Engineering Perspective 3(3):27-34, 2023

e-ISSN:2757-9077

[Link] Research Paper

Capacity Analysis Based on Vehicle Trajectory Data on a Weaving Bottleneck in

Nanjing

Jean Claude Sugira1* , Jean Pierre Nsengimana2 , Marc Nshimiyimana2

1
School of Transportation Engineering of Southeast University, Nanjing, China
2
School of Civil Engineering of Southeast University, Nanjing China

ABSTRACT
This paper investigates the capacity of weaving bottlenecks in Nanjing, where multiple traffic streams merge or diverge closely,
causing significant traffic disruptions. Utilizing trajectory data from 862 vehicles collected through UAV cameras, the study assesses
the bottleneck's capacity and performance measures, including traffic flow, speed, and lane occupancy. The analysis employs a com-
bination of the Highway Capacity Manual (HCM) 2010 guidelines and PTV Vissim simulation software. The HCM 2010 framework
provides a standardized approach for evaluating traffic flow characteristics and determining the level of service, while Vissim allows
for the evaluation of diverse traffic scenarios.

The results reveal that the bottleneck operates at Level of Service E (LOS E), characterized by high congestion, reduced speeds,
and prolonged travel times. However, introducing ramp metering to the area improves conditions, resulting in a transition to Level of
Service D (LOS D). This study emphasizes the potential of ramp metering to enhance bottleneck performance. Nevertheless, future
research should explore other strategies, such as intelligent transportation systems and improved public transportation services, to
encourage modal shifts and reduce private vehicle usage. By addressing these challenges, cities like Nanjing can mitigate traffic con-
gestion and create more efficient and sustainable transportation systems. The findings provide valuable insights for urban planners
and policymakers seeking evidence-based solutions to optimize traffic flow and mobility in busy cities. By adopting the HCM 2010
guidelines and simulation tools like Vissim, this study contributes to the development of efficient traffic management strategies, ulti-
mately fostering more livable and accessible urban environments.

Keywords: Traffic congestion; bottleneck capacity; weaving bottleneck; HCM 2010; PTV Vissim; LOS; Ramp metering
History Author Contacts
Received: 03.07.2023 *Corresponding Author
Accepted: 05.09.2023 e-mail addresses: sugiraclaude3@[Link], pkinseb08@[Link], marc1shyaka@[Link]
How to cite this paper: Sugira, J.C., Nsengimana, J.P., Nshimiyimana, M., (2023), Capacity Analysis Based on Vehicle Trajectory
Data on a Weaving Bottleneck in Nanjing. Engineering Perspective, 3(3), 18-27. [Link]
/[Link].71385

1. Introduction capacity and proposing mitigation strategies using various

Urban areas worldwide are grappling with the persistent methodologies. The Highway Capacity Manual (HCM) 2010,
challenge of traffic congestion, which has far-reaching developed by the Transportation Research Board (TRB), serves as
implications for travel efficiency, environmental sustainability, a fundamental reference for evaluating the operational
and quality of life. Effective management of congested performance and level of service (LOS) of transportation
bottlenecks is crucial to mitigate the negative impacts of facilities.(Mahdi et al., 2022) The HCM 2010 guidelines provide
congestion and optimize the performance of transportation standardized frameworks for assessing key parameters such as
networks. The capacity analysis of bottleneck sections plays a travel time, speed, lane changes, and flow rates to determine
vital role in understanding traffic flow dynamics and identifying bottleneck capacity and LOS(Ryus, Vandehey, Elefteriadou,
appropriate strategies for congestion reduction(Skabardonis et al., Dowling, et al., 2011).In parallel, the utilization of advanced
2016). Numerous studies have focused on analyzing bottleneck simulation tools has revolutionized capacity analysis by allowing

27
J.C. Sugira et al. Engineering Perspective 3(3): 27-34, 2023
researchers to create virtual models of road networks and assess future research projects to address several challenges. These
traffic flow under diverse scenarios(Raju et al., 2018). PTV include the development of more accurate and comprehensive
Vissim, a widely adopted microscopic traffic simulation software, models, the incorporation of real-time data for dynamic capacity
offers a powerful platform for analyzing traffic dynamics, analysis, the integration of intelligent transportation systems for
simulating different traffic conditions, and evaluating improved traffic management, and the evaluation of sustainable
performance measures within bottleneck sections(Suthanaya & transportation solutions to reduce congestion and environmental
Upadiana, 2019; WSDOT, 2014). This study aims to conduct a impacts. In light of the outlined research trends, strengths,
comprehensive capacity analysis of a weaving bottleneck in weaknesses, knowledge gaps, and the need for further research,
Nanjing, employing a combined methodology that incorporates this study aims to contribute to the existing body of knowledge by
the HCM 2010 guidelines and PTV Vissim simulation software. conducting a capacity analysis of this bottleneck in Nanjing,
The primary objectives are to characterize traffic flow patterns, employing a combined methodology of the HCM 2010 guidelines
assess the bottleneck's LOS, and propose effective congestion and PTV Vissim simulation software. The findings from this
reduction strategies to enhance traffic management. study will shed light on the specific characteristics of the weaving
The remainder of this paper is structured as follows: Section 2 bottleneck and provide insights for effective congestion reduction
presents materials and methodology, Section 3 provides results strategies and improved traffic management in similar contexts.
and discussion, outlining the application of the HCM 2010
guidelines and the use of PTV Vissim for simulation and presents 2. Data collection
the results of the capacity analysis, discussing the findings
Secondary data was collected online from the website seutraf-
specific to the weaving bottleneck in Nanjing. Finally, Section 4
[Link], where it was recorded using an Av high-resolution cam-
offers concluding remarks and suggestions for future research and
era. The data was captured within a duration of 450 seconds, en-
implementation. Research on bottleneck capacity analysis has
compassing trajectory data for 862 vehicles traveling unidirec-
witnessed significant advancements in recent years. Many studies
tional on a 5-lane highway, with the study area extending over a
have utilized the HCM 2010 guidelines as a foundation for
length of 386 meters.
assessing and understanding the operational performance and
LOS of bottlenecks.(Rouphail & Williams, n.d.) These studies
have successfully applied the HCM 2010 methodologies to
evaluate key parameters such as flow rates, speed, and lane
changes, enabling a standardized assessment of bottleneck
capacity. However, while the HCM 2010 provides a robust
framework, some limitations exist in terms of its applicability to
specific types of bottlenecks and the representation of certain
traffic conditions(Raju et al., 2018). Moreover, the use of
simulation tools, such as PTV Vissim, Anylogic, Ainsum, and
Sumo, has significantly contributed to the analysis of bottleneck
capacity and the exploration of congestion reduction
strategies(Wang et al., 2014). Simulation-based studies have
enabled researchers to create virtual models of road networks, Figure 1. Bottleneck section under study ([Link]
evaluate various traffic scenarios, and estimate performance In this traffic configuration, there are three continuous lanes and
measures with a high degree of accuracy. These studies have two merging ramps that eventually converge into a downstream
highlighted the advantages of simulation-based approaches, segment with three lanes as shown above in Figure 1. The col-
including the ability to capture real-world complexities, lected dataset encompasses various attributes for each vehicle, in-
investigate dynamic traffic patterns, and assess the impact of cluding acceleration, vehicle ID, velocity, latitude, longitude, ve-
different operational strategies. However, it is important to hicle length, vehicle width, and lane ID, all observed at different
consider the inherent limitations of simulations, such as the time intervals. Accompanying this dataset is a video recording
accuracy of input data and calibration of model parameters, to capturing the progression of each vehicle, from its entry into the
ensure reliable results(Bottleneck-Removal-1-Pg, n.d.). Despite bottleneck area to its exit. The analysis of this data reveals a sig-
the progress made in bottleneck capacity analysis, there are still nificant frequency of lane changes within the merging section.
knowledge gaps that warrant further investigation. For example, Many drivers opt to reduce their speed or decelerate upon ap-
there is a need for more research on the capacity analysis of proaching the merging influence zone. Once they navigate past
specific types of bottlenecks, such as complex urban intersections. the merging zone, they resume acceleration and regain the free
Additionally, the interaction between bottleneck capacity and flow speed of 110 km/h. Notably, during the merging area, the
emerging technologies, such as connected and autonomous free flow speed is determined to be 45 km/h, a finding drawn from
vehicles, requires further exploration to understand their influence the comprehensive data collection. Graphical representation of
on traffic flow and capacity(Zhao et al., 2017). collected vehicle speeds is seen on Figure 2.
The existing research in bottleneck capacity analysis and
congestion mitigation strategies has provided valuable insights
and methodologies. However, there remains a need for current and

28
J.C. Sugira et al. Engineering Perspective 3(3): 27-34, 2023
ulations were integrated to provide a comprehensive understand-
ing of the weaving bottleneck in Nanjing. The combined method-
ology allowed for a multi-dimensional analysis of the bottleneck,
considering both the observed real-world data and the simulated
scenarios. This integrated approach provided valuable insights
into the traffic flow dynamics, capacity constraints, and potential
congestion reduction strategies for the weaving bottleneck in Nan-
jing.

4. Results and discussion

4.1. Using HCM (2010)
The Highway Capacity Manual (HCM) 2010 is a widely used
resource for analyzing and evaluating the operational performance
of highways and transportation systems(Ryus, Vandehey,
Elefteriadou, Dowling G, et al., 2011). It provides guidelines and
methodologies for conducting various types of transportation
Figure 2. Vehicle ID vs Speed analysis, including data analysis. The data collected were ana-
3. Materials and Methodology lyzed by doing calculations based on the guidelines of HCM
(2010) where different flowrates, speeds, and other parameters
This study employed a mixed-methods approach to conduct a
have been calculated as detailed below:
comprehensive capacity analysis of bottlenecks. The methodol-
ogy consisted of two main components: the application of the
Calculation of fHV = heavy vehicle adjustment factor;
Highway Capacity Manual (HCM) 2010 guidelines for capacity
analysis and the utilization of PTV Vissim simulation software for 𝟏
fHV = (𝟏)+𝑷𝑻 (𝑬𝑻 −𝟏)+𝑷𝑹 (𝑬𝑹 −𝟏)
(1)
traffic simulation. To assess the capacity of the weaving bottle-
neck, the HCM 2010 guidelines were utilized as a standardized
where:
framework. The guidelines provided a systematic approach for
ET and ER are passenger car equivalents for trucks buses and rec-
evaluating key parameters such as flow rates, speed, and lane
reational vehicles (RVS), respectively
changes. Data collection for the capacity analysis involved the ex-
PT is the Proportion of trucks and buses, and RVS, respectively in
traction of recorded vehicle trajectory data at the bottleneck loca-
the traffic stream
tion. The collected data included vehicle speed, vehicle width, ve-
The terrain is in a class of level terrain hence, ET (Truck and buses)
hicle length, headways, lane positions, lane changes, and others.
is 1.5, and ER for (RVS) is 1.2. The proportion of Trucks and buses
Using the HCM 2010 methodology, the extracted trajectory data
in collected data is 0.006 and 0.0015 respectively.
were analyzed to calculate various performance measures such as
the level of service (LOS), speeds, and flow rates. These measures 𝟏
Then fHV = (𝟏)+𝟎.𝟎𝟎𝟔(𝟏.𝟓−𝟏)+𝟎.𝟎𝟎𝟏𝟓(𝟏.𝟐−𝟏)
= 0.99
provided insights into the operational performance of the bottle-
neck and facilitated the characterization of traffic flow patterns
Based on existing data, we have 6965 Veh/hr. where the free-flow
and congestion levels. Additionally, the HCM 2010 guidelines al-
speed in the downstream segment is 110 km/h and the free-flow
lowed for the identification of potential capacity constraints and
speed at the merging influence area is 45km/h.
bottlenecks within the weaving section. In parallel to the capacity
analysis, traffic simulation using PTV Vissim software was con-
Conversion of volume (Veh/h) to flowrate pc/h;
ducted to further investigate the traffic dynamics and assess the
performance of different congestion reduction strategies. PTV Using this equation: v=
𝑽
(2)
Vissim provided a powerful platform for creating a virtual model (𝐏𝐇𝐅)(𝒇𝑯𝑽 )(𝒇𝒑 )

of the road network and simulating various traffic scenarios. The

Where V: is the hourly volume (veh/h), V is Peak 15-min flow
simulation model was calibrated and validated using the collected
rate in an hour (pc/h), fHV : Heavy vehicle adjustment factor, fp is:
trajectory data to ensure its accuracy in representing real-world
Driver population factor.
traffic conditions.
Once the simulation model was established, different opera- 𝑯𝒐𝒖𝒓𝒍𝒚 𝑽𝒐𝒍𝒖𝒎𝒆
Peak Hour Factor (PHF) = =1
tional strategies were evaluated to identify effective congestion 𝐏𝐞𝐚𝐤 𝐟𝐥𝐨𝐰 𝐫𝐚𝐭𝐞 (𝐰𝐢𝐭𝐡𝐢𝐧 𝐡𝐨𝐮𝐫)
reduction measures. These strategies included ramp metering and
Checking for maximum flow entering merge influence area;
others. Through the simulation experiments, the impacts of each
strategy on traffic flow, capacity, and LOS were assessed, ena- 4776 2120
VF = = 4824 pc/h, VR = = 2141 pc/h,
bling the identification of the most promising approaches for con- (1)(0.99)(1) (1)(0.99)(1)
gestion mitigation. The capacity analysis results obtained from the
HCM 2010 guidelines and the findings from the PTV Vissim sim- V23 = V12x1.12 = 4863 x 1.12=3330 pc/h

29
J.C. Sugira et al. Engineering Perspective 3(3): 27-34, 2023
all other cases, including cases in which VR12 exceeds its stated
2944
V12 = =2974pc/h, , limit, LOS is determined by estimating the density in the ramp
(1)(0.99)(1)
influence area.
VR23 = V23+VR=3330+2141=5471pc/h and
𝑽𝑭𝑶 = 𝑉𝑅 + 𝑉𝐹 =4824+2141= 6965 pc/h, 4.1.1. Calculating level of service (LOS)
Where: 𝑉12 : Flowrate entering ramp influence area (pc/h) Computing density;
VR12 = sum of flow rates for ramp (VR) and vehicles entering ramp
This equation is used to estimate the density in the merge influ-
influence area (pc/h);
ence area
VR = flow rate on-ramp (pc/h)
𝐷𝑅 = 3.402 + 0.00456𝑉𝑅 + 0.0048𝑉12 − 0.01278𝐿𝐴 (3)
For this analysis, we have two conditions; LAeff = 2LA1 + LA2
Checking for 1st condition:
where: 𝐷𝑅 : Density of merge influence area (pc/km/ln), 𝑉𝑅 : On-
1st condition states that the total departing freeway flow (v) may
ramp peak 15-min flowrate (pc/h), 𝑉12 : Flowrate entering ramp
exceed the capacity of the downstream freeway segment. Failure
influence area (pc/h), 𝐿𝐴 : Length of acceleration lane (m)
(LOS F) is expected. Here no further calculations are needed, and
queues will form upstream from the merged segment. When the DR=3.402+0.00456*2141+0.0048*2974-0.01278*240
downstream freeway capacity is exceeded, LOS F exists regard- =24.37pc/km/ln. Basing on exhibit 25-7 in HCM(2000) and also
less of whether the flow rate entering the ramp influence area ex-
ceeds its capacity. For the given road section, the free flow speed as it shown in table 2, we were able to determine the Level of
is 110 km/h at the downstream street with 3 lanes in the same di- Service of the section under study.
rection. Hence the estimated capacity (𝑉𝐹𝑂 ) of the downstream
segment is 6965 pc/h which is less than 7050 pc/h which is sug- Table 2. LOS criteria for merging and diverging (Ryus, Vandehey,
gested by HCM LOS F will not exist. Elefteriadou, Dowling, et al., 2011)
LOS Density (pc/km/ln)
Table 1. Capacity values for merge area (Highway Capacity Manual
A ≤6
2000 RAMPS AND RAMP JUNCTIONS CONTENTS, n.d.) B >6-12
Freeway Maximum Downstream Freeway Flo V R1 2 C >12-17
Free-Flo w, v(pc/h) (pc/ D >17-22
w Number of Lanes in One Direction h/ln) E >22
Speed(km 2 3 4 >4 F Demand exceeds capacity
/h)
120 4800 7200 9600 2400 4600 The merging area is under LOS E as the density of the merging
110 4700 7050 9400 2350 4600 influence area is greater than 22pc/km/ln (see Table 2).
100 4600 6900 9200 2300 4600
90 4500 6750 9000 2250 4600 4.1.2. Characteristics of LOS E;
LOS E is indicative of moderately congested traffic flow. It im-
plies that the roadway is operating at or near its capacity, and ve-
VR12 is the Maximum Downstream Freeway Flow. Checking for
hicles are traveling at speeds slightly below the posted speed limit.
maximum flow entering merge influence area;
Vehicles in LOS E experience reduced speeds compared to free-
For our case, we have VR12 =V12+VR=2697+2141=4838pc/h, here
flow conditions, with the average speed typically around 45-50%
based on Table 1 (Exhibit 25-7 in HCM 2010)The maximum flow
of the posted speed limit. Traffic density in LOS E is relatively
entering the influence area exceeds the desired maximum flow as
high, and vehicles are closely spaced. The roadway is operating at
recommended by HCM, but there will be no LOS F according to
or near its maximum capacity. Travel time in LOS E is longer than
this second condition:
under free-flow conditions, with moderate delays and periodic
fluctuations in speed due to congestion. Vehicles in LOS E may
2nd condition states that when the total flow entering the ramp in-
frequently need to accelerate and decelerate due to the changing
fluence area exceeds its maximum desirable level but the total
traffic conditions. This is particularly evident in areas with merg-
freeway flow (v) does not exceed the capacity of the downstream
ing lanes, intersections, or other areas where traffic flow is dis-
freeway segment. In this case, locally high densities are expected,
rupted(Highway Capacity and Level of Service, 1991; Volosenko
but no queuing is expected on the freeway. The actual lane distri-
& Laurinavičius, 2020).
bution of entering vehicles is likely to consist of more vehicles in
Although lane changes may still occur in LOS E, they may be less
the outer lanes than is indicated by the models herein. Overall, the
frequent and more difficult due to the high traffic volume and lim-
operation will remain stable, and LOS F is not expected to occur.
ited gaps between vehicles. The roadway is operating at or near
But when the total downstream flow exceeds the basic freeway
its maximum capacity, and small disruptions or fluctuations in
capacity of the downstream segment, LOS F exists. In such cases,
traffic flow can quickly lead to congestion or breakdown condi-
no further computations are needed, and LOS F is assigned. For
tions. LOS E may be perceived as moderately uncomfortable for
30
J.C. Sugira et al. Engineering Perspective 3(3): 27-34, 2023
drivers. The close proximity of vehicles and the need for frequent and researchers to model and analyze traffic operations and be-
adjustments in speed and position can contribute to a higher level havior in various transportation systems. PTV Vissim enables us-
of driver stress. Safety concerns may arise in LOS E due to the ers to simulate and visualize the movement of individual vehicles
increased traffic density and closer spacing of vehicles, with a and their interactions within a detailed virtual environment (Pla-
slightly higher risk of rear-end collisions, lane merging conflicts, nung Transport Verkehr GmbH - Copyright, 2022; Utomo et al.,
and other traffic-related incidents compared to lower levels of ser- 2020).
vice (Abdullah & Sadullah, 2017) Setting Up the VISSIM Model for the Study Area;
4.1.3. Computing speeds as supplemental information; VISISM needs to prepare to run the simulation of estimating the
capacity of the study area. There are a few steps in VISISM mi-
Finding average speed in the outer lane Ramp influences area; crosimulation that have been described below.
𝑉𝑅12
1. An overlapped geometry of the study section was drawn in
𝑀𝑆 = 0.321 + 0.0039𝑒 1000 − 0.004(LAeff𝑆𝐹𝑅 /1000)
where the total section of the bottleneck was redrawn in VISSIM
SFR is the free-flow speed of the ramp (km/h); as it is shown in Figure 3.
LAeff = 2LA1 + LA2=240m 2. Next, five types of vehicles were selected. Car and heavy goods
vehicles (HGV) are considered traditional vehicles, For all types
where: LAeff Is the effective length of the acceleration lane of vehicles, driving behavior was determined as Freeway (free
𝑆𝐹𝑅 = 45𝑘𝑚/ℎ, Here it indicates that the capacity of the ramp lane selection)
roadway itself should be 3500 pc/h for Two-Lane Ramps accord- 3. The desired speed was customized as a requirement. In this pa-
ing to Exhibit 25-3 in HCM2010 per, a total of 8 types of speed distributions were selected with
𝑀𝑆 =0.321+0.492-0.0432=0.77 where: Ms is the intermediate lower and upper bounds. The calibration, validation, and
speed determination variable for the merge area simulation setup for various free-flow speeds describes a detailed
𝑆𝑅 = 𝑆𝐹𝐹 − (𝑆𝐹𝐹 − 67)𝑀𝑆 speed distribution.
4. The next part includes vehicle composition, In the vehicle input
𝑆𝑅 =110-(110-67) *0.77=76.89 Km/h part, traffic demand needs to be set as input. This demand value
was found in the collected data. For capacity checking, traffic de-
𝑉𝐹 −𝑉12 4824−2974
𝑉𝑂𝐴 = = = 925pc/h mand was kept at 6965veh/h for all five lanes. For vehicle inputs,
𝑁𝑂 2
each lane was given a number of vehicles that is similar to that
Where VOA = average per-lane flow rate in outer lanes at begin of one contained in collected trajectory data. The first 3 minutes were
ramp influence area spent for warm-up, which means running different vehicles ini-
SR = space mean speed of vehicles within ramp influence area tially and counting for the next 60 minutes.
(km/h); for merge areas, this includes all vehicles in VR12; for di- 5. Finally, the simulation parameter setup and evaluation parame-
verge areas, this includes all vehicles in V12 ter setup were completed to get the result. A total of 3 data
Here 𝑉𝑂𝐴 lies between 500pc/h to 2300 pc/h, then the second for- collection points were set up at the end of each lane to get the
mula in HCM has been used; VISSIM output.
S𝑂 = 𝑆𝐹𝐹 – 0.0058(V𝑂𝐴 – 500) , 𝑆𝑂 = 110 − 0.0058(925-
500) =107.535Km/h
𝑉 +𝑉𝑂𝐴 𝑁𝑂
𝑺𝒘 = 𝑉1212 𝑉𝑂𝐴 𝑁𝑂 = 86.31Km/h
+
𝑺𝑅 𝑺𝑶

where: SO = space mean speed of vehicles traveling in outer lanes Figure 3. First simulation in Vissim without Ramp metering
(Lanes 3 and 4, where they exist) within 450-m length range of
ramp influence area (km/h);
SFF = free-flow speed of freeway approaching merge or diverge 4.3. Simulation Results
area (km/h); The results of this simulation were characterized by congested
VR12 = sum of flow rates for ramp (VR) and vehicles entering traffic conditions in the merging influence areas as it is shown in
ramp influence area (pc/h); Figure 3 and in table 3, The simulation results proven it, as it is
VOA = average per-lane flow rate in outer lanes at the beginning characterized by with long queues, slow speeds, and significant
of ramp influence area delays. The simulation results demonstrate a decrease in average
NO = number of outside lanes in one direction (not including ac- and maximum speeds compared to other free-flow conditions, in-
celeration or Deceleration lanes or Lanes 1 and 2), dicating reduced mobility and capacity utilization on the road net-
VF = total approaching freeway flow rate (pc/h), and work. The congestion leads to increased travel times, with longer
V12 = demand flow rate approaching ramp influence area (pc/h) journey durations, moreover, it was associated with reduced traf-
fic flow rates. In order to propose a method by which the per-
4.2. Simulation of collected data using PTV VISSIM formance of this bottleneck can be improved, the simulation was
repeated with the same conditions as the first one. But by applying
PTV Vissim is a leading software application for microscopic traf-
ramp metering at the entrance ramps as it is shown in Figure 4,
fic simulation. It is widely used by transportation professionals
31
J.C. Sugira et al. Engineering Perspective 3(3): 27-34, 2023
Ramp metering was applied to section of the road under study. 5. Results interpretation
Ramp metering systems typically consist of traffic signals in- After conducting the analysis using the Highway Capacity
stalled at the entrance ramps. After setting ramp metering to Manual (HCM) methodology and simulating the performance of
VISSIM software the simulation was repeated again. After the the bottleneck using VISSIM, the results indicated that the Level
ending of the second simulation, the results showed reduced queu- of Service (LOS) of the bottleneck was initially evaluated as E.
ing and improved overall performance of the bottleneck as its ca- This LOS suggests significant congestion and delays, indicating a
pacity falls under LOS D. And also, simulation results demon- suboptimal traffic flow. To improve the performance of the bot-
strated a noticeable reduction in delays and smoother traffic pro- tleneck, a ramp metering strategy was implemented using
gression through the bottleneck. VISSIM simulation. The ramp metering approach involved con-
trolling the flow of vehicles entering the mainline from the on-
ramp. By regulating the rate at which vehicles merge onto the
mainline, it aims to alleviate congestion and enhance overall traf-
fic flow efficiency. The comparison of results before and after the
implementation of ramp metering provides significant insights
into the effectiveness of this traffic management strategy. The av-
erage free-flow speed for the downstream segment saw a notable
increase from 110 km/h to 120 km/h after ramp metering was in-
troduced. This improvement signifies enhanced traffic flow and
better mobility within the downstream segment, attributed to the
controlled entry of vehicles through ramp metering.
The merge area speed, a critical factor in bottleneck perfor-
mance, exhibited a positive change as well. Prior to ramp metering,
the merge area speed was 76.89 km/h, which notably rose to 87.4
km/h post-implementation. This change indicates that the intro-
duction of ramp metering has facilitated smoother merging of ve-
hicles at higher speeds, effectively reducing the potential for con-
gestion and delays in this critical zone.
Figure 4. Second simulation process after application of Ramp metering The queue length, often indicative of the level of congestion
and delays experienced by drivers, showed a substantial reduction
from 102 meters before ramp metering to a significantly dimin-
The effectiveness of these strategies was assessed through com-
ished 40 meters after its application. This reduction in queue
prehensive performance metrics derived from the simulation re-
length translates to shorter waiting times for vehicles at the merge
sults. Figure 4 also proves the successful contribution of Ramp
point, resulting in improved traffic flow and a less congested road-
metering to the reduction of Queue length, as there are some
way environment.
empty spaces between moving vehicles.
One of the most critical aspects, delay per vehicle, experienced a
substantial decrease from 85 seconds to 32 seconds following the
Table 3. VISSIM simulation results before and after the application of
implementation of ramp metering. This decrease indicates that ve-
Ramp metering
hicles spent significantly less time in traffic queues, leading to
Performance Indicator Before R After Ra
shorter travel times and an overall enhancement in traffic effi-
amp Met mp Mete
ciency. Additionally, the density of vehicles on the roadway, a vi-
ering ring
Average free flow Speed for dow 110 120 tal measure of traffic congestion, exhibited a positive change.
nstream segment (km/h) Density decreased from 24.37 pc/km/ln to 20.2 pc/km/ln after
Average Merge area speed (km/h) 76.89 87.4 ramp metering. This reduction implies that vehicles were better
spaced out, contributing to smoother traffic flow and a less con-
Queue Length (m) 102 40
gested road network.
Delay (s/vehicle) 85 32
Overall, the Level of Service (LOS) transitioned from E to D
Calculated Density (pc/km/ln) 24.37 20.2
after ramp metering was introduced. This shift underscores a sub-
stantial enhancement in the overall traffic conditions, with re-
The simulation demonstrated a significant improvement in the duced congestion, lowered delays, and improved traffic flow. Top
performance of the bottleneck, as evidenced by reduced queue of Form
lengths, improved average speeds, and decreased travel delays. The successful implementation of ramp metering suggests its
These positive outcomes validate the efficacy of the implemented effectiveness in mitigating congestion and improving the bottle-
strategies and highlight the potential benefits of using PTV Vissim neck's performance. The results align with previous studies and
simulation as a decision-support tool for bottleneck manage- empirical evidence supporting the positive impact of ramp meter-
[Link] of Form ing on traffic flow and capacity enhancement. These findings
demonstrate the practical application of the HCM methodology in

32
J.C. Sugira et al. Engineering Perspective 3(3): 27-34, 2023
evaluating bottlenecks and identifying potential solutions for im- ing the Department of Transportation Engineering, have signifi-
proving traffic performance. Moreover, the VISSIM simulation cantly contributed to the success of this study. We are sincerely
provided a valuable platform to assess the effectiveness of the thankful for everyone's collective efforts in making this endeavor
ramp metering strategy and its impact on the bottleneck's LOS. possible.

6. Conclusions Conflict of interest declaration:

In conclusion, the application of the Highway Capacity Manual The authors declare that they have no conflicts of interest con-
(HCM) methodology combined with VISSIM simulation has pro- cerning this article.
vided valuable insights into the performance of a bottleneck and
the effectiveness of a ramp metering strategy. The initial evalua- Credit Author Statement
tion of the bottleneck indicated a Level of Service (LOS) E, high- Sugira Jean Claude: Conceptualization of the study, devising the
lighting significant congestion and delays. However, after imple- methodology, developing the necessary software tools for data
menting the ramp metering strategy in the simulation, the LOS analysis, writing of the original draft.
improved to D, indicating a noticeable enhancement in traffic flow Nshimiyimana Marc: Conceptualizing the research, overseeing
and a reduction in congestion. These results align with previous the project's progress, and writing the original draft.
studies [2] that have demonstrated the effectiveness of ramp me- Nsengimana Jean Pierre: Curating the trajectory data collected
tering in improving traffic performance. The successful applica- from the website "[Link].", visual representations, and in-
tion of the HCM methodology and VISSIM simulation empha- terpretation of the results.
sizes the importance of utilizing advanced tools and techniques
for traffic analysis and mitigation strategies. By employing these References
methods, transportation planners and engineers can gain valuable 1. Abdullah, R., & Sadullah, A. F. M. (2017). A study on traffic
insights into the causes of congestion and identify effective solu- weaving segment level of service on Malaysia urban highway.
tions for enhancing traffic flow and reducing delays. AIP Conference Proceedings, 1865.
The results strongly suggest the implementation of a ramp me- [Link]
tering strategy at the bottleneck location to improve traffic flow 2. Highway Capacity and Level of Service. (1991).
and reduce congestion. Ramp metering should be further studied [Link]
and optimized by considering factors such as traffic demand pat- 3. Highway Capacity Manual 2000 RAMPS AND RAMP
terns, control algorithms, and coordination with adjacent intersec- JUNCTIONS CONTENTS. (n.d.).
tions. 4. Mahdi, M. B., Alrawi, A. K., & Leong, L. V. (2022). Compat-
Continuous Monitoring and Evaluation: Regular monitoring ibility between delay functions and highway capacity manual
and evaluation of the bottleneck's performance are crucial to iden- on Iraqi highways. Open Engineering, 12(1).
tify any changes in traffic conditions and assess the effectiveness [Link]
of the implemented ramp metering strategy. This will enable 5. Planung Transport Verkehr GmbH - Copyright, P. (2022).
transportation authorities to make timely adjustments and fine- PTV Vissim 2023 - First Steps Tutorial. [Link]
tune the system for optimal performance. 6. Raju, N., Arkatkar, S., & Joshi, G. (2018). Examining Effect
Integration of ITS technologies, such as real-time traffic moni- of Bottleneck on Multi-Lane Roads at Midblock Sections Us-
toring, adaptive control systems, and advanced data analytics, can ing Simulation. 697–705.
further enhance the effectiveness of the ramp metering strategy. [Link]
These technologies can provide valuable data for decision-making, 7. Rouphail, N. M., & Williams, B. M. (n.d.). ABSTRACT
optimize control parameters, and facilitate proactive traffic man- ANXI JIA. Stochastic Capacity at Freeway Bottlenecks with
agement. Comprehensive Analysis of Alternative Strategies: Application to Travel Time Prediction. (Under the direction
While ramp metering has shown promising results, it is essential of Dr.
to explore alternative strategies and conduct a comprehensive 8. Ryus, P., Vandehey, M., Elefteriadou, L., Dowling G, R., &
analysis to identify the most effective combination of measures Ostrom K, B. (2011). New TRB Publication: Highway Capac-
for further improving the bottleneck's performance. This may in- ity Manual 2010. TR News, 273.
clude considering changes in lane configurations, signal timing 9. Ryus, P., Vandehey, M., Elefteriadou, L., Dowling, R. G., &
adjustments, or other operational and infrastructure improvements. Ostrom, B. K. (2011). Highway capacity manual 2010. TR
By implementing these recommendations, transportation agencies News, 273.
can strive towards achieving a more efficient and sustainable 10. Skabardonis, A., Papadimitriou, F., Halkias, B., & Kopelias,
transportation system, providing better mobility and reducing P. (2016). Operational Analyses of Freeway Off-Ramp Bot-
congestion for the traveling public. tlenecks. Transportation Research Procedia, 15.
[Link]
Acknowledgment 11. Suthanaya, P. A., & Upadiana, N. (2019). Traffic management
Despite not receiving official, assistance, our friends and fami- of Udayana University Sudirman campus intersection using
ly's belief in our capabilities has been a constant motivation. The Vissim software. MATEC Web of Conferences, 276.
research participants and the resources available at SEU, includ- [Link]

33
J.C. Sugira et al. Engineering Perspective 3(3): 27-34, 2023
12. Utomo, E. W., Ramadhan, F., & Imran, A. (2020). VISSIM
Simulation-Based Analysis for Improving Traffic Conditions
in Bandung Indonesia. International Journal of Simulation:
Systems, Science & Technology. [Link]
sst.a.21.03.05
13. Volosenko, S., & Laurinavičius, A. (2020). Level of service
evaluation methods: Possible adaptation for Lithuania. Baltic
Journal of Road and Bridge Engineering, 15(2).
[Link]
14. Wang, X., Hadiuzzaman, M., Fang, J., Qiu, T. Z., & Yan, X.
(2014). optimal ramp metering control for weaving segments
considering dynamic weaving capacity estimation. Journal of
Transportation Engineering, 140(11).
[Link]
15. WSDOT. (2014). Protocol for VISSIM Simulation. Washing-
ton State Department of Transportation, 1(September).
16. Zhao, J., Ma, W., & Xu, H. (2017). Increasing the Capacity of
the Intersection Downstream of the Freeway Off-Ramp Using
Presignals. Computer-Aided Civil and Infrastructure Engi-
neering, 32(8). [Link]

34