With the advent of deep learning systems, the potential applications of machine learning and artificial intelligence have expanded to include image interpretation. Specifically, AI-driven image recognition has proven to be a valuable tool in emergency situations. When a disaster occurs, both rapid response times and accurate information are crucial. Image recognition technology helps quickly identify the type of disaster and determine an effective intervention strategy, thereby facilitating rescue efforts and minimizing damage. In this article, the authors analyze two scientific literature databases on the topic, using specific keywords and the VOSviewer software. While the scientific community largely agrees that image recognition technology is essential for providing accurate and rapid responses to various hazards, the authors argue that current technological advancements can be harnessed synergistically to address all disasters that threaten public safety.

Urban evacuation management is very complicated due to infrastructural limitations, the unpredictability of crowd behavior, and the need for speedy decisions. The present paper discusses some of the challenges involved and points at the potential of some emerging technologies, such as Artificial Intelligence (AI), Internet of Things (IoT), and Digital Twins (DW) for improving the processes of evacuation. AI allows behavioral predictions and optimization of traffic flow, IoT supplies real-time environmental data, while Digital Twins enable the simulation and analysis of different scenarios. Yet, it is followed by ethical concerns of data privacy, social challenges like unequal access, and technical limitations involving high costs and model accuracy. This study emphasizes that more research needs to be developed along the lines of developing more advanced AI models and assessment of social and ethical implications from the use of these tools. This research establishes that one of the most relevant variables in developing better preparedness for urban evacuations, ensuring safety, and building resilience in cities is their use of integrated emerging technologies and AI.

Digital twins are the new way of taking emergency management one step further by simulating and analyzing in real time the dynamics of evacuations in cities. Advanced systems create a precise digital twin of physical environments that integrate diverse data streams and predictive algorithms to model human behavior and adapt to rapidly evolving situations during crises. This means a huge improvement compared to traditional approaches in evacuation planning, which generally have serious lacks in flexibility and accuracy in such complex urban environments. The concept is important since rapid urbanization coupled with increasing population density and higher frequencies of natural and artificial disasters can pose a serious challenge. Effective evacuation strategies are of prime importance to reduce casualties. Digital twins have the potential to change such practices in enhancing situational awareness, predicting bottlenecks, and optimizing response times. However, the implementation challenges include data accuracy, privacy issues, and socio-behavioral complexities in modeling. By overcoming those barriers through collaborative interdisciplinary approaches, robust data infrastructure, and standardized frameworks, digital twins can prove to be really transformational for emergency management and provide an approach that is more robust and adaptive against crises for the cities.

The flood crisis management in Rende, which is located in southern Italy, especially with its proximity to the Crati River, requires innovative solutions for rapid and efficient emergency evacuation. This study presents the potential of intelligent response technologies to maximize emergency response, minimize casualties, and enhance evacuation planning. Seven various evacuation scenarios are examined, ranging from the latest technologies such as AI-driven early warning systems, real-time mapping, self-driving automobiles, intelligent transportation, and drone-assisted rescue missions. We evaluate all these scenarios based on their execution speed, safety, people coverage, cost-effectiveness, tech feasibility, flexibility, and impact on infrastructure. Utilizing the fuzzy hierarchical analysis process (FAHP), this research systematically ranks these options to determine the most effective flood response strategies. The research identifies AI-driven early warning systems and preemptive evacuation plans as the most cost-effective, fastest, and risk-reduction measures. Smart transport networks coupled with autonomous vehicles can also increase evacuation effectiveness by a significant percentage. The report emphasizes connecting digital solutions and traditional emergency response systems with the aim of making Rende a more flood-resilient town. With these innovative technologies, Rende can become a model for Italy's flood crisis management, demonstrating how smart city infrastructure and AI-informed decision-making can transform disaster response. This research provides policy-relevant findings for policymakers, emergency planners, and city planners who want to implement technology-based crisis management in flood-prone regions.

Abstract Drivers’ behavior is one of the most important factors affecting road transportation safety. In particular, studying this issue in relation to young people aged 25 and below becomes more sensitive because they are not experienced and there are some age-related elements. Moreover, a significant percentage of beginner drivers fall into this age category, which can lead to risky behavior. Overconfidence, indiscipline, careless driving, or speeding tendencies may contribute greatly to their vulnerability to hazards on the roads. Hence, there is a need for further research to establish the constraints and possible risks involved with young drivers to improve road safety. Hence, this study aims to analyze the potential hazards associated with youth driving behavior in order to facilitate the development of relevant remedies through a thorough understanding of their behavior for safe transportation on the roads. To achieve this goal, a multi-criteria decision-making approach has been used. The proposed approach uses measurement of options and ranking based on the compromise solution method in an intuitive fuzzy environment to evaluate and rank risks. In addition, through consultation with experts and experienced technicians, a selection of 17 potential hazards were identified from existing risk factors. These risks are classified into three groups: working on the phone, distractions, and non-compliance. The present stud shows that risky driving and driving in reverse represent the highest level of risk, while speeding represents the lowest level of risk among young drivers.

The critical examination of driver behavior within road transport emerges as a cornerstone for augmenting fuel efficiency and achieving environmental sustainability. This study delves into the transformative potential of artificial intelligence (AI) in scrutinizing and refining driver behavior to optimize fuel consumption. Leveraging AI's advanced learning capabilities, the study reveals how dynamic analysis of driving patterns and real-time feedback can lead to significant reductions in fuel usage. By processing complex data such as acceleration, braking habits, and idle durations through sophisticated machine learning models, AI systems can identify inefficiencies and suggest behavioral adjustments to drivers. The article explores the multifaceted challenges inherent in the adoption of AI for this purpose, including data privacy concerns, the dependability of AI under varying traffic and weather conditions, and the human factor. Concurrently, it shows the expansive opportunities AI presents, such as enhanced predictive accuracy and the capacity for personalized driver training systems. Through a discussion of these challenges and opportunities, the research underscores the role of AI in forging a path toward greener roads. This study gives a detailed look at some AI methods and how they can be used in the real world. Understanding the connection between driver behavior, AI, and fuel use in a deeper way can enable other researchers to develop new transportation solutions that are more environmentally friendly.

Finding the most important causes of traffic crashes has been a major focus of transportation safety research for a long time. Many unpredictable and uncertain variables heavily influence the severity of the crash. Among these variables, the number of vehicles involved in crashes (NVIC) is one of the most important factors. Consequently, this study aims to evaluate the NVIC through intelligent techniques. For this purpose, we utilized a combination of the Harmony Search (HS) algorithm and the Differential Evolution (DE) algorithm with the k-means algorithm as machine learning approaches to evaluate the NVIC through a clustering analysis. In this study, a dataset of 35 case studies was used. This comprehensive database has seven model inputs, including Daylight, Type of crashes, Weekday, Location 2nd level, Speed limit, Avg Speed, and Annual Average Daily Traffic. Subsequently, the clustering results were compared and assessed against the actual outcomes. The obtained results from the analysis demonstrated that while both techniques displayed acceptable performance, the HS-K average approach converged faster than the DE-K average approach. In addition, the results of both models showed that daylight and type of crash had a greater impact on the samples where at most one car was involved in the crash, while the other parameters had a greater impact on the samples where more than one car was involved in the crash.

The academic and technical sectors of road transportation both find analysis of road network performance to be an essential area of research. There has been an increase in recent years in the number of studies investigating the function and potential applications of machine learning (ML) in the analysis of performance on road networks. Although machine learning has been shown to be useful in assessing performance on road networks, there is still a dearth of in-depth quantitative and qualitative research on the topic. The key goal of this research is to assess machine learning's quantitative and qualitative applications in transportation engineering's study of driver behavior analysis. In this paper, we looked at the research that has been done on how to use machine learning (ML) to study driving behavior analysis (DBA) from 1995 to 2022. All of the reviewed studies in this research were taken from the Web of Science (WOS) platform and analyzed. The analysis and discussions in this study try to provide a broad view of the changes that have occurred in the development process of these studies to other researchers and can be useful for showing the opportunities and challenges confronted by researchers in the use of ML in the analysis of driver behavior.

Crashes and the resulting damage are major issues in most countries' road transportation. Several factors play a role in causing road crashes and have been identified by researchers. One of the most significant factors in causing road accidents is the role of humans as drivers. Human errors play a key role in road crashes. So, the main aim of this research is to find, study, and rank some of the most significant human factors so that, with a good understanding of how important they are, solutions can be found to reduce road crashes. For this purpose, after examining and identifying the most common human factors, 15 factors that can have the greatest impact on road accidents were selected. An evaluation model was also used to evaluate and rank these factors using a combination of the Fuzzy Delphi Method (FDM) and the Analytic Hierarchy Process (AHP), which is one of the best ways to make decisions. The study was conducted on urban roads in Cosenza, located in the southern region of Italy. The obtained results indicated that the Drug and alcohol use (A9) and the Gender (A2) had the greatest and least impact on causing road crashes, respectively. Finally, some solutions were proposed to improve road safety in current situation.

In measuring road safety, accident severity is a key concern. Crash severity prediction models inform researchers about the severity of a crash based on a variety of criteria. To date, an enormous amount of research has been done on crash severity, and several models have been suggested to forecast crash severity utilizing existing test data or simulated datasets created using regression or classification analysis. In this study, a new approach was developed to determine the level of road crash severity (LRCS) using a large amount of real-existing data (1627 cases) by applying machine learning methods to the roads of Calabria in southern Italy. This study has three main goals: building prediction models based on classification approaches with the highest accuracy; comparing the performance of two supervised learning methods, including artificial neural networks (ANN) and convolutional neural networks (CNN), for predicting the LRCS; as well as determining the role of each of the contributing parameters in the LRCS and presenting a ranking by performing a sensitivity analysis. Finally, based on the accuracy values, it has been found that there isn't a salient difference between the predicted models. But it should be noted that 68.4% accuracy for the testing dataset implies the CNN model was superior to the ANN model, which scored 61.74% accuracy. This is acceptable if the minor variation in modeling accuracy is desired. Also, the results of the sensitivity analysis showed that the number of vehicles and the road element had the highest and lowest effect rates on the LRCS, respectively.

Collection and analysis of traffic data are the most critical challenges in traffic control in transportation networks. The dramatic growth of new technologies such as tiny devices equipped with elements of computing, sensors, and actuators in the smart environment of the internet of things (IoT) has led to the collection of large volumes of data. Identifying and combining data obtained from various smart devices plays a key role in data processing. Hence, data fusion is required to collect and extract usable data from numerous sources. This study discusses the challenges and opportunities of using data fusion to estimate travel time under wireless sensor networks (WSN s) in the Intelligent Transportation Systems (ITS). For this purpose, the three main categories of data fusion, namely the probability-based method, the artificial intelligence-based method, and the evidence theory-based method, were considered. Consequently, discussing the advantages and drawbacks of data fusion techniques can give researchers a better perspective of travel time estimation with better accuracy.

Two new technologies are going to shape the future of traffic management and control: “connected” and “autonomous” vehicles. The possibility of gathering information from vehicles travelling on the road and the use of this information for better traffic management is receiving a growing attention from the scientific community. The purpose of this study is to present the architecture of a dedicated system for traffic signal real time regulation based on data coming from “connected” vehicles. The system presented in this paper has been implemented on the field in a dedicated experimental intersection site to assess performances of the system in actual use. Results of the experiments carried on with the presented system will establish the feasibility and the advantages of FCD adaptive traffic signals and will help to develop better regulation algorithms for this kind of new “connected” intersections.

Traffic microsimulation has been used extensively to evaluate consequences of different traffic planning and control policies in terms of travel time delays, queues, pollutant emissions and every other common measured performance while at the same time traffic safety has not been considered in common traffic microsimulation packages as a measure of performance for different traffic scenarios. The absence of safety evaluation with microscopic simulation models is due to a lack of complete and established models for the simulation of potential crashes. Commonly-used procedures for safety evaluation with microscopic simulation have been based on traffic conflict theory which until recently was not extended to single vehicle crashes and to crashes between vehicles which are moving on non-conflicting trajectories. This paper presents some applications of a new procedure based on potential crash events simulation for the evaluation of safety levels in microsimulated traffic scenarios which takes into account also potential crashes with road side objects and barriers.

This paper presents a simple concept which has not been, up to now, thoroughly explored in scientific research: the use of information coming from the network of Internet connected mobile devices (on vehicles) to regulate traffic light systems. Three large-scale changes are going to shape the future of transportation and could lead to the regulation of traffic signal system based on floating car data (FCD): (i) the implementation of Internet connected cars with global navigation satellite (GNSS) system receivers and the autonomous car revolution; (ii) the spreading of mobile cooperative Web 2.0 and the extension to connected vehicles; (iii) an increasing need for sustainability of transportation in terms of energy efficiency, traffic safety, and environmental issues. Up to now, the concept of floating car data (FCD) has only been extensively used to obtain traffic information and estimate traffic parameters. Traffic lights regulation based on FCD technology has not been fully researched since the implementation requires new ideas and algorithms. This paper intends to provide a seminal insight into the important issue of adaptive traffic light based on FCD by presenting ideas that can be useful to researchers and engineers in the long-term task of developing new algorithms and systems that may revolutionize the way traffic lights are regulated.