Implementation of Artificial Intelligence in Traffic Engineering Management.
A Short Review Bahri Sangadji# # National Road Implementation Center.
North Maluku AbstractAi Several technologies, including fuzzy logic.
Q-learning, neural networks, the internet of things, and evolutionary algorithms, are commonly used to solve this topic in an effort to find answers to traffic problems in metropolitan areas.
Furthermore, these techniques highlight the optimization function, resulting in a traffic signal control design that may optimize the number of cars in a system, with implications for minimizing waiting times and the volume of vehicles in waiting situations.
The likelihood of obtaining a reduction in the amount of noise and pollution created by motorized vehicles is the next impact of this.
Based on these considerations, a number of previously described ways may be classed as one strategy in producing a design or design that is directed to the built environment and in accordance with sustainable development goals.
This research is an approach to reviewing the processes that may be established artificially in the traffic system so that it can adapt to the situations around it in real time.
KeywordsAi Put your keywords here, keywords are separated by semi colon.
INTRODUCTION
Increased urban traffic has a significant influence on road traffic congestion, and hence on the time it takes road users to reach their destination.
Expanding the road and expanding its capacity will not enough since the intersection will become Congestion is unavoidable, yet by regulating the intersection, there is still space for improvement.
The primary goal of traffic regulation in metropolitan settings is to ensure the safety of road users.
Furthermore, for the convenience aspect, various factors such as delays, traffic jams, pollution, noise, and the effect of enhancing service delivery to traffic are taken into account.
The major criteria of the designed control system are that the signal not enable unclear traffic movement and that it be open about how and when the displayed signal is updated.
Furthermore, two additional aspects of regulating the control system are addressed: judgments concerning the order of signal indications so that the system may be successfully optimized and the construction of control logic for signal production.
Fixed time and traffic-response are the two most popular techniques to dealing with traffic control signals.
Fixed-time, on the other hand, lacks the adaptability to traffic demands.
As a result, the goal of this work is to conduct a literature review on artificial intelligence techniques to operating traffic light systems.
Approaches Related to Artificial Intelligence Artificial intelligence and machine learning have shown to be excellent methods for improving the effectiveness of traffic control systems .
In terms of enhancing traffic system performance, three trustworthy methodologies are QLearning.
Neural Networks, and Fuzzy Logic Systems .
Q-Learning The concept behind this strategy is to observe movement at each junction so that the environment may interact with it in order to develop the best control mechanisms for lowering the duration of waiting lineups for both automobiles and pedestrians .
Furthermore, the outcomes of these observations are transmitted with other intersections via a distributed network in order to produce a global optimal timetable for the whole system .
Furthermore, the current traffic management regulations are included into the learning algorithm in order to give solutions while testing in the actual condition stage .
Furthermore.
Zeng and Zhang .
published a work in 2018 that built a form of a deep reinforcement learning agent that uses real-time GPS data to learn how to regulate traffic lights at isolated crossroads .
They constructed it using a mix of Recurrent Neural Network and Deep Q-Network, dubbed DRQN .
, and compared its performance on the partially observed traffic scope to that of the Deep Q-Network (DQN) standard .
II.
METHODOLOGY
This study employs a literature review to examine past studies on the advancement of artificial intelligence in traffic Furthermore, conclusions may be taken about the procedures that can be put in place to create the quantity of traffic volume.
RESULT AND DISCUSSION
Another effective transportation system has evolved by combining deep learning techniques with an approach function to train an adaptive traffic signal controller .
The n-step asynchronous Q-Learning algorithm is utilized in the learning stage, together with two hidden layers of the artificial neural network .
This method can minimize the average overall delay by 40% .
Based on these considerations, the optimization method based on Q-Learning is a type of adaptive artificial intelligence used to improve traffic system performance .
Neural Network This technique is used to address the situation of growing vehicle traffic in metropolitan areas, which causes substantial traffic congestion .
The goal of this strategy is to forecast traffic .
, which is a critical component of effective traffic management .
Furthermore, the modeling developed by this artificial neural network can solve difficult issues and non-linear scenarios .
The model is then utilized to compare against a traffic data set collected during a monitoring operation .
Furthermore, because of the intricacy of the challenge in optimizing traffic management systems.
Cellular Neural Networks (CNN) are being used as a solution to this problem .
The key benefit of this mechanism is that it is very quick at solving complicated and even non-linear problems.
CNN, too, is hardware platform agnostic .
Another benefit is that the modeler may create a full-scale model that takes non-linear elements and complexity into Finally, the CNN strategy in this context is a coordinated area traffic control with several junctions, both urban and highway .
Each lane has sensors fitted to identify things 200 meters away from the light signal.
To support traffic signals in conjunction with an IoT strategy .
RFID (RadioFrequency Identificatio.
devices such as RFID tags and RFID sticker readers, parking area sensors, and video systems are employed .
As a result, it is envisaged that the sensor object observes the vehicle in both tiny and large-scale movement intensities, as well as while it is halted or waiting for a traffic light .
The studied position or entity is then sent to a local smart server (LSS), which has more processors and storage capacity than the sensor .
The procedure is carried out through the closest node.
In order to achieve this aim.
LSS-based sensors provide real-time context aware computing and flow processing over Wi-Fi .
Genetic Algorithm The Genetic Algorithm is commonly used to solve traffic signal system difficulties .
The method that is often employed in traffic has various issues, such as extended wait periods, which have an influence on the number of cars accumulated .
Furthermore, one type of algorithm that can be used for optimization is the Genetic Algorithm .
This system employs a strategy based on evolutionary natural and biological selection theory, which is accomplished in multiple stages, namely:
Castro et al.
have offered a fictitious conditional network with adaptive biological behavior in accepting system circumstances and the ability to adjust control behavior .
Fuzzy Logic This method can overcome language and traffic data challenges that are difficult to forecast in order to regulate traffic .
Time, day, season, weather, and unusual conditions are some of the characteristics considered in estimating traffic behavior.
If these issues are not thoroughly explored, they might cause delays .
Furthermore, another technique to reduce traffic delays is to use fuzzy control to dynamically determine the maximum green time.
The fundamental benefit of this method is its ability to respond to real-time traffic conditions at isolated junctions, resulting in the construction of an efficient traffic management system .
The application of fuzzy logic is also considered to be an effective solution to traffic congestion problems, as well as a means of determining the reasons of traffic bottlenecks .
IV.
CONCLUSIONS
Finally.
Fuzzy Logic.
Q-learning.
Neural Network.
Internet of Things, and Genetic Algorithm are mechanisms that may be employed in traffic signal optimization to generate the number of cars in an intersection in order to decrease waiting times and vehicle volume build-up.
This research examines the approaches that have been employed to improve the function of traffic signals in combating vehicle volume REFERENCES