersection. Following data were collected in the field:
3. The Need and the Objective of the
Study • Intersection geometry, including lane usage and
link distances
3.1 The Need of the Study
• Existing Intersection Turning Movement Counts
According to the transportation professionals’ concern
the increase in urban traffic congestion has become a • AM and PM peak hour observation (2 hours @ 15
serious matter. In urban road networks, the intersections minute periods) for four days
are very large and very close to each other. Due to • Classified count of vehicles
which more traffic congestion is observed especially
during peak hours. Researchers are making efforts for • Field studies including travel time runs and
obtaining rational quantification of congestion and approach delay studies. These data were useful
formulating appropriate measures for mitigation of for the calibration of the computer model and for
congestion for urban roads, also efforts to improve the comparison to corresponding data in and after
traffic operations at urban intersection are made. In fact, condition

391
Performance Improvement of a Signalized Intersection (A Case Study of New Baneshwor Intersection)

• Currently functioning Signal Timing and Phasing

Data

• Additional data (as required

4.2 Data Extraction & Preparation of Input

Parameters for Microscopic Simulation
Model
The videos recorded at the site were played back on a
computer to extract the required data. From the videos,
following data were extracted: Figure 1: Sample of calibration of Old Baneshwor lane
for delay
• Classified vehicle count

• Directional movement of each vehicle R squared = 0.8495 (i.e. 84.95 percent of variance of
• Relative flow field data is explained by the variance of Vissim output

• Signal Timing and phasing data Regression equation: VISSIM OUTPUT EXPECTED=
0.8424 * FIELD + 29.683
• Size of vehicles for modelling in VISSIM

• Analytical Computation of Delay and Travel time 5.2 Validation of model

The model was compared between Vissim output
4.3 Development of Model in VISSIM expected and Vissim output for Day4 collected at the
same site under similar conditions and was performed
Links were created using connector to represent road
for validation which is shown in figure 2.
segments that carry them through movements and
general curvature of the roadway. A connector is a type
of link used to join two areas of a single link or to join
two areas of two links which have additional
characteristics that affect driver behavior, specifically
lane changing, so it is important when coding to take
this into consideration and eliminate the excessive use
of connectors. They are the blue print for creating a new
traffic network. They were imported from aerial
photograph. Aerial image was acquired from a virtual
map tool such as Google earth.
Figure 2: Sample of validation of Old Baneshwor lane
for delay
5. Calibration & Validation of model

5.1 Calibration of model 6. Data Analysis and Results

Calibration was performed using driving behaviour
6.1 Traffic Volume
parameters in Vissim. The field data of Day 1, Day 2 &
Day 3 were used to Vissim as inputs. Simulation runs Traffic survey showed that during the morning peak
were performed for those inputs. Observations were hour, there are more traffic demand flow from Tinkune
made to see if the simulation matches the corresponding as compared to the Maitighar, Old Baneshwor and
field values of delay, travel time & traffic volume which Sankhamul i.e. from Tinkune to Maitighar at
is shown in figure1. New-Baneshwor Intersection. This happens due to the

392
 Proceedings of IOE Graduate Conference, 2017

arrival at the offices and business center of Kathmandu with any other lane and hence 20 cycles for Tinkune
valley which are located around the Ratnapark, lane, 20 cycles for Maitighar lane(38.5 % of total traffic
Pulchowk, Kupandole, etc. area. Total traffic volume at volume), 19 cycles for Sankhamul lane, 18 cycles for
New-Baneshwor intersection is 14,876 PCU from 9:20 Old Baneshwor lane is found out maintaining demand
A.M. to 11:05 A.M for DAY 4. Similarly for DAY 4, of each lane by traffic police when analyzed for 1 hour
total traffic volume at Tinkune is 7127 PCU which is and 30 minutes.
more than minimum twice than that of Maitighar (3205
PCU), Old Baneshwor (2461 PCU) and Sankhamul
(2084 PCU) which is shown in figure 3.
It was also observed that during the evening peak hour,
there are more traffic demand flow from Maitighar as
compared to the Tinkune, Old Baneshwor and
Sankhamul i.e. from Maitighar to Tinkune at
New-Baneshwor Intersection. This happens due to the
departure from the offices and business center of
Kathmandu valley. Total traffic volume at
New-Baneshwor intersection is 12,056 PCU from 5:20
P.M. to 6:50 P.M for DAY 4. Similarly for DAY 4, total
traffic volume at Maitighar is 4646 PCU which is more
than minimum 1.25 than that of Tinkune (3694 PCU),
Old Baneshwor (2136 PCU) and Sankhamul (1578
PCU) which is shown in figure 4.
Figure 3: Directional movement (Morning Peak)
Table 3 shows the total traffic volume of morning and
evening peak time in terms of PCU which was carried
out for 105 minutes and 90 minutes respectively.

6.2 Vehicle Composition

Traffic survey showed that motorcycle was found to be
one of the major modes of transportation.Typically at
New-Baneshwor intersection motorcycle/Bicycle was
found to be 77 %, car/Jeep/Van/Pick up/Light Van/Micro
Bus 18 %, Mini Bus/Truck/Bus 4 % and Tempo 1 % in
composition.

6.3 Traffic Police Signal Data

In the Existing scenario, during morning peak time it is
seen that traffic volume of Tinkune Lane is higher when
compared with any other lane and hence 23 cycles for
Tinkune lane(47.91 % of total traffic volume), 15 cycles Figure 4: Directional movement (Evening Peak)
for Maitighar lane, 16 cycles for Sankhamul lane, 15
cycles for Old Baneshwor lane is found out maintaining
demand of each lane by traffic police when analyzed for 6.4 Level of Service (LOS)
1 hour and 45 minutes.
Level of Service for both morning and evening peak are
Similarly, during evening peak time it is seen that traffic found to be F which were calculated using HIGHWAY
volume of Maitighar Lane is higher when compared CAPACITY MANUAL [11] for which passenger car

393
Performance Improvement of a Signalized Intersection (A Case Study of New Baneshwor Intersection)

unit for different vehicles were adopted from Nepal Road is checked for delay and traffic volume at U-turn section
Standard 2070 [12]. due to the probability of increment or decrement of delay
and traffic volume at U-turn section of existing scenario
6.5 Evaluation of results for different respectively. It is seen that after providing U-turnat
alternatives 300m, there is no any huge variation on delay as well
as traffic volume with respect to existing scenario which
6.5.1 Different Alternatives Solutions are listed Table.
Total 5 alternative solutions options were performed on Again from the above option 2, 3 & 5, it is observed that
Vissim which are liste the overall delay for Three phases signal planning with
flyover by providing U-Turn at 300m is effectively
Option 1: Four-Phases Signal Planning on Existing decreased by 81.92% and 80.1% in morning and
Scenario of fixed timing signal 80, 120, 150 & evening peak time respectively maintaining Level of
200 sec. [13] Service C, in addition, each lane of Maitighar,Tinkune,
Option 2: Three phases signal planning of 120 sec Old Baneshwor & Sankhamul are found to have delay
by providing U-Turn at 150, 200, 300 & 400m and travel time decreased by minimum 60% in the
verifying queue pocket area for U-Turn [13] morning and evening peak time. Thus, among all the
above options, option 5 is the most effective in the
Option 3: Fly over with existing scenario[6] [13] [14] reduction in congestion. The best signal planning for
Option 4: Four phases signal planning with flyover of option 5 that was adopted in Vissim is listed in Table 2
120 sec.[6] [13] [14]
Option 5: Three phases signal planning with flyover
by providing U-Turn at 150, 200, 300 & 400m Table 2: Adopted Signal Planning for Option 5
verifying queue pocket area for U-Turn [13] [14] Priority Time (sec.)
Maitighar & Tinkune 50
Comparsion chart of option 2 & 5 with existing scenario Old Baneshwor 35
is presented in Figure5, Figure6, Figure7 & Figure8. Sankhamul 35
From the above Option 1 & 2, it is observed that the
overall delay for Three phases signal planning of 120
sec by providing U-Turn at 300m is effectively
decreased by 69.93% and 78.9% in morning and
evening peak time respectively maintaining Level of
Service D, in addition, each lane of Maitighar,Tinkune, Table 3: Traffic Volume of morning and evening days
Old Baneshwor & Sankhamul are found to have delay interms of PCU
and travel time decreased by minimum 30% in the Peak Date Day Total
morning and evening peak time.The best signal Time
planning for option 2 that was adopted in Vissim is Morning 2073 -3-22 Wednesday 12483
listed in Table 1. Morning 2073 -5-12 Sunday 13741
Table 1: Adopted Signal Planning for Option 2 Morning 2073 -7-05 Friday 13988
Morning 2073 -7-09 Tuesday 14876
Priority Time (sec.)
Evening 2073 -9-08 Sunday 10923
Maitighar & Tinkune 70
Evening 2073 -9-10 Tuesday 10947
Old Baneshwor 15
Evening 2073 -9-11 Wednesday 11460
Sankhamul 35
Evening 2073 -9-13 Friday 12056

Similarly, after protecting the right turning traffic of

Maitighar and Tinkune by providing U-turn at 300 m it

394
 Proceedings of IOE Graduate Conference, 2017

6.5.2 Comparison chart

Figure 8: Comparsion of travel time between different

alternatives (Evening peak)

Figure 5: Comparsion of delay between different 6.5.3 Verifying Queue Storage Length
alternatives (Morning peak)
Maximum queue length for storage legth that must be
provided (L1) = 91.7m
Comfortable Deceleration Length (L2) = 55m (F0r
Design Speed 60km/hr.)
Minimum Total Length for U-Turn that ust be provided
= L1 + L2 = 91.7 + 55 = 146.7m less than 300m (O.K.)

7. Conclusion
It was found that the Level of Service for both morning
and evening peak are found to be F.
Four phases signal planning effectively decreases delay
and travel time by 2.0254% and 9.76% in morning and
evening peak time respectively, however, Maitighar
Figure 6: Comparsion of delay between different main is found to increase by minimum 60% in the
alternatives (Evening peak) evening peak time and Sankhamul during morning and
evening peak and are observed to be non-effective in the
reduction in congestion.
After applying three phases signal planning by providing
U-Turn at 300m, it was observed that the overall delay
and travel time was effectively decreased by 69.93% and
78.9% in morning and evening peak time respectively
maintaining Level of Service D, in addition, each lane is
found to decreased by minimum 30% in the morning and
evening peak and later checked for the traffic volume
and delay at U-turn and found that there is no any huge
variation on delay as well as traffic volume with respect
to existing scenario.
Four phases signal planning with flyover observed that
the overall delay and travel time was effectively
Figure 7: Comparsion of travel time between different decreased by 59.37% and 69.06% in morning and
alternatives (Morning peak) evening peak time respectively maintaining Level of

395
Performance Improvement of a Signalized Intersection (A Case Study of New Baneshwor Intersection)

Service D, in addition, each lane is found to have compare the results with each output.
decreased by minimum 20% in the morning and
evening peak.
Acknowledgments
Finally after applying three phases signal planning with
flyover by providing U-Turn at 300m, it was observed The authors are grateful to every single person who
that the overall delay and travel time was effectively contributed directly or indirectly in the research.
decreased by 81.92% and 80.1% in morning and References
evening peak time respectively maintaining Level of
Service C, in addition, each lane was found to have [1] Bernard Enright Martin Rogers. Highway Engineering.
decreased by minimum 60% in the morning and WILEY Blackwell, 3rd edition, 2016.
evening peak time. Thus three phases signal planning [2] Wolfgang S. Homburger and James H.Kell.
Fundamentals of traffic engineering. Berkeley,
with flyover by providing U-Turn at 300m was most Calif. Institute of Transportation Studies, University of
effective in the reduction in congestion. California, 9th edition, 1977.
[3] Jaume Barceló. Fundamentals of traffic engineering.
Springer-Verlag New York, 1st edition, 2010.
8. Recommendation
[4] G. Dhinakaran R. Prasanna Kumar1. Estimation
of delay at signalized intersections for mixed traffic
Based on the analysis, congestion at New-Baneshwor conditions of a developing country. 2012.
intersections, it is an extremely difficult task to cover [5] Auttakorn SALA. Assessment of traffic flow benefits
all aspects of a topic in a limited time frame. Following of flyovers:a case study. 2013.
tasks are recommended for future studies in a similar [6] Rahmat Mudiono. Research of flyover as a solution to
topic: congestion of intersection junction a case study: Jalan
jatingaleh semarang. 2015.
[7] Narabodee Salatoom1 and Pichai Taneerananon. A
• The vehicular volumes considered in this study study of a flyover-bridge - improved intersection. 2015.
were based on the present data collected on field.
[8] Kamran Ahmed. Evaluation of low cost technique
It is suggested to accurately forecast the future “indirect right turn” to reduce congestion at urbanized
traffic volumes and attempt three phases signal signalized intersection in developing countries. 2011.
planning with flyover by providing U-Turn, [9] Yi Qi Lei Yu. Left turn lane design and operation. 2007.
underpass too for the stability over time in year. [10] YANG Hongxuc CHEN Tianzi, JIN Shaochenb.
Comparative study of vissim and sidra on signalized
• While conducting the traffic signal of intersection. 2013.
New-Baneshwor intersection on VISSIM that was [11] USA: Transportation Research Board (TRB). Highway
controlled by traffic police manually, it seems that Capacity Manual 2000. National Research Council,
the fixed timing signal plan on VISSIM were 2000.
limited to 125 signal combination which [12] Nepal Road Standard 2070. Department of Road.
restricted the analysis period up to 1 hr. and 45 [13] Suresh Nama Avijit Maji, Akhilesh Kumar Maurya.
minutes for which 15 minutes were taken for Performance-based intersection layout under a flyover
for heterogeneous traffic. 2015.
seeding period and results total up to 1.5 hrs.
analysis period. So it can be conducted by using [14] Joni Arliansyah Rhaptyalyani H. Della, Hanafiah.
Traffic performance analysis of u-turn and fly over
other traffic flow simulation software’s such as u-turn scenario;a case study at soekarno hatta road,
SimTraffic, Sidra Intersection, CORSIM ,etc. and palembang, indonesia. 2015.

396