Article citation information:
Aydin, M.M.,
Potoglou, D., Cipcigan, L. Working principle and performance
evolution of camera-based intelligent signalized intersections: Samsun city, Türkiye
example. Scientific
Journal of Silesian University of Technology. Series Transport.
2023, 121, 5-17.
ISSN: 0209-3324. DOI: https://doi.org/10.20858/sjsutst.2023.121.1.
Metin Mutlu AYDIN[1], Dimitris POTOGLOU[2], Liana CIPCIGAN[3]
WORKING PRINCIPLE AND PERFORMANCE EVOLUTION OF CAMERA-BASED INTELLIGENT
SIGNALIZED INTERSECTIONS: SAMSUN CITY, TÜRKIYE EXAMPLE
Summary. In the
current literature, it is clearly seen that most of the traffic chaos is
generally observed at intersections of the urban roads in cities. On the other
hand, many current traffic studies and research prove that fixed-time
signalized intersections cannot have a good ability to control and manage
current traffic flow at signalized intersection legs. For this aim, intelligent
intersections were developed and started to be used in many cities all over the
world in the last decade. These new intelligent intersection systems suggest
dynamic signal times for all intersection legs by using real-time measured
traffic data. These systems generally use cameras or loop detectors, which are
located in the proper places on a signalized intersection leg and record
vehicle movements. Within the scope of this study, a performance
comparison was made for before and after the camera-based intelligent intersection
applications at three isolated pilot signalized intersections within the scope
of the "Smart City Traffic Safety" project, which is one of the
largest Intelligent Transportation System projects in Turkey. After the system
was activated, it was observed that the drivers had impatient behaviors in the
beginning and had difficulty getting used to these new systems. After the
signal cycle was regulated with the learning of artificial intelligence, it was
seen that the drivers had more patience and more observant behaviors. It was
also obtained from the analysis results that these new intelligent systems
resulted in an average 16% decrease in control delays and a 20% decrease in
vehicle speeds.
Keywords: intelligent
intersection, traffic chaos, control delay, dynamic signal timing, urban roads
1. INTRODUCTION
Intelligent signalized intersections (ISIs)
will play an important role in traffic management and the improvement of
delays, vehicle speeds and emissions by ensuring road safety at signalized intersections
[1]. These systems can also be named "smart signalized intersections,
traffic-actuated signalized intersections," and camera-based or
sensor-based ISIs. They can obtain data from the cameras or sensors, process
the data by taking advantage of low-atency, high-bandwidth communications,
detect and track objects, and provide intelligent feedback and input to the
main control and operating systems. These systems have a key role in smart
cities. The connection and collection among intersections, roads and
corresponding vehicle and pedestrian traffic fully define the real-time
dynamics of a smart city [2-4]. ISIs will be at the core of artificial
intelligence-powered traffic control and management systems for the future of
cities. The World Economic Forum 2019 reported that big cities will double the
number of vehicles up to 2040. This report led to the conclusion that traffic
related problems with road flow tend to increase, and smart systems are a
possible solution to these problems. They could be effective in the creation of
mechanisms that allow current cities to be “smarter” than they are
now. Thus, many countries have started to transform their cities into smart
cities. There has been a big interest among the national road authorities in
promoting research focused on transformation [5]. Especially, developing
intelligent traffic light systems has an important role in this transformation
because they have many important points for vehicle connection and traffic
chaos [6, 7].
The current number of traffic cameras in all
cities is estimated to be billions, and they could be a great help in
developing ISIs, since they would provide various information from the control
system, which would include the detection and counting of vehicles and
pedestrians in the different study areas. Thus, many studies have been
conducted to obtain signalized intersection traffic data and propose flexible
and dynamic traffic light times [8]. For this purpose, the first video image
processing system, called the Spatial Image Processing Traffic Flow Sensor, was
developed by [9] to detect traffic queue length on roads. Then, [10] developed
image processing systems to measure traffic parameters such as volume, speed,
vehicle length, and queue length. In another study, [11] started to use virtual
loops to measure various parameters in traffic flows. In these loops, the size
and position of each loop can be adjusted by users to collect proper volume,
speed, occupancy, and vehicle classification data. In their study, [8] also
proposed a video image processing system to determine the traffic signal cycle
failure by tracking the queue formation to increase the performance of the
current traffic system. In a similar vehicle tracking research [12], suggested
a new method to track vehicles by matching different regions with vehicles via
video recordings. The research aims to uncover vehicle attributes, such as the
geographical location, length, and speed, through the collection of images by a
properly calibrated and high-definition traffic camera. In a conducted study of
[13], average stop number and control delay of vehicles were tried to be
estimated by using the image analysis technique (IAT). The total control delay
was aimed at being obtained by adding all the delays of stopped vehicles in an
examined time interval.
Roadside vehicle and environment monitoring
systems for signalized intersections are constant platforms, which generally
consist of pole-mounted or standard cameras placed in high locations, and are
connected to a control center or a device [14, 15]. Today, camera systems are
extremely cheap, small and smarter than the previous [16]. In addition, the
increase in processor power as well as the emergence of a new generation of
embedded architectures which allow real-time site applications have spawned a
huge interest in camera-based detection and tracking systems in the cities [17].
Especially for signalized intersections, most of these camera-based traffic
systems require one or more traffic cameras to be mounted at highly elevated
locations. In the current site applications, single-camera-based traffic
systems are mostly preferred to monitor signalized intersections. Most of them
work on multiple traffic camera-based systems, and each camera in the system
works independently, and then they perform a high-level fusion for the
observation and data collection from the examined intersections [18, 19]. In
addition, to get more accurate results from the systems, important
pre-processing steps of the systems such as the calibration of cameras, are
necessary before operating them.
The latest developments and studies on
intelligent transportation systems (ITS) clearly show that the utilization of
IT systems is of great importance to making cities “smarter”. Thus,
many cities around the world have started to apply to these systems to get
smart city titles, give good service to their citizens, and save humans and the
world. In its project, Samsun Metropolitan Municipality (SMM) has also started
to apply one of the biggest ITS project in Turkey. Intelligent Signalized
Intersections (ISIs) are the most important part of this project. In this
study, used ISI systems and the transformation process are introduced in detail
with all their properties. Then, the completed six ISIs performance evaluation
analysis results are shared. Results indicate that ISIs have a great effect on
the transformation of cities into smarter cities, and they have a positive
effect on reducing traffic chaos, controlling delays, vehicle speed and
emissions at intersections, which have the highest complexity in urban road
networks.
2. SMART CITY TRAFFIC SAFETY (SCTS)
PROJECT
“Smart City Traffic Safety
Project” is a candidate project to become the biggest Intelligent
Transportation System (ITS) project in Turkey. It is conducted with the
collaboration of Samsun Metropolitan Municipality (SMM) and ASELSAN for 2021
summer. The project aims to transform 72 fixed-time or non-signalized
intersections into intelligent signalized intersections, average speed
detection system in main corridors, parking violation detection system in
roadside parking areas, red light violation detection system in sections with
signalized lights, and initially 20 electric buses into public transportation
hubs. Thus, the SCTS project can be named one of the biggest ITS project in
Turkey.
3. INTELLIGENT SIGNALIZED
INTERSECTIONS (ISI) APPLICATIONS
As a part of the implementation of
the project, 72 ISI will be implemented to manage traffic and reduce air
pollution throughout the city by decreasing waiting times and delays at
signalized intersections. The locations of all intelligent signalized intersections
are given in Fig. 1. Geometric arrangements, infrastructure, and installation
processes for all digital equipment started in August 2021, and are planned to
finish in August 2023. Real site before-and-after transformation and
construction images for some ISIs are also given in Fig. 2.
Fig. 1. Locations of 72 intelligent
signalized intersections at urban roads in Samsun/Türkiye
After Before
(a)
(b)
(c) (d)
Fig. 2. Real site images for (a-b)
before and after transformation
and (c-d) construction duration
All work
packages and applied steps for the transformation of all these intersections
can be summarized as given in the flowchart below (Fig. 3).
Fig. 3. Work packages and steps for the
transformation of 78 intersections
3.1. Definition of System Properties
and Working Principles
Intelligent signalized intersection
(ISI) systems with cameras are an instant traffic control and operating system.
It includes computer-based vision solutions. This system mainly uses a single
fish-eye camera. With the help of the locations of cameras, this system can
obtain various traffic-related data, such as entry and exit directions of
vehicles, vehicle classes, average speed, etc. After the data collection
process, the system process obtained data, and it controls the signalized
intersection instantly and regularly in real time. The ISI system with traffic
cameras manages the total cycle time of the signalized intersections by
regulating green durations at intersection legs in an adaptive manner. It
includes a fully adaptive analysis method which helps optimize traffic control
in real time. Thus, camera-based intelligent signalized systems ensure high
contributions to the economy via reducing fuel consumption of vehicles; to
human health by reducing gas emissions; and to driver psychology by reducing vehicle
delays and total travel time in traffic. ISI systems also include an
intelligent traffic controller (ITC) device, an image processing device (IPD),
fish-eye camera, a camera pole (the height of the unit may show changes), and
analysis software.
4. METHODOLOGY
4.1. Definition of System Properties and
Working Principles
To evaluate the performance of
camera-based intelligent signalized intersections at urban roads of Samsun
city, total of six new ISIs were obtained from the reporting software. The
locations of the examined ISIs are shown in Fig. 5.
As mentioned before, almost all intersections
underwent geometric modification during the project. The previous geometric
properties and new intersection design features of all these examined
intersections are also seen in Fig. 6. The previous signalized intersection
types and the new types of examined ISIs are also given in Table 1.
Tab. 1
Previous
and new types of examined intersections
|
Inter.
No |
Previous
Type |
Current
Type (ISI) |
Average
Daily Traffic (veh/day) * |
|
1 |
4-leg
signalized traffic circle |
4-leg signalized |
18567 |
|
2 |
4-leg
signalized traffic circle |
4-leg signalized |
29645 |
|
3 |
4-leg signalized |
4-leg signalized |
30565 |
|
4 |
4-leg signalized traffic circle |
4-leg signalized roundabout |
55054 |
|
5 |
4-leg signalized traffic circle |
4-leg signalized |
43575 |
|
6 |
3-leg unsignalized roundabout |
4-leg signalized roundabout |
35972 |
* Data
belongs to 2019-year real site measurements.
In the study, all vehicle data was
obtained by counting the properties of the system software using the image
processing method. All vehicle numbers were determined by the counting of
fish-eye cameras’ real-time recordings. During the counting process,
vehicle types are classified in 4 groups: passenger cars, minibuses,
buses/midibuses, and trucks/lorries by the system software using the image processing
method.
(a) (b) (c)
(d) (e) (f)
(g)
Fig. 4. Real site images from the
ISI system (a) ITC, (b) IPU, (c) Camera pole and Fish-eye camera, (d) Editing
of virtual loops, (e) Triggers on intersection entering lanes, (f) Triggers
inside Intersection and (g) Reporting of obtained data and findings [20]
Fig. 5. Locations of examined
6 ISIs on urban roads of Samsun city
(a)
(b) (c)
(d) (e) (f)
Fig. 6. Previous geometric
properties and new intersections design features of
examined 6 ISIs
4.2. Definition of System Properties and
Working Principles
As can be seen from the descriptive statistical
analysis in Table 2, the highest Average Daily Traffic (ADT) flow is obtained
in ISI-4 and the lass is observed in ISI-1. Table 2 also shows that the most
observed vehicle type was the passenger car and the least observed was the
truck/lorry as expected on urban roads.
Tab. 2
Vehicle and traffic flow statistics
for the examined 6ISIs
|
Intersect.
No |
Vehicle
Type |
|||||||
|
Passenger
Car |
Minibus |
Bus/Midibus |
Truck/Lorry |
|||||
|
ADT |
σ |
ADT |
σ |
ADT |
σ |
ADT |
σ |
|
|
1 |
15384 |
2022 |
2953 |
422 |
1139 |
117 |
1199 |
121 |
|
2 |
23599 |
1984 |
5066 |
428 |
1289 |
150 |
1093 |
187 |
|
3 |
22515 |
21557 |
7322 |
6748 |
1666 |
1749 |
1351 |
1448 |
|
4 |
45771 |
44299 |
7290 |
6795 |
3195 |
3824 |
1810 |
2333 |
|
5 |
22361 |
18618 |
12528 |
12528 |
8860 |
8611 |
1813 |
1793 |
|
6 |
22378 |
4055 |
7928 |
1308 |
5049 |
513 |
2454 |
467 |
ADT:
Average Daily Traffic, σ: Standard Deviation, unit: Veh/Day for ADT and σ.
The given results in Table 3 were
also calculated by the developed software in the system. From the analysis,
changes (decrease) in control delays, speeds and emissions were also calculated
by the system.
Tab. 3
Obtained
results after the application of new intelligent intersection systems
|
Intersection
No |
Decrease
in |
|||
|
Control
Delays (sec.) (%) |
Average
Speeds (km/h) (%) |
CO2 (gr) |
PM10 (gr) |
|
|
1 |
11 |
14 |
415 |
602 |
|
2 |
20 |
22 |
886 |
905 |
|
3 |
16 |
19 |
845 |
867 |
|
4 |
22 |
25 |
3499 |
5176 |
|
5 |
14 |
17 |
876 |
859 |
|
6 |
12 |
15 |
869 |
843 |
|
Average
( |
16 |
19 |
1232 |
1542 |
As can be seen from Table 3, changes
for control delays, average vehicle speeds and emission values show differences
at different ISIs. Traffic volume, driver characteristics and behaviors,
vehicle types, and intersection types can show the most important parameters in
these results. It can be seen from the table that the highest decrease in
control delays is observed at ISI-4 and the least at ISI-1. Similar results are
also obtained for average vehicle speeds in the same ISIs. These new
intelligent systems resulted in an average 16% decrease in control delays and a
19% decrease in average vehicle speeds. It can also be seen from Table 3 that
these new intersection management systems have an important role in reducing
emissions at signalized intersections. These systems reduced an average of 1232
gr CO2 and 1542 gr PM10 daily. It can be concluded that ISIs have a
great effect on reducing traffic chaos, control delays, vehicle speed, and
emissions.
5. CONCLUSIONS AND SUGGESTIONS
In the last decade, there has been a
considerable increase in the application of Intelligent Signalized
Intersections (ISIs). Performance of new ISIs plays a vital role in the safety
and quality of travel on arterial networks and urban roads. On the other hand,
the collection of intersection performance data, such as vehicle control delay
and queue length, is a time-consuming and labour-intensive task. In this paper,
new ISIs and the transformation of signalized intersections are introduced in
detail. Then, a performance evaluation of the pilot new ISI applications on
urban roads in Samsun City was made. Evaluation results indicated that these
new intelligent systems resulted in an average 16% decrease in control delays
and a 19% decrease in average vehicle speeds, and these results may vary
according to intersection, driver, and vehicle characteristics. It is also
determined from the results that these new intersection management systems have
an important role in reducing emissions, which has a vital impact on climate
change at signalized intersections. According to calculated results from the
examined 6 new ISIs, transformation to intelligent intersections reduces on
average 1232 gr CO2 and 1542 gr PM10 daily. Thus, it can be clearly
revealed that ISIs have a great effect on reducing traffic chaos, controlling
delays, vehicle speeds, and emissions, as well as making cities
“smarter”.
In this study, only completed 6 ISIs
performance evaluation were made. Rest 72 ISIs transformation process still
continues. It is thought that after the remaining systems are completed, there
will be significant reductions in traffic chaos, control delays, vehicle speed
and emissions throughout the city, the intersections and road networks in the
entire city will work as a whole and integrated. Thus, citizens will face less
complexity and emissions caused by city traffic.
This study was conducted under a
research project titled “i-gCar4ITS: Innovative and Green Carrier
Development for Intelligent Transportation System Applications” which was
supported by the British Council. The authors would like to thank the British
Council for this support. The authors also thanks Samsun Metropolitan
Municipality, Ondokuz Mayıs University and Cardiff University for their
partnerships and support.
References
1.
Yang
S., E. Bailey, Z. Yang, J. Ostrometzky, G. Zussman, I. Seskar, Z. Kostic. 2020.
“Cosmos smart intersection: Edge
compute and communications for bird's eye object tracking”. In: 2020 IEEE International Conference on
Pervasive Computing and Communications Workshops (PerCom Workshops): 1-7. IEEE.
2.
Karagiannis
G., O. Altintas, E. Ekici, G. Heijenk, B. Jarupan, K. Lin, T. Weil. 2011.
“Vehicular networking: A survey and tutorial on requirements,
architectures, challenges, standards and solutions”. IEEE
Communications Surveys & Tutorials 13(4): 584-616.
3.
Czech
Piotr. 2017. „Physically disabled pedestrians - road users in terms of
road accidents”. Advances in Intelligent Systems and Computing 505:
33-44. DOI: https://doi.org/10.1007/978-3-319-43991-4_4. Springer, Cham. ISBN:
978-3-319-43990-7; 978-3-319-43991-4. ISSN: 2194-5357. In: Sierpinski Grzegorz
(eds), Intelligent transport systems and travel behaviour, 13th Scientific and
Technical Conference „Transport Systems Theory and Practice”,
Katowice, Poland, September 19-21, 2016.
4.
Czech
Piotr. 2017. „Underage pedestrian road users in terms of road
accidents”. Advances in Intelligent Systems and Computing 505:
33-44. DOI: https://doi.org/10.1007/978-3-319-43991-4_4. Springer, Cham. ISBN:
978-3-319-43990-7; 978-3-319-43991-4. ISSN: 2194-5357. In: Sierpinski Grzegorz
(eds), Intelligent transport systems and travel behaviour, 13th Scientific and
Technical Conference „Transport Systems Theory and Practice”,
Katowice, Poland, September 19-21, 2016.
5.
Olaya-Quiñones
J.D., J.C. Perafan-Villota. 2021. “A smart algorithm for traffic lights
intersections control in developing countries”. In: IEEE Colombian Conference on Applications of
Computational Intelligence: 93-106. Springer, Cham.
6.
Mu H., L. Liu, X. Li. 2018. “Signal preemption control of emergency
vehicles based on timed colored petri nets”. Discrete Dynamics in
Nature and Society: 1-15.
7.
Wang
X.B., K. Yin, H. Liu. 2018. “Vehicle actuated signal performance under general traffic at an
isolated intersection”. Transportation Research Part C: Emerging Technologies
95: 582-598.
8.
Zheng
J., Y. Wang, N.L. Nihan, M.E. Hallenbeck. 2006. “Detecting cycle failures at signalized intersections using video
image processing”. Computer‐Aided Civil and Infrastructure Engineering 21(6): 425-435.
9.
Higashikubo
M., T. Hinenova, K. Takeuchi.
1998. “Traffic queue length
measurement using an image processing sensor”. Sumitomo Electric
Technical Review 43: 64-68.
10. Fathy M., M.Y. Siyal. 1995.
“Real-time image processing approach to measure traffic queue
parameters”. IEEE Proceedings: Vision, Image, and Signal Processing
142(5): 297-303.
11. Yin Z., Y. Fan, H. Liu, B. Ran.
2004. “Using image sensors to
measure real-time traffic flow parameters, Preprint CD-ROM”. In: 83rd Transportation Research Board Annual
Meeting. January 2004. Washington, DC, USA.
12. Gupte S., O. Masoud, F.K.R. Martin,
N.P. Papanikolopoulos. 2002. “Detection and classification of
vehicles”. IEEE Transactions on Intelligent Transportation Systems
3(1): 37-47.
13. Saito M., J. Walker, A. Zundel.
2001. “Using image analysis to estimate average stopped delays per
vehicle at signalized intersections”. Transportation Research Record:
Journal of the Transportation Research Board 1776: 106-113.
14. Saunier N., T. Sayed. 2007.
“Automated analysis of road safety with video data”. Transportation
Research Record 2019(1): 57-64.
15. Messelodi S., C.M. Modena, M. Zanin.
2005. “A computer
vision system for the detection and classification of vehicles at urban road
intersections”. Pattern Analysis and Applications 8(1): 17-31.
16. Rinner B., W. Wolf. 2008. “An
introduction to distributed smart cameras”. Proceedings of the IEEE
96(10): 1565-1575.
17. Shi Y., F.D. Real. 2009. “Smart cameras: Fundamentals and
classification”. In: Smart
cameras: 19-34. Springer, Boston, MA.
18. Hu Z., C. Wang, K. Uchimura. 2007. “3D
vehicle extraction and tracking from multiple viewpoints for traffic monitoring
by using probability fusion map”. In: 2007 IEEE Intelligent Transportation Systems Conference: 30-35.
IEEE.
19. Subedi S., H. Tang. 2018.
“Development of a multiple-camera 3D vehicle tracking system for traffic
data collection at intersections”. IET Intelligent Transport Systems
13(4): 614-621.
Received 25.06.2023; accepted in
revised form 14.08.2023
Scientific Journal of Silesian University of Technology. Series
Transport is licensed under a Creative Commons Attribution 4.0
International License
[1] Faculty of Engineering, Ondokuz Mayıs
University, Kurupelit Kampüsü, 55217 Samsun, Turkey. Email: metinmutluaydin@gmail.com.
ORCID: https://orcid.org/0000-0001-9470-716X
[2] School of Geography and Planning, Cardiff University, CF10
3WA, Cardiff, United Kingdom. Email: potogloud@cardiff.ac.uk. ORCID: https://orcid.org/0000-0003-3060-7674
[3] School of Engineering, Cardiff University, CF24 3AA,
Cardiff, United Kingdom. Email: cipciganlm@cardiff.ac.uk. ORCID: https://orcid.org/0000-0002-5015-3334