1. BAYERO UNIVERSITY, KANO
FACULTY OF ENGINEERING
DEPARTMENT OF CIVIL ENGINEERING
ADVANCED TRAFFIC ENGINEERING : CIV8331
ASSIGNMENT
ON
REVIEW OF MICROCSCOPIC TRAFFIC FLOW
MODELS USING ARTIFICIAL INTELLIGENCE
BY:
IBRAHIM HALLIRU
SPS/20/MCE/00021
TEL: +2348039291976
EMAIL: ibrahimhalliru91@gmail.com
SUPERVISED BY
ENGR.PROF.HM AL-HASSAN
2. INTRODUCTION
ARTIFICIAL INTELLIGENCE
Artificial intelligence elicits strong feelings. For starters, there's our
preoccupation with intelligence, which appears to grant humans a unique
status among living things. "What is intelligence?" "How can intelligence be
measured?" and "How does the brain work?" are some of the questions that
come up. When trying to comprehend artificial intelligence, all of these
questions are important. The primary question for engineers, particularly
computer scientists, is the intelligent machine that behaves intelligently like a
person (Mackie, 2017).
AI was first defined by John McCarthy in 1955 as "the purpose of AI is to
construct robots that behave as if they are intelligent." Consider the following
situation as a way to put this definition to the test. On a four-by-four-meter
square enclosed platform, fifteen or so small robotic vehicles are moving.
Different behavioral tendencies can be seen. Some cars travel in small
groupings. Others glide through the area with ease, avoiding any potential
collisions. Others seem to be led by a leader. Also, aggressive tendencies can
also be seen.
3. AI AND TRANSPORTATION
Engineers in the automobile industry have come a long way in the past 130
years. In the current world, it is expected that one out of every two persons
possesses an automobile. These vehicles are quite dependable. This makes us
extremely mobile, which we take use of in our work, daily lives, and leisure
activities. We are also reliant on it. We can no longer live without a car,
especially in rural places with inadequate public transportation. The next wave
of improved road transit convenience is now upon us. In a few years, we will
be able to purchase electric self-driving cars, also known as robotic cars that
will drive us to practically any location independently. The robotic car's
passengers would be able to read and work or sleep during the trip. This is
already feasible on public transportation, but passengers in a robotic car could
do so at any time and on any route.
4. Vehicles that are self-driving and capable of operating independently may be able
to travel without passengers. This will result in a new level of convenience:
robotic taxis. We will be able to order the ideal taxi, in terms of size and
equipment, for any possible transit reason using a smartphone app. We will have
the option of traveling alone in the taxi or sharing a ride with other passengers.
We will no longer require our own vehicle. All associated obligations and
expenses, such as refueling, technical service, cleaning, parking, buying and
selling, garage rent, and so on, are nullified, saving us money and time.
The level of passenger safety will be significantly higher than it is now. Experts
believe that future accident rates will be between zero and 10 percent lower
than they are now. Emotional driving (often known as "road rage"), inattentive
driving, and driving while under the influence of drugs or alcohol will all be
prohibited. The following are some of the benefits and drawbacks of AI in
transportation:
5. ADVANTAGES
Artificial intelligence (AI) offers
the benefits of permanence,
reliability, and cost-effectiveness
while also addressing uncertainty
and speed in solving problems or
making decisions.
When AI vehicles are
used, CO2 emissions
will be greatly
reduced.
AI aids cost-cutting by
reducing the amount
of time that employees
are required to work.
DISADVANTAGES
AI requires a large
data sets which takes
longer training period.
Public means of
transportation will
drastically be reduced
and or loss of jobs for
commercial/taxi
drivers.
AI vehicles (robotics
automobiles) must be
networked, which will
open the door to a
slew of hacker attacks.
6. SUITABILITY OF AI IN SOLVING TRANSPORTATION PROBLEMS
Traffic issues share a number of aspects that make them accessible to AI-based
solutions. To begin with, transportation problems typically require both
quantitative and qualitative data. Even though we frequently deal with
qualitative data in transportation, researchers frequently encounter systems
whose behavior is difficult to model using conventional methods, either because
the interactions among the various system components are not fully understood
or because the human component of the system introduces a great deal of
uncertainty. Building empirical models based on observable data is critical for
such complicated systems. Given its ability to approximate universal functions,
NNs are ideal instruments for creating such models.
7. AIM AND OBJECTIVES
• The aim of this review is to identify AI
strategies that can be used to solve
microscopic traffic models. The objectives
are;
• To comprehend artificial intelligence in
respect to traffic inputs in order to provide
meaningful outputs.
• To present research on artificial intelligence-
based traffic models.
8. STATEMENT OF PROBLEM
The continuing growth of urban motor traffic has increased traffic congestion in
most cities. Because travel demand outpaces road capacity, the problem will only
get worse unless substantial improvements in traffic management are
implemented. As a result of recent improvements, modeling and simulation are
becoming more important instruments for traffic control and future transportation
networks. Simulation is required not only as a planning tool to assess the benefits
of proposed next-generation transportation systems, but also as a tool to design
scenarios, optimize control, and predict network behavior from an operational
standpoint. In general, modeling and simulation provide engineers and researchers
with a complete picture of a hypothetical transportation system, as well as the
ability to assess present problems (and provide prospective solutions) swiftly and
effectively.
9. LITERATURE REVIEW
Research findings, over the years, have proven that traditional models
cannot handle a large volume of traffic data. One major theoretical issue
that has dominated the field of transportation for many years is that if a
large volume of traffic datasets is not handled (divided into inputs and
outputs) properly, this can decrease the accurate prediction of traffic flow
at road intersections or freeways. However, much uncertainty still exists in
the determination of which artificial intelligence methods effectively
resolve traffic issues, especially from the perspective of the traffic flow of
individual vehicles at the microscopic level.
In this regard, independent simulators are commonly used in transportation
research to answer concerns about human factors and safety. Driving
Simulators (DS) are a type of simulator that is used to track a driver's
behavior, performance, and attentiveness. Traffic Simulators (TS) are
10. another tool that may be used to better plan, build, and run transportation
networks. Microscopic, mesoscopic, and macroscopic traffic simulation models
are common. Individual vehicle speeds and locations are common measures
used in microscopic models to predict the state of individual cars.
Macroscopic models combine the description of traffic flow into a "larger
picture," and speed, flow, and density are frequently used as metrics.
Mesoscopic models combine macro- and microscopic elements (Boxill and Yu,
2000).
In traffic dynamics, different statistical and machine learning methods have
been used to solve microscopic flow parameters, such as Neural Network
(NN), Intelligent Driver Model (IDM), Fuzzy System (FS), and so on (Zhu, 2021).
Some of the models mentioned in this review include:
11. NEURAL NETWORKS (NN)
Neural networks are networks of nerve cells in the brains of humans and
animals. The human brain has about 100 billion nerve cells. We humans owe
our intelligence and our ability to learn various motor and intellectual
capabilities to the brain’s complex relays and adaptivity. For many centuries
biologists, psychologists, and doctors have tried to understand how the brain
functions. Around 1900 came the revolutionary realization that these tiny
physical building blocks of the brain, the nerve cells and their connections,
are responsible for awareness, associations, thoughts, consciousness, and the
ability to learn (Kyamakya, 2006).
A neural network model which is the branch of artificial intelligence is
generally referred to as artificial neural networks (ANNs). ANN teaches the
system to execute task, instead of programming computational system to do
definite tasks. To perform such tasks, Artificial Intelligence System (AI) is
generated. It is a practical model which can quickly and precisely find the
patterns buried in data that replicate useful knowledge.
12. Mathew & Ravishankar, (2012) create three distinct neural network
architectures with different inputs used to predict vehicle-type-dependent
following behaviors. To put it another way, the initial neural network
architecture has only two inputs: leader velocity and gap distance. The
other two neural networks are based on the first and include additional
inputs, such as vehicle types that lead and follow. They claim that the two
neural networks with vehicle type inputs were able to accurately gather
field velocity data.
Mirbaha et al., (2017) uses NN to analyze effective parameters at the
microscopic level between two deceleration and congestion phases.
Artificial neural networks results determine that increasing spacing
difference of the follower vehicle between two phases will result in
increasing the follower maneuverability based on under reaction pattern.
Since, the follower driver has no considering propagated wave and no
disregarding speed drop; this condition causes to reduce excessive safe
spacing in traffic oscillation. The follower driver performs faster responses,
reducing the time between the two phases, because of developing safe
spacing.
13. INTELLIGENT DRIVER MODEL
The Intelligent Driver Model (IDM), a microscopic car-following model is well-
known to reproduce all essential traffic dynamic phenomena observed on
freeways. The time-continuous Intelligent Driver Model (IDM) is probably the
simplest complete and accident-free model producing realistic acceleration
profiles and a reasonable behavior in essentially all single-lane traffic situations.
The IDM is an Adaptive Cruise Control (ACC) system intended for adjusting the
driver’s longitudinal desired velocity and safety time gap. The ACC is a new
technology which requires the use of sensors for detecting the velocity and
distance of the vehicle in front of a vehicle equipped with the ACC system and
enables ACC vehicle to adjust its velocity accordingly.
14. The IDM, developed by Kesting et al. (2008), is expressed by:
Where: an is the maximum acceleration of the vehicle n [m/s2];
vo
n the desired velocity of the vehicle n [m/s];
sn the distance gap [m]:
15. The IDM is derived from a list of basic assumptions, characterized by the following
requirements:
The acceleration fulfills the general conditions for a complete model.
The equilibrium bumper-to-bumper distance to the leading vehicle is not less than a
safe distance.
Braking strategy shows how vehicles are controlled when obstacles or red traffic
lights are approached.
Transitions between different driving modes (e.g., from the acceleration to the car-
following mode) are smooth.
The model should be as mean as possible. Each model parameter should describe
only one aspect of the driving behavior (which is favorable for model calibration).
Furthermore, the parameters should correspond to an intuitive interpretation and
assume reasonable values.
16. RELATED STUDIES
Olayode et al., (2021), develop an ANN-PSO capable of modelling and analyzing vehicular
traffic flow at a signalized road intersection. For the accurate modelling of traffic flow, it
was important to understand and identify some traffic flow parameters and have clarity on
how traffic flow affects the movement of vehicles at a signalized road intersection. In his
study, successful development of ANN-PSO model for the traffic flow of vehicles at a
signalized road intersection was achieved. The vehicular speed, the time of the day, the
traffic volume, and the traffic density of each class of vehicles (light vehicles, long trucks,
medium trucks, and short trucks) have been considered as the traffic flow inputs and
outputs variables.
Park et al., (2021), considered isolated intersection and coordinated intersection traffic
signal control models constructed using DQN, an emerging reinforcement learning
technology that offers a solution to complex and diverse problems. Iterative experiments
were performed using a microscopic traffic simulation in the scenarios constructed to
evaluate and verify the completed models.
17. CURRENT RESEARCH
Converting traffic sensors into intelligent agents that can detect and report
traffic accidents or predict traffic conditions are just a few examples of AI
technology in use today (Schleiffer, 2002). Researchers recently discovered
that AI outperforms several existing algorithms in assessing and anticipating
traffic conditions based on microscopic traffic data collected from vehicles on
their path, as envisioned in the vehicle–infrastructure integration or linked
vehicle program.
18. FUTURE RESEARCH
Further research is desirable to study sequential prediction. Not all traffic
incident-related factors can be acquired at the time when the incident is
reported, on the contrary, more detailed information is gradually acquired
over time. A more realistic prediction method that continuously updates
model variables and results over time will provide more accurate estimation
and reliable references.
According to studies on lane changing, Advanced Driver Assistance Systems
(ADAS) must be equipped with concepts and approaches that allow for the
prediction of future circumstances. An examination of driving behavior
prediction under real-time traffic conditions is very important to carry out
driving erroneous analysis at the microscopic level in order to optimize
warning and control strategy of driver aid systems and decide assistance
systems more precisely (Ding et al., 2013).
19. CONCLUSION
The main motivation for using AI models is the ability to learn and
incorporate the uncertainties from real driving data. This means, after
learning from the driving behavior data, the model could generate vehicles
states to reproduce any style of driving. In this review, it is seen how
microscopic parameters were used and modeled using AI by different
researchers.