Innovative Infrastructure Solutions (2021) 6:125

https://doi.org/10.1007/s41062-021-00493-1

TECHNICAL PAPER

Analysis of delay and queue length variation at three‑leg signalized

intersection under mixed traffic condition
Dungar Singh1 · Tazim Ameen1 · Abdullah Ahmad1

Received: 27 October 2020 / Accepted: 23 February 2021 / Published online: 28 March 2021
© Springer Nature Switzerland AG 2021

Abstract
The rapid growth in the economy has increased personal vehicle usages which is responsible for hefty road traffic conges-
tion in all major cities. In delay variations at a signalized intersection, there is a significant impact on travel time among
the number of various other attributes. Even though the impact of the above-mentioned distinctions on the perspective for
design, planning, and analysis of signal control is having an important insinuation nevertheless one of the major problems
in the urban area at present scenario is because of the delay due to traffic congestion. Measurement of delay depends upon
various parameters such as signal cycle, saturation flow rate, arrival and departure of the class of particular automobiles,
types of the intersection, the grade of an intersection, amount of traffic volume, the queue length of traffic, composition of
vehicles, and type of traffic control system. Vehicle movement related to a signal indicator, non-stopped movement during
the green time, stop and queue movement during red time include a delay that might be due to over-saturation flow in traffic
or pedestrian flow, etc. The aim of the paper is to measure the delay variability and find out what amount of delay accounts
for the three-leg signalized intersection in total travel time. Delay variability helps in predicting total travel time variability
and helps in estimating accurate travel time. It is seen that the stopped delay is the primary affecting constraint in the inter-
section point which is responsible for the long queue length of the street and it is predominantly founded on the position of
the vehicle in the queue line and the manual control of the length of the cycle at a signalized intersection.

Keyword Travel time · Conflict zone · Delay time variation · Cycle length · Dilemma zone

Introduction the immense flow of traffic. Unnecessarily high traffic flow,

ineffective road capacity, geometric condition, poor pedes-
Intersection delay has a massive influence on the proper trian infrastructural facilities, vehicle parking on main roads,
operation of an intersection The delay of traffic signals is and significant lateral movement because of the absence of
one of the serious challenges affecting road users in develop- lane discipline, and drivers prefer to use the transverse gaps
ing countries such as India in their regular lifestyle. Gener- to reach the front of the queue, is one of the factors caus-
ally, at intersections, the cause of the delay is due to vari- ing congestion which results in delays and excessive queue
ous motives. Vehicle delay is the time loss witnessed by a lengths. Jaipur is the capital city of Rajasthan and among
vehicle while traversing an intersection, also delineated by the fastest-growing cities in India. The selected intersection
the lag time among both arrival and departure times of vehi- is Laxmi Mandir which is at a remoteness of 1 km from
cles. The delay at the intersection consists of three parts viz. the Gandhi Nagar railway station in Jaipur city. The junc-
deceleration delay, stopped delay, and acceleration delay [1]. tion carries traffic from all three directions i.e., Tonk Road,
During peak periods, vehicle drivers suffer delays due to Sawai Ram Singh Road, and Sahkar Marg Roadside which
are considered as leg-1, leg-2, and leg-3, respectively, in this
* Tazim Ameen paper. The area is surrounded by different textile markets,
[email protected]; [email protected] coaching centers, service centers like banks, travel agencies,
Abdullah Ahmad etc. Handling traffic operations at this junction have always
[email protected]; [email protected] been a challenging task for the municipal corporation and
the enforcing agencies due to a very high volume of mixed
1
Department of Civil Engineering, National Institute traffic. The flow of traffic via this junction is governed by a
of Technology, Srinagar, India

13
Vol.:(0123456789)
125 Page 2 of 11 Innovative Infrastructure Solutions (2021) 6:125

type of fixed traffic signal. As far as the intrinsic worth of the duration of the cycle at the intersection. Saw et al. [12]
this intersection, analysis of delay of traffic movement is developed a model to study delay components at signalized
considered desirable at this intersection. The intersection intersections by introducing two-wheeler and three-wheeler
in the present time has an oversaturated flow which results as probe vehicles as a part of mixed traffic flow. Adopting
in higher delay and finally increases delay variability. The GPS receivers that provide an ability to precisely calculate
intersection is the location where the driver tries to increase intersection delay, comprising entirely of deceleration delay,
or decrease the speed based on traffic flow and signal indica- stopped delay, and acceleration delay, Alkaissi, and Hussain
tion. Sometimes when the flow is less the aggressive drivers obtained a quite practical delay time model [1]. Chaudhry
attempt to increase the speed and try to exit the intersec- MS and Ranjitkar P [13] suggested an empirical model to
tion before green lights turn red. The distance covered in integrate this growing trend in the saturation flow rate for the
which the driver thinks to accelerate or retard is termed as study of capacity and signal timing. This research explores
a dilemma zone. The speed reduces considerably at a sig- queue discharge activities at signalized intersections relying
nalized intersection. The point at which the speed of the on data obtained from six signalized intersections in Auck-
driver decreases and reaches zero and finally accelerating land, New Zealand. These field evaluations support what has
up to the normal speed is considered as an intersection con- been discerned anywhere else that the saturation flow rate
flict area. This study deals with finding the causes of delay enhances with green time. Ameena S et al. [14] have made
and measurement of these delays at various sections of the an effort to determine the average space headway using a
intersection. model-based approach with a unique citation to congestion
measurement for Intelligent Transportation Systems (ITS)
applications. The density within the segment was made use
Literature review of for the prediction of space headway. This research study
made use of the Extended Kalman Filter (EKF) for calcu-
A study of the available literature shows that a great deal of lation structure. Preethi P et al. [15] proposed a report, in
emphasis has been paid to delaying analysis at the signalized which Webster’s delay model is updated to match the road
intersection, research has been carried on delay models in traffic situations in India. Data on traffic flow parameters,
developed countries. Webster [2] generated a delay model signal timing information, and vehicle delay are obtained
for under-saturated flow at signalized intersections with from various signalized intersections with widely different
homogeneous traffic scenarios based on established queu- field conditions by means of a video survey and a Global
ing analysis. Furthermore, several researchers improved Positioning System (GPS). The semi-empirical adjustment
Webster’s delay model for oversaturated flow conditions, term of the Webster model has been updated and adjusted
and the principal ones are the Robertson model (Hurdle, on the basis of site investigation of delay for various control
1984), the Akcelik model (Roess et al., 2006), and the HCM constraints at Intersections with signalization. In this study,
2010 model [3–5]. Nevertheless, all these models have been the adjustment term was calculated and modeled as per the
designed for homogeneous traffic situations. Nedevska et al. Artificial Neural Network approach. Maurya AK et al. [16]
[6] analyzed delay models and the level of service of the in their research paper provided an understanding of the
road for an urban and rural area with a defined delays analy- layout of speed and headway distribution in diverse traffic
sis facility. However, these models have been proposed for situations encompassing multiple motorized and non-motor-
homogeneous traffic situations and have not calculated pre- ized automobiles with different speeds. Shawky M et al. [17]
cise delays for heterogeneous traffic conditions [7]. Arasan came up with research on examining the start-up delay at
and Jagadeesh [7] under heterogeneous traffic conditions, signalized intersections in Abu Dhabi City, UAE. The study
measured the saturation flow and the delay in traffic at a further deals with the effect on traffic efficiency and protec-
signaled intersection. A framework was designed by Arasan tion of the demonstrating lead/lag phase-in scheme. They
and Naidu [8] to determine the average delay under het- concluded that lead/lag phasing has gradually improved traf-
erogeneous traffic scenarios at a signalized intersection. fic efficiency by minimizing the computational intersection
Darma et al. [9] estimated a delay at a signalized intersec- delay but has an adverse effect on safety. Partha PD et al.
tion with a number of parameters influencing control delay. [18] studied the queue discharge attributes at signalized
Hoque and Imran [10] suggested a paradigm shift to the intersections under congested circumstances. The report pre-
delay model of Webster to accommodate the highly diverse sents the speed, headway, and acceleration behavior of vehi-
traffic situation, and by changing the adjustment term, a mul- cles throughout the queue release after green onset under
tiple linear regression model was suggested. Nithya et al. complex traffic situations. Maini P and Khan S [19] made
[11] suggested a model based on GPS data of stopped delay an analysis of the speed of vehicles moving through the
modeling at the intersection and calculated the dimension of intersection area for four multiple motorized vehicles viz.
impact affecting the queue of total control delay effect and two-wheeler, three-wheeler, cars, and buses. The analysis

13
Innovative Infrastructure Solutions (2021) 6:125 Page 3 of 11 125

indicates that throughout the intersection domain the speed taken and is segmented into five slots to note the traffic
at which vehicles travel does not differ greatly depending on speed behavior. The sectioning of the stretch is done prior
the type of vehicle. Radhakrishnan S and Ramadurai G [20] to the intersection proper and thereafter until the vehicle
introduced a novel method to calculate the discharge head- regains normal speed.
ways in mixed traffic scenarios which are characterized by At the intersection, the video survey was conducted for
a lack of lane discipline. To model the discharge headway, a 3 h in the morning. (8:00 A.M to 11:00 A.M) and 2 h of the
linear mixed-effect regression was used. evening time (5:00 P.M to 7 P.M). In this research paper, the
morning peak hours from 9:00 A.M. to 11:00 A.M. are key
study hours. The following categories of vehicles have been
Research objectives considered viz.; two-wheeler, three-wheeler, four-wheeler,
Light Commercial Vehicle (LCV), and Multi-Axle Vehicle
The principal aim of the research is to determine the delay at (MAV). The information about the existing cycle time at
the signalized 3-leg intersection and to analyze the variation the signalized intersection is also collected. This study is
of queue length and focus on: particularly focused on the traffic flow from Tonk Road (leg-
1) toward the Sawai Ram Singh Road (leg-2) signal system
• Causes of delay at the signalized intersection and find the at the intersection. The cycle length of the signal phase is
remedial solution for the delay causes. 110 s of which 48 s are green periods and 60 s of red peri-
• Traffic flow behavior concerning signalized intersection ods followed by 2 s of the amber period for the main road.
and to section the stretch along the major road of the Free left turn of vehicles toward the Sahkar Marg Road have
intersection concerning the speed distribution. not been considered in this study, since it is not conflicting
• Delay and queue length measurement and their variabil- with other vehicles coming from all other approaches of the
ity corresponding to signal cycle length at the intersec- signalized intersection. The three-phase signal system of the
tion. intersection is presented in Fig. 1. The phase lengths of red,
• Speed assessment across the intersection with respect to green, and yellow phases are shown with their correspond-
delay time. ing colors in Fig. 1. Also for leg-2, two values of the green
phase are shown one corresponding to through traffic and
the other for the right turn, the value for the right turn being
Data collection and study area the minimum of the two.

Laxmi Mandir is having the one of busiest signalized

intersections in Jaipur city and having large vehicular and
pedestrian traffic. The video graphic survey was carried Classified volume count survey
out using a portable digital video camera deployed on the
top of the four-storied building at a height of 10 m above The video-graphic investigation has been conducted to
the ground on the opposite side of Sahkar Marg Road. In know the traffic flow and composition pattern for the
which the two legs namely Tonk Road (leg-1) and Sawai morning peak hour period from 9:00 AM to 11:00 AM.
Ram Singh Road (leg-2) are major roads. These two roads The video camera was fixed at a high-rise building having
carry a high intensity of traffic including commercial a wide area coverage at the intersection and traffic record-
vehicles. The third leg leads to Sahkar Marg Road (leg-3) ing was done using a camera. Then the traffic data were
which passes through many residential and commercial extracted for the arriving vehicles with respect to the sig-
establishments. The data have been analyzed for the vari- nal cycle interval. The number of vehicles was counted to
ous delay components and delay variability with reference clear the signalized intersection whose signal cycle length
to study at a signalized intersection and also for signalized is 110 s. Nearly 17 signal cycles were recorded randomly
intersection impact. The study stretch of about 900 m is during the peak period in this study.

Fig. 1  The three-phase signal

system at Laxmi Mandir inter- leg-1 48 sec 2 sec 60 sec
section

leg-2 48 sec 20 sec 2 sec 40 sec

leg-3 70 sec 38 sec 2 sec

13
125 Page 4 of 11 Innovative Infrastructure Solutions (2021) 6:125

Speed and delay survey of the heavy vehicle restriction is quite good and can be
ignored here. The traffic composition at the intersection for
In this study, the speed profile of the corridor was collected leg-1 and leg-2 has been shown in Fig. 5 and 6, respectively.
using probe vehicles (two-wheeler, three-wheeler, and four- Average discharge queue profile of the main road in the
wheeler) to note the moment-to-moment variations in speed. cycle length of 110 s which consists of green periods of
This technique provides a speed traverse of the study cor- leg-1 (Tonk Road) 0 to 50 s, right turn of Leg-3 (Sahkar
ridor of 900 m at the intersection stretch along leg-1 to leg-2. Marg Road) 50 to 90 s, and right turn of leg-2 (Sawai Ram
The schematic diagram of the selected intersection is shown Singh Road) 90 to 110 s all cycle data have been presented
in Fig. 2 in Fig. 7. It is clear from Fig. 7 that the variation of the dis-
charge rate observed at the end of the green period reduces.
Nearly 246 PCUs have been cleared during the entire green
period of the cycle.
Traffic volume and composition

The traffic flow of each approach is presented in pcu/hr with Probe vehicle speed variability
due consideration of the Passenger Car Unit (PCU) factor for
the signalized intersection mentioned in IRC SP: 41(1994) These constraints could be extracted from video graphic
[21]. The traffic flow of the major approach in the direc- footage. The speed profile using the probe vehicles has been
tion of Tonk Road (leg-1) to Sawai Ram Singh Road and carried out during the peak hour period at the intersection
Sawai Ram Singh Road (leg-2) to Tonk Road is 4470 pcu/ for the stretch of 900 m inclusive of the conflict area. To
hr and 3402 pcu/hr, respectively. The right-turning traffic examine this multifaceted traffic flow, video graphic data
from Sahkar Marg Road (leg-3) to Tonk Road is 4195 pcu/ trips were carried out in peak periods using two-wheeler,
hr whereas the right-turning traffic from the main road Sawai three-wheeler, and four-wheeler as probe vehicles. The vehi-
Ram Singh Road (leg-2) to Sahkar Marg Road (leg-3) side cle movement mechanism can be realized for its variation
is 1466 pcu/hr. The right-turning traffic from Sahkar Marg through the formation of five behavior zones with corre-
Road and the through traffic from Sawai Ram Singh Road sponding traffic specifications as shown in Table 1.
make a total volume of 7597 pcu/hr toward Tonk Road. The The stretches around the intersection are divided into five
volume flow diagram for the intersection is shown in Fig. 3 different zones as shown in Fig. 8 in order to have an under-
and the actual traffic scenarios are revealed in Fig. 4. The standing of traffic behavior precisely. In the previous study
traffic composition at this intersection indicates the domina- (Saw et al. [12]), they also adopted the same criteria to study
tion of two-wheelers to an extent of 50% of the heterogene- the traffic behavior at the intersection by dividing the stretch
ous traffic. The mode-wise share of the main road as shown into upstream and downstream sections. A typical aver-
in Fig. 5 and Fig. 6 for four-wheelers and three-wheelers age speed profile was taken by the probe vehicle as shown
is almost 36% and 9%, respectively. In the city area, entry in Fig. 8 which indicates the formation of three segments as

Fig. 2  Schematic diagram of

study at Laxmi Mandir intersec-
tion

13
Innovative Infrastructure Solutions (2021) 6:125 Page 5 of 11 125

Fig. 3  Volume Flow diagram for Laxmi Mandir intersection

1.03 1.19 0.05 0.05 0.05 0.42

Two wheeler
0.83
Auto

Car/Jeep/Taxi

Mini Bus

48.02 Bus
41.23
LCV

Two Axle

MAV

Tractor
7.2
Cycle

Fig. 4  Actual Traffic condition at Laxmi Mandir intersection

Fig. 5  Traffic composition at leg-1 approach at Laxmi Mandir inter-
section

an upstream side, downstream side, and middle intersection Zone‑wise vehicle movement
conflict area Two regions are formed in namely pre and post-
conflict regions concerning the area of intersection as depicted At the signalized intersection, speed transformation is tak-
in Fig. 8. The conflict area i.e., Section III, where conflict takes ing place right from mid-block to the intersection conflicts
place and due to the movement of various types of vehicle and area (S I to S III). When a driver moves ahead of a certain
pedestrian together at the intersection which results in conflict distance in dilemma and decisive zone (S II) during which
points and thus increases the queue length at the signalized the driver observes the circumstances ahead and accord-
intersection. ingly agrees to accelerate, decelerate or move with the
same speed depends upon the intersection signal indicators

13
125 Page 6 of 11 Innovative Infrastructure Solutions (2021) 6:125

1.48 0.07
at the intersection. Deceleration mainly starts when the
2.03 0.21
0.49
Two wheeler vehicle is about to join the queue in the dilemma zone (S
1.62
II) and also in the conflict area zone (S III). The down-
Auto
stream area consists of two sections normalizing zone (S
Car/Jeep/Taxi IV) and mid-block zone (S V), respectively. Vehicle move-
ment starting from the stopped condition at the intersec-
Mini Bus tion belongs to the conflict area (S III). Further, the accel-
30.36 eration of vehicles takes place in the normalizing zone (S
Bus
IV) and attains normal speed. Similarly, S V pertaining
53.08
LCV to the mid-block zone of downstream, there again speed
transformation can be observed in the post-intersection
Two Axle
area.
10.66 Tractor

Cycle Results and discussion

Section‑wise Speed Variations of probe vehicles

Fig. 6  Traffic composition at leg-2 approach at Laxmi Mandir inter-
section The speed variation depends on various factors such as the
driver’s behavior and the driver’s intelligence. The driver
takes a decision based on traffic volume, signal indication,
queue situations, and geometric condition of the intersection.
161
160 The vehicle queue depends on the saturation flow pattern
140
of the vehicle’s arrival. The vehicle queue starts when the
flow signal is red and it was observed that mostly two-wheel-
120 ers accumulate near the stop line as start queue line of the
vehicles, further as soon as the red signal ends vehicles in
flow (pcu)

100
the queue line start moving and will exit the Intersection at
80 73 normal speed. Due to high vehicle flow and merging and
60
diverging conflict between vehicles, makes the remaining
vehicles move at the crawling speed. This leads to a lengthy
40 conflict area which is the S III (Intersection conflict area
20
zone). In the S I (Mid-block zone upstream) and the S V
12
(Mid-block zone downstream), the speed of the vehicle is
0 higher than compared to the remaining sections as clearly
0 - 50 50 - 90 90 - 110
portrayed in Fig. 8. As the vehicle mainly three-wheelers and
Green phase of cycle length (seconds)
four-wheelers leaving the S I (Mid-block zone upstream) and
entering the S II (Dilemma and Decision zone) which is a
Fig. 7  Flow during the green period at the Laxmi Mandir intersection congested area, the speed decreases as it would be difficult
for a three-wheeler and four-wheeler to accelerate in few
and queuing length of traffic. During this time at the inter- seconds. In the S II zone which is a dilemma and decision
section, the flow of a vehicle and conflict between vehicle zone where mostly two-wheeler drivers adapt to acceler-
and pedestrian leads to a longer length of the conflict area ate the vehicle is at high speed as compared to the S III

Table 1  Section zoning particulars

Section Nomenclature Zone Distance in meters (m) Traffic specification

Section I SI Mid-Block Zone (upstream) 0 to 250 Speed ranging from 25 to 21 kmph

Section II S II Dilemma and Decision Zone 250 to 350 Speed ranging from 21 to 18 kmph
Section III S III Intersection Conflict Area Zone 350 to 550 Speed ranging from 18 to 15 kmph
Section IV S IV Normalizing Zone 550 to 650 Speed ranging from 17 to 20 kmph
Section V SV Mid-Block Zone (downstream) 650 to 900 Speed ranging from 20 to 27 kmph

13
Innovative Infrastructure Solutions (2021) 6:125 Page 7 of 11 125

Average Speed of Probe Vehicle (kmph) 33 variation in Tables 2, 3, and 4 has been shown in Fig. 9.
30
Pre-conflict region Post-conflict region
The speed variation of a two-wheeler and, a four-wheeler is
27 higher in the mid-block zone (S I & S V) because of the high
24 accelerating powers. For a four-wheeler, the speed is higher
21 in the normalizing zone (S IV) than the intersection conflict
18 zone (S III) as the four-wheeler has higher speed due to high
15 accelerating power and thus can accelerate in a short dura-
12 Decision zone tion which leads to an increase in their average speed. The
Dilemma and

Normalizing
9 speed of the four-wheeler has reduced significantly in the
Mid-Block zone Intersection
Mid-Block zone
conflict zone (S III) as a four-wheeler driver has less maneu-

zone
6 (upstream) conflict area
(downstream)
3 zone verability compared to a three-wheeler and two-wheeler. The
0 two-wheelers and three-wheelers try to penetrate the queue
250 350 550 650 900
and to fill the available gaps.
Distance (m)
The data presented in Tables 2, 3, and 4 represent the
sectional speed behavior of the two-wheeler, three-wheeler,
Fig. 8  Typical Speed profile of probe vehicle along the section-wise
at Laxmi Mandir intersection and four-wheeler, respectively, with their coefficient of vari-
ance being maximum in section III ( S III).

Table 2  Statistical measures: Section I Section II Section III Section IV Section V

Sectional speeds (Kmph) of a
two-wheeler Average 22 21 16 14 23
Min 21 20 13 13 22
Max 25 25 20 18 28
Standard deviation 2.82 3.53 4.94 3.53 4.24
Coefficient variance 0.12 0.16 0.3 0.25 0.18

Table 3  Statistical measures: Section I Section II Section III Section IV Section V

Sectional speeds (Kmph) of a
three-wheeler Average 20 21 17 15 19
Min 18 18 13 13 16
Max 21 22 18 16 20
Standard Deviation 2.12 2.82 3.53 2.12 2.82
Coefficient Variance 0.1 0.13 0.2 0.14 0.14

Table 4  Statistical measures: Section I Section II Section III Section IV Section V

Sectional speeds (Kmph) of a
four-wheeler Average 22 20 17 18 21
Min 18 19 14 15 19
Max 22 22 19 19 23
Standard Deviation 2.82 2.12 3.53 2.82 2.82
Coefficient Variance 0.12 0.1 0.2 0.15 0.16

(Intersection conflict zone) and S IV (Normalizing zone).

Delay and queue length variation
As from Fig. 8, the S III (Intersection conflict zone) clearly
shows the effect of signalized intersection where the speed
Delay in the Highway Capacity Manual, HCM 2000 con-
is lowest as related to other sections. The deviation and coef-
text is defined as the difference between a travel time
ficient of variations of probe vehicles are higher in the S III
actually experienced in the field and the reference travel
(Intersection conflict zone).
time that would result during ideal conditions [22]. Delay
Similarly, the speed profile of the two-wheeler, three-
mainly consists of three components viz., Deceleration
wheeler, and four-wheeler as observed from sectional speed

13
125 Page 8 of 11 Innovative Infrastructure Solutions (2021) 6:125

two-wheeler 65
30 three-wheeler Four-wheeler
four-wheeler 60 Two-wheeler
Three-wheeler
55

Control delay in seconds per pcu

27
Speed of vehicle (kmph)

50
24
Intersection 45
21 conflict zone 40

18 35
30
15
25
12 20
15
9
250 350 550 650 900 10
Distance (m) 5
0
0 10 20 30 40 50 60 70 80 90 100 110 120
Fig. 9  Sectional speed variation of probe vehicles from leg-1
approaching to leg-2 side at Laxmi Mandir intersection Cycle length (seconds)

Fig. 10  Control Delay variation at the leg-1 of Laxmi Mandir inter-

delay, Stopped delay, and Acceleration delay. Generally, section
a vehicle decelerates while entering the intersection irre-
spective of the signal indicator which contributes to decel-
eration delay. The stopped delay is for the vehicle which The sectional speeds have not been directly correlated
stops during the red period which is further enhanced by with the traffic delay but instead, they are related with the
the vehicles that start to accumulate near the stop line signal cycle lengths i.e., during the red phase the vehicles
to form the queue lane until the signal turns green and will be at standstill condition depicting minimum or zero
acceleration delay is the time taken by a vehicle to move speed thereby causing delays, and during the green phase
from stop condition to regain the normal speed condition. with the vehicles in the movement of a certain speed, thereby
However, vehicles in the queue lane have additional tending to minimize the delays. Thus, it has been shown that
delay components depending upon the vehicle’s location delays have a direct correlation with signal cycle lengths
in the queue lane and crawling distance up to the stop line. as shown in Fig. 10. The values of correlation for delays
Vehicles that are moving with a uniform or normal speed of two-wheeler, three-wheeler, and four-wheeler with the
on an approach roadway toward an intersection tend to signal cycle length as found from the data points presented
decelerate while entering Mid-block Zone (S I). Figure 10 in Fig. 10 have been shown in Table 5.
depicts that at the stop line of leg-1 when the vehicle halts As from Table 5, there is a very much significant correla-
during the red interval of signal, there is a consequent tion among the probe vehicles with the signal cycle length
increase in the deceleration delay. The stopped delay com- thereby indicating that by varying the signal cycle length the
ponent, as is depicted from Fig. 10 suggests that the main delay of vehicles is very much affected. Thus showing the
part of the total control delay is the stopping time for vehi- dependence of vehicle delay on signal cycle length.
cles. Further, when the signal changes to green, vehicle As from Fig. 10, the graph between the control delay and
departure begin which is the cause of acceleration delay. cycle length goes on decreasing with a different rate in the
As can be seen from Fig. 10 that the slope of the graphs red phase of the signal and in the green phase of the signal.
of all the probe vehicles is negative indicating that with The evidence here depicted is of the slope of the graphs for
the onset of the red period there is a decrement in control different vehicles which is negative depicting the decrease
delay and toward the end of the green period, the control in the delay for a particular vehicle. The vehicle arriving at
delay is minimum of all. Furthermore, it can be observed the start of the red phase will have more delay corresponding
that at the end of the red period, all of the probe vehicles to the vehicle arriving later. This variability of delay with
have almost the same delay, with the graph of two-wheeler
a little bit higher than the other two as the two-wheelers
Table 5  Correlation values of probe vehicles
generally tend to be at the front near the stop line because
of their high maneuverability. It is also clear from Fig. 10 Delay time (s)
at the termination of the red phase of the signal duration at Two-wheeler Three-wheeler Four-wheeler
leg-1, the graph of control delay vs. cycle length of all the
probe vehicles follows a slope that is less steep than that Signal cycle − 0.963 − 0.9765 − 0.9805
length (s)
of what follows in the green phase of the signal.

13
Innovative Infrastructure Solutions (2021) 6:125 Page 9 of 11 125

respect to signal time is clearly understood from the graphs delay has been shown in Fig. 11. The control delay at the
having a negative slope. The steepness of the slope confirms signalized intersection is a primary function of cycle length
here the rate of the variability of delays in the signal phases. and has been thus related with a given cycle length of 110 s
This is clearly understood from the graph shown in Fig. 10 at the intersection of our study in Fig. 11.
that up to the end of the red phase of the signal, the vehicle Under the normal situation circumstances, traffic is arriv-
is undergoing the deceleration delay and stopped delay and ing at the intersection approach uniformly and there is a
once the green phase is on, the vehicle undergoes an accel- constant headway between each vehicle but during the red
eration delay until it regains its normal speed condition. interval. The queue of vehicle types lying in wait at the inter-
The percentage-wise composition of control delay com- section starts to expand and thus form a vehicle queue. At
prising of deceleration delay stopped delay and acceleration the completion of the red interval, the queue achieves its full
capacity, as clearly seen from the graph in Fig. 12. Further,
the queue starts to clear once the signal goes green, as vehi-
deceleraon delay (0 - 10seconds) stop delay (10 - 60 seconds) cles depart from the intersection and thus the graph moves
acceleraon delay (60 - 110 seconds)
down as shown in Fig. 12.

9% Speed assessment across the intersection

with respect to delay time

It has been already shown that a strong correlation exists

between the control delay of probe vehicles and the signal
45% cycle length. The variation of delay with respect to the
sectional outline of the intersection can be employed to
assess how the speed profile of vehicles along the inter-
section outline. The variation of delay with the sectional
46% outline of the intersection has been shown in Fig. 13. From
the graph of sectional outline distance in meters versus
delay, the speed of the vehicle can be computed from the
gradient of the graph as shown in Fig. 13. It is evident
from the graph that delay begins to develop as soon as
the vehicle enters Section I (S I) having a certain speed
Fig. 11  Composition of Control delay at leg-1 of Laxmi Mandir inter- which reduces as the vehicle enters Section II (S II). This
section is confirmed by the slope of the graph in Section II (S II)

Fig. 12  Queue length varia-

tion at leg-1 of Laxmi Mandir
intersection

13
125 Page 10 of 11 Innovative Infrastructure Solutions (2021) 6:125

d(D)
= 4 × 10−6 T 4 − 1.2 × 10−3 T 3 + 0.1204T 2 − 4.4621T + 49.586
Sectional outline of intersection (D) (m)
dT
(1)
where D is the sectional outline distance in meters (m)
900
and T is the delay time in seconds (s) and d(D)dT
is the sec-
SV tional speed across the intersection in m/s. The effort was
650 to study the behavior of speed variation of vehicles with
S IV
550 respect to delay. The actual behavior as pertinent from the
Stopped Delay S III Fig. 14 is discrete in nature as the section-wise study was
350
250
S II taken. Our main aim was to capture the similar behavior
SI
with an appropriate accuracy but by a continuous function as
shown in Fig. 14. The fourth-order fitting curve was adopted
0 10 20 30 40 50 60 70 80 90 100 110 120 as its R2 value was found to be quite appropriate for the
Delay (T) (sec/pcu) calculated data. The scope of the Eq. (1) is limited to the
data studied in this paper. For deriving the more generalized
Fig. 13  Variation of delay with the sectional outline of the intersec- equation a bulk of data is required which can be achieved in
tion the future course of action.

which is less steep than that of Section I (S I). The moment

the vehicle reaches the Intersection conflict area i.e., Sec- Conclusions and recommendations
tion III (S III), the graph becomes parallel to the delay
axis, thereby confirming zero slope and thus no speed of Traffic operations in uniform conditions are extremely flex-
vehicles. The length of this portion of the graph represents ible, depending on the geometry of the highway, vehicle
the stopped delay and that in Section I (S I) and Section types, saturation flow, driver behavior, pedestrian character-
II (S II) gives us the deceleration delay. After the signal istic, etc. due to heterogeneity the traffic operations become
turns green, the vehicles in the queue will start leaving much more complicated. It becomes arduous to evolve vari-
the intersection thereby increasing the velocity which is ous parameters that define the traffic operations and to quan-
again confirmed by the increasing slope of the graph until tify each of them smoothly. This paper presents one of the
they attain the uniform velocity. The portion of the graph parameters and an approach to evolve it and the conclusions
in Section IV (S IV) and Section V (S V) represents the drawn after having a thorough study. In a three-leg intersec-
acceleration delay of vehicles. tion from the study of various types of vehicle movement
The graph between the derivative of the sectional out- delay, speed characteristic behavior of traffic, and further
line with respect to delay was plotted and it was assessed analysis of data. It is noted that the reduction in the vehicle
that a fourth-order curve as represented in Eq. (1) proved speed is the cause for the delay and formation of long queue
to be a better fit with R2 = 0.928 where R2 is the coeffi- lengths. The average speed for probe vehicles drastically
cient of determination. reduces to 13 kmph in the intersection conflict area (S III).

Fig. 14  Speed variation with 60

delay time SPEED (CALCULATED)
50
SPEED (FITTED CURVE)

40
(m/s)

R² = 0.9281
30
( )/

20

10

0
0 20 40 60 80 100 120
-10
T (s)

13
Innovative Infrastructure Solutions (2021) 6:125 Page 11 of 11 125

The study emphasizes the three zones namely ’dilemma and 2. Webster FV (1958) Traffic signal settings, road research technical
decision zone’, ’intersection conflict zone’, and ’speed nor- paper no. 39. road research laboratory, Her Majesty’s Stationery
Office, Berkshire, England
malizing zone’ attached with a mid-block section on both 3. Hurdle VF (1984) Signalized intersection delay models- a primer
ends. The dilemma zone is about deciding to stop or accel- for the uniformed. in: signalized intersection delay models- a primer
erate or decelerate the vehicle based on the signal cycle and for the uniformed. Transportation Research Record, pp 96–105
queue conditions. Conflict zone where vehicle merging and 4. Roess R, McShane W, Prassas E (2012) Traffic engineering, 2nd
Editio. Prentice-Hall, New Jersey
diverging and pedestrian movement at a signalized inter- 5. Transportation Research Board (2010) Highway capacity manual
section takes place and is the cause of the increase in cycle 2010. National Research Council, Washington, USA
length at the intersection which in turn causes an increase 6. Nedevska I, Ognjenović S, Gusakova E (2016) Methodology for
in queue length. analysing capacity and level of service for signalized intersections
(HCM 2000). MATEC web conference. https://​doi.​org/​10.​1051/​
It is inferred that the slope of graphs of all the probe matec​conf/​20168​605026
vehicles is negative indicating that with the onset of the red 7. Arasan VT, Jagadeesh K (1995) Effect of heterogeneity of traffic on
period there is a decrement in control delay and toward the delay at signalized intersections. J Transp Eng 121:397–404. https://​
end of the green period, the control delay is the minimum doi.​org/​10.​1061/​(ASCE)​0733-​947X(1995)​121:​5(397)
8. Arasan V, Naidu V (1999) Estimation of Delay to Mixed Trafficat
of all. These shreds of evidence are exactly agreeing with Signalized Intersections: An Empirical Approach. Indian Highw
the actual scenario which the traffic has at the intersection. 27:31–43
The approach length variation is not affecting the average 9. Darma Y, Karim M (2005) Control delay variability at signalized
delay of the traffic at the signalized intersection. At the inter- intersection based on HCM method. Proc East 5:945–958
10. Hoque S, Imran A (2007) Modification of Webster’s delay formula
section approach, deceleration delay constitutes to about under non-lane based heterogeneous road traffic condition. J Civ Eng
9%, stopped delay constitutes about 46% and acceleration 35:81–92
delay constitutes about 45%. Thus, stopped delay has the 11. Nithya S, Senthurkumar D, Gunasekaran K (2014) Delay Measure-
maximum delay component in the control delay constitut- ment of Manually Controlled Intersection Using GPS. Open Civ Eng
J 8:130–135. https://​doi.​org/​10.​2174/​18741​49501​40801​0130
ing more than half of the total control delay. The delays in 12. Saw K, Katti BK, Joshi G (2018) Vehicle movement mechanism and
long queue lengths of vehicles at the intersection reaching delay variability analysis at signalized intersection by probe vehicle
a maximum of about 18.5 m. The remedial solution for the data under mixed traffic condition: a case study in India. Eur Transp
stopped delay which constitutes 46% of the total control 68:1–14
13. Chaudhry M, Ranjitkar P (2013) Capacity and Signal Timing
delay is to maintain the optimum speed of vehicles enter- Analysis of Signalized Intersection with Increasing Saturation Flow
ing the domain of the intersection by having suitable signal Rate. In: Proceedings of 92nd Annual Meeting of 17 Transportation
cycle lengths for the vehicles on the major road. It is also Research Board. Washington, DC
suggested that by selecting the suitable signal cycle lengths 14. Salim A, Vanajakshi L, Subramanian SC (2010) Estimation Of Aver-
age Space Headway Under Heterogeneous Traffic Conditions. Int J
as per the traffic flow corresponding to the particular road so Recent Trnads Eng Technol 3:6–10
that queue length formations are reduced to their minimum 15. Preethi P, Varghese A, Ashalatha R (2016) Modelling Delay at
values. Since the study is based on a fixed type traffic signal Signalized Intersections under Heterogeneous Traffic Conditions.
system thereby it is also suggested that to have a minimum Transp Res Procedia 17:529–538. https://​doi.​org/​10.​1016/j.​trpro.​
2016.​11.​107
value of delay and hence queue length, a sensor-based traf- 16. Maurya AK, Das S, Dey S, Nama S (2016) Study on Speed and
fic signal system may be adopted. Future research should be Time-headway Distributions on Two-lane Bidirectional Road in
performed on different types of intersections such as four-leg Heterogeneous Traffic Condition. Transp Res Procedia 17:428–437.
intersections and multi-leg intersections with a wide range https://​doi.​org/​10.​1016/j.​trpro.​2016.​11.​084
17. Shawky M, Al-Ghafli A, Al-Harthi H (2017) An analysis of the
of traffic and geometric conditions. start-up delay and safety for signalized intersections: Impact of
left-turn phasing sequences. Adv Intell Syst Comput 484:205–218.
Supplementary Information The online version contains supplemen- https://​doi.​org/​10.​1007/​978-3-​319-​41682-3_​18
tary material available at https://d​ oi.o​ rg/1​ 0.1​ 007/s​ 41062-0​ 21-0​ 0493-1. 18. Dey PP, Nandal S, Kalyan R (2013) Queue discharge characteristics
at signalised intersections under mixed traffic conditions. Eur Transp
Compliance with ethical standards 55:1–21
19. Maini P, Khan S (2002) Discharge Characteristics of Heterogeneous
Traffic at Signalized Intersections. In: Proceedings, 4th International
Conflict of interest On behalf of all the authors, the corresponding au- Symposium on Highway Capacity. pp 258–270
thor states that there is no conflict of interest. 20. Radhakrishnan S, Gitakrishnan, (2015) Discharge headway model
for heterogeneous traffic conditions. Transp Res Procedia 10:145–
References 154. https://​doi.​org/​10.​1016/j.​trpro.​2015.​09.​064
21. Indian Road Congress (1994) At-grade intersections in rural and
1. Alkaissi ZA, Hussain RY (2020) Delay time analysis and model- urban areas. Indian Road Congress, New Delhi, India
ling of signalised intersections using global positioning system (gps) 22. Transportation Research Board (2000) Highway capacity manual
receivers. IOP Conf Ser Mater Sci Eng. https://​doi.​org/​10.​1088/​ 2000. National Research Council, Washington, DC, USA
1757-​899X/​671/1/​012110

13