Optimization of Traffic Signal Using Linear

Programing Problem Linear Programing problem

Author M. Surname* and Bauthor D. Surname III
Department of Science, XXX University, Country
* Corresponding author: [email protected]

Abstract

Perspiciatis unde omnis iste natus error sit voluptatem accusantium doloremque
laudantium, totam rem aperiam, eaque ipsa quae ab illo inventore veritatis et quasi
architecto beatae vitae dicta sunt explicabo. Nemo enim ipsam voluptatem quia
voluptas sit aspernatur aut odit aut fugit, sed quia consequuntur magni dolores
eos qui ratione voluptatem sequi nesciunt. Neque porro quisquam est, qui dolorem
ipsum quia dolor sit amet, consectetur, adipisci velit, sed quia non numquam eius
modi tempora incidunt ut labore et dolore magnam aliquam quaerat voluptatem.
Ut enim ad minima veniam, quis nostrum exercitationem ullam corporis suscipit
laboriosam, nisi ut aliquid ex ea commodi consequatur? Quis autem vel eum iure
reprehenderit qui in ea voluptate velit esse quam nihil molestiae consequatur, vel
illum qui dolorem eum fugiat quo voluptas nulla pariatur?
Keywords: keyword 1, keyword 2

1 Introduction
Traffic signals at crossing points exist to keep traffic safe for automobiles as well as pedes-
trians traffic. Crossing points pose the safety risk to traffic because this is where traffic
meets that travels in cross directions. Traffic signals duration must be manipulated in
order to optimize the flow of the traffic through the crossing points. Furthermore, im-
proving traffic signals at crossing points and reducing vehicle idling can also contribute
significantly to lowering fuel consumption and emissions [1] [2].
When designing traffic signals with a fixed timing pattern, the commonly used Webester
formula is often used to calculate the ideal cycle length that minimizes delays, but this
formula tends to give too long of a cycle time when traffic is very congested. [3]. Since
traffic congestion at an intersection is mostly caused by the formulation of long queues,
this study chooses to employ an alternative approach by considering queue length as the
primary measure to address traffic congestion, even though delay-based approaches are
typically more common to determine the optimal signal plan[2]. Using non-standardized
traffic signal durations will improve vehicular and pedestrian traffic flow [4] at intersections
considering the rate of traffic flow and combined with different widths of crosswalks.
Using the number of idling vehicles at the intersection as a measure of current traffic
demand at the intersection of Tatarpur Chowk in Bhagalpur, Bihar, was selected for
analysis in order to determine whether the traffic flow through this intersection could
be optimized during the afternoon rush hour, which is from 3 PM to 5 PM. The signal
control at the intersection applies a pre-timed control of a 125-second cycle length. The
study solves the problem by simplifying the traffic system at the Tatarpur Chowk by
utilizing the linear programming to examine the existing conditions, such as fluctuations

Proceedings of the Samahang Pisika ng Pilipinas

th
39 Samahang Pisika ng Pilipinas Physics Conference
20–22 October 2021
1
of vehicular arrival rates from all directions that were omitted. Alternatively, average
vehicular arrival rates during the two-hour peak were used to determine the total number
of vehicles idling at the intersection.

2 Literature Review
Reference [5] presents a solution that leverages video processing. The live feed is first
processed before being transmitted to the servers, where a C++-based algorithm analyzes
the data to produce results. A comparison between hard-coded and dynamically coded
approaches revealed that the dynamic algorithm performed 35% better.
Reference [6] presents an Arduino-UNO-based system designed to alleviate traffic con-
gestion and reduce waiting times. It captures images using a camera and processes them
in MATLAB, where saturation and hues are removed to generate a threshold image for
traffic density estimation. The Arduino communicates with MATLAB via USB and pre-
installed simulation packages, adjusting green light durations based on vehicle count and
density. However, this approach has several limitations: overlapping vehicles make count-
ing difficult, and non-vehicular objects such as billboards, poles, and trees interfere with
detection since they also appear in black and white.
Reference [7] introduces a fuzzy logic-based traffic light system that adapts to real-
time traffic conditions. It utilizes two fuzzy controllers—each with three inputs and one
output—to manage primary and secondary driveways. Simulations using VISSIM and
MATLAB demonstrated improved traffic flow in low-density scenarios.
Reference [8] proposes an intelligent traffic light system that combines artificial neural
networks (ANN) with a fuzzy controller. Cameras installed at intersections capture im-
ages, which are first converted to grayscale and normalized. A sliding window technique
is then applied for segmentation, allowing vehicle counting irrespective of size. The ANN
processes the segmented images, and its output is used by the fuzzy controller to deter-
mine red and green light durations. The system achieved an average error rate of 2% with
an execution time of 1.5 seconds.
Reference [9] employs a support vector machine (SVM) algorithm alongside image
processing techniques. It captures frames from live video feeds, processes them using
OpenCV, and converts them to grayscale before applying SVM. This system not only
estimates traffic density but also detects red light violations.
Reference [10] suggests an adaptive traffic light timing system that utilizes image pro-
cessing and traffic density analysis. It consists of a microcontroller-controlled traffic light
timer, high-resolution image sensors, MATLAB for processing, and UART-based trans-
mission. However, it does not account for emergency vehicles or detect accidents at
intersections.
Reference [11] reviews various methodologies for traffic light management systems,
noting that they all follow a common framework: selecting input data, extracting traffic
parameters, processing the data, determining traffic density, and updating the system
parameters accordingly.

3 Signalized Intersection
Before drilling into the details of the study, a discussion on the nature of the Tatarpur
chowk signalized intersection shall be presented. At Tatarpur Chowk intersection, there
are two lanes in each direction of the road. From MM College Bhagalpur, a vehicle can
either go straight to Railway station or can take left turn to go Kotwali Chowk. And a

Proceedings of the Samahang Pisika ng Pilipinas

th
39 Samahang Pisika ng Pilipinas Physics Conference
20–22 October 2021
2
vehicle can go to either MM College Bhagalpur by taking a right turn or can go to railway
station road by taking a left direction Kotwali Chowk.
The pre timed 125-second cycle length is adopted for the intersection during the study
period. Besides, all red times are assigned zero for all phases at the intersection in this
study. In addition, the startup delays at the starting of the green indication and the
clearance delays are assumed zero and so are neglected in this models. Subsequently, the
term green time is composed of green and yellow time. On the other hand, the term red
time in this study accordingly means effective red time itself.
The signal control at the intersection consists of three phases in which out of these
three phases, two sides of movements simultaneously consolidate vehicles in their lanes.
When the green time of any phases of given movement ends, the vehicle on those lanes
are signaled to stop, and vehicles on the next lane are signaled to proceed. The system
runs in this manner until all three phases are activated, starting from the MM College
Bhagalpur right turn, Raliway Station left turn, and Kotwali Chowk right turn, in that
order, to complete the cycle.

[ Picture]

Here fig-... illustrates the two letter notation for vehicle arrival rate variables which are
explained as follow.
SR- Average vehicle arrival rate at the Railway station road.
NR- Average vehicle arrival rate at the MM College Bhagalpur road.
TR- Average vehicle arrival rate at the Kotwali Chowk road.
Their effective green time are denoted by G1 , G2 and G3
G1 - Effective green time for TR phase.
G2 - Effective green time for SR phase.
G3 - Effective green time for NR phase.

4 The Linear Programming Model

The goal is to reduce to the minimum the number of automobiles idling at the intersection,
considering the typical arrival rates of vehicles and existing effective green duration for
vehicles to turn right or left or making through movement. In order to approximate
the number of vehicles, the study multiplied the arrival rate of automobiles travelling in
each movement by the duration of the corresponding red time assigned to the affected
traffic lanes. Put another way, the queue length for each movement may be calculated by
multiplying the rate of arrival by the corresponding idle time.
As an example, vehicles turning left from MM College Bhagalpur to Kotwali Chowk
and going through from MM College Bhagalpur to Ralilway Station are in the lane NR,
and the duration of the corresponding green signal leveled as G3 . The duration of this
cycle is G2 +G1 + G3 in that order. Thus, the resultant queue length is NR(G2 + G1 ).
Therefore, the objective function is,
Min. TR(G2 + G3 ) +SR(G1 + G3 ) +NR(G2 + G1 ).

5 Identification of Constraints
Effective green times should be maximized for those traffic lanes having the relatively
higher rates of traffic flow. However, pedestrian safety and vehicle flow rate, along with
queue length, need to be considered as essential constraints for the traffic signal time at
the intersection.

Proceedings of the Samahang Pisika ng Pilipinas

th
39 Samahang Pisika ng Pilipinas Physics Conference
20–22 October 2021
3
5.1 Pedestrian Constraints
According to the US Federal Highway Administration, the standard walking speed of
pedestrians should be 4 feet per second for safe crosswalk countdown time at the intersec-
tion [12]. The width of the road at Tatarpur Chowk is 8 meters i.e, around 26 feet. This
length is constant for all the directions. In order to address safety concerns, the study
uses the longest crosswalk on each street in order to establish a safe duration of the green
time given to pedestrians. The minimum green time, corresponding to pedestrian safety,
are determined by dividing the road width by the suggested pedestrian walking speed of
4 feet per second. So the Pedestrain constraints are,
26
G2 > 4
26
G3 > 4

i.e

G2 > 6.5
G3 > 6.5

5.2 Queue Length Constaints

In order to set the limit of the maximum number of vehicles waiting to make the turn
for each movement, constraints related to queue length for all the considered three turns
are set in this study. To optimize vehicular flow through the intersection, a maximum
of ten idling vehicles was used as a limiting factor. Thud, the maximum number of
vehicles in each lane waiting for a green indication is 10 multiplied by the number of lanes
available for the traffic to proceed in each movement. Hence, the following constraints
are established to control queue lengths at the intersection. Here, each direction of the
intersection contains two lanes. Thus,

NR(G2 + G1 ) ≤ 20
SR(G1 + G3 ) ≤ 20
TR(G2 + G3 ) ≤ 20

5.3 Constraints Related to Vehicle Flow

Constraints related to vehicle flow were established for all the vehicles making through,
left, and right turns. For automobiles making through movement, the absolute distance
for the final vehicle to proceed across the intersection was the width of the perpendicular
road with an additional length of ten vehicles. as the standard length has been taken of the
vehicle from the Indo-HCM. Since we have data of the cars, bikes, bus and autorikshaw.
Thus, the average length of these vehicles is 15.50 feet. Besides the standard vehicle length
ranges form ..... feet for bikes to .... feet for bus. Thus, the vehicle length of 15 feet was
estimated to calculate the constraints in this study. So this produced an estimated travel
distance of 26+(10∗15) feet = 176 feet for a vehicle proceeding through on either Railway
Station or MM College Bhagalpur, as shown in fig..../
The variation in vehicle turning radius due to differences in size and power must also
be considered. However, when the turning radius is the same for all vehicles, the study
assumes that each curve distance can be approximated by a quarter of a circle’s perimeter,
where a single radius is used. So the curve distance is 14 (2πr) .
For example, the curve distance for a vehicle coming from Raliway Station and making
the right turn to go Kotwali Chowk is approximated using a turning radius of 13 feet.

Proceedings of the Samahang Pisika ng Pilipinas

th
39 Samahang Pisika ng Pilipinas Physics Conference
20–22 October 2021
4
Thus the corresponding curve distance is estimated to be 12 π ∗ 13 = 20.42 feet shown in
fig..../
In this study, the speed limit of vehicle has been taken ...... kmph unless otherwise
posted. However, a vehicle ldling at an intersection starts form a speed of zero when the
signal turns green. In addition, vehicles either going right or left end to travel slower
than vehicles making through movement. In order to provide a realistic speed of vehicles
turning left or right through this intersection, a speed of ......feet per second was applied.
The minimum green time under the vehicle flow constraint for each straight movement is
determined by dividing the total distance for the last car to travel across the intersection
by ...... the speed limit. Correspondingly, the vehicle flow constraints for each left of
right turn estimate the minimum green time by adopting the ......speed to divide the
combination of 10 vehicle length plus the curve length. The longer duration for each
movement, governed by the same green time, is proposed in the model. So, the vehicle
flow constraints are established as follows:

G1 >
G2 >
G3 >
G2 >

6 Data Collection
I extend my sincere gratitude to the Smart City Management Bhagalpur authority for
providing the raw data of the Tatarpur intersection which has been invaluable for the
research. They collected data by video cameras. Given the existing traffic signal patterns
at the intersection the number of vehicle were assigned designation specifying the lane in
which they arrived at the intersection, their intended direction of travel and the existing
cycle of G2 +G1 + G3 . The vehicle arrival rate of each movement can be ascertained by
dividing the total arrivals by the duration of data collection time. The statistical analysis
shows that the vehicle arrival rates are: NR = 0.1413 veh/sec, SR = 0.0429 veh/sec and
TR = 0.1020 veh/sec.

7 Linear Programming Code

The study utilized linear programming and used pyomo as a diagnostic tool. The author
acknowledge that stochastic simulation process would have been more accurate since
flucturation of vehicle arrival rates are neglected when linear programming is adopted as
the study tool. However, the study opts to scope merely on the two hour P.M peak in
which the vehicle arrivals rates at the intersection are perceived to be consistently high
therefore the fluctuations were assumed to have less impact and negligible in this study
since the short time interval of similar traffic demand was segmented [13]. The full .........
programming code used which includes the objective function , constraints for vehicle flow
rates, constraints for queue lengths and crosswalk countdown time constraints is presented
in the fig [code fig. ].

8 Result
References
[1] A. Stevanovic, J. Stevanovic, K. Zhang, and S. Batterman, Optimizing traffic control
to reduce fuel consumption and vehicular emissions: Integrated approach with vissim,

Proceedings of the Samahang Pisika ng Pilipinas

th
39 Samahang Pisika ng Pilipinas Physics Conference
20–22 October 2021
5
 cmem, and visgaost, Transportation Research Record 2128, 105 (2009).
[2] F. Mannering et.al, Principles of Highway Engineering and Traffic Analysis, 6th
Edition, John Wiley Sons, Inc., Hoboken, NJ, (2016).
[3] A. Y. Zakariya and S. I. Rabia, Estimating the minimum delay optimal cycle length
based on a time-dependent delay formula, Alexandria Engineering Journal 55, 2509
(2016).
[4] X. Li, G. Li, S.-S. Pang, X. Yang, and J. Tian, Signal timing of intersections using
integrated optimization of traffic quality, emissions and fuel consumption: a note,
Transportation Research Part D: Transport and Environment 9, 401 (2004).
[5] Khushi, Smart control of traffic light system using image processing, in 2017 In-
ternational Conference on Current Trends in Computer, Electrical, Electronics and
Communication (CTCEEC) (IEEE, 2017), 99–103.
[6] A. Vogel, I. Oremović, R. Šimić, and E. Ivanjko, Improving traffic light control by
means of fuzzy logic, in 2018 international symposium ELMAR (IEEE, 2018), 51–56.
[7] Y. Suhweil, M. Al Yaman, et al., Smart controlling for traffic light time, in 2017 IEEE
Jordan Conference on Applied Electrical Engineering and Computing Technologies
(AEECT) (IEEE, 2017), 1–5.
[8] R. Soman and K. Radhakrishnan, Traffic light control and violation detection using
image processing, Traffic 8 (2018).
[9] A. Kanungo, A. Sharma, and C. Singla, Smart traffic lights switching and traffic
density calculation using video processing, in 2014 recent advances in Engineering
and computational sciences (RAECS) (IEEE, 2014), 1–6.
[10] . Subhdeep Chakraborty, Siddharth Srivastava. et.al, ”adaptive traffic light timer
controller”, IIT KANPUR, NERD MAGAZINE (2022).
[11] M. S. Shinde and S. Jagtap, Intelligent traffic management systems, A Review 2
(2016).
[12] C. Traffic, Manual on uniform traffic control devices, US Department of Transporta-
tion, Federal Highway Administation (2009).
[13] L. Zhang, Y. Yin, and Y. Lou, Robust signal timing for arterials under day-to-day
demand variations, Transportation Research Record 2192, 156 (2010).

Proceedings of the Samahang Pisika ng Pilipinas

th
39 Samahang Pisika ng Pilipinas Physics Conference
20–22 October 2021
6