The decision-making process for driving maneuvers is critical for the large-scale deployment of automated vehicles (AVs). This paper puts forward a deep reinforcement learning (DRL) framework, termed integrated convolutional-attention decision-making (ICADM), designed to address behavioral decision-making challenges in complex traffic scenarios. ICADM utilizes two complementary feature extraction pathways, namely convolutional neural network (CNN) and multi-head attention, each focusing on distinct aspects of data representation to enhance decision-making efficiency for AVs. To combine these features, ICADM employs a dynamic fusion method that models their correlations, aiming to reduce information redundancy in the state space and enhance the quality of the state representation. Operating within a continuous action space, ICADM enables more precise and adaptive driving maneuvers. Moreover, a unified reward function is introduced to optimize driving policies, balancing safety, efficiency, and passenger comfort. The simulation results validate the effectiveness of ICADM, demonstrating its superior performance compared to the baseline DRL method. ICADM improves driving safety by increasing the average time-to-collision (TTC) to 13.92 times that of the baseline, while also enhancing driving efficiency with a 2.17% increase in average speed and improving passenger comfort with a 15.32% reduction in the comfort index (CI). Comparative experiments and ablation studies further confirm the enhanced decision-making capability of ICADM in complex scenes.

The development of intelligent transportation systems (ITS) in urban areas currently faces a critical misalignment between technological applications and the developmental stages of cities. Most studies on ITS to date have focused on solutions tailored to developed cities, which overlook the practical needs of growing cities. The above problem reduces the applicability of advanced technologies in resource-constrained contexts. To fill the gap, our study constructs a five-tier ITS evaluation framework for cities evolving from city 1.0 to 5.0. Through the synthesis of the literature and practical case analysis, a three-view research paradigm is proposed, "issue-science-engineering" (ISE). It integrates three key elements of ITS in urban areas: the infrastructure network, the transportation network, and the management network, which identify five distinct evolutionary stages of ITS for cities. By pinpointing common critical challenges across these stages, the study distills four priority research themes of ITS for academia: data-driven collaborative optimization, system resilience under complex disruptions, human-centered equity assurance, and incremental transitions toward sustainability. Furthermore, the study establishes a matching mechanism between technological pathways and city developmental stages, which could offer actionable insights for collaboration between industry and academia. Our study provides a decision-making framework for the development of ITS in cities at different stages. We hope that the study could contribute to a more balanced global advancement in this field.

Autonomous vehicles (AVs) will operate alongside human-driven vehicles for an extended transition period, during which aggressive human driving may become a major source of risk. This study proposes an integrated safety-control framework that combines real-world-data-driven behavior modeling with deep reinforcement learning to design longitudinal AV control strategies for mixed traffic. Aggressive, general, and defensive driving patterns are calibrated from the CitySim dataset, and dynamic aggressiveness is incorporated into an improved car-following model. A proximal policy optimization algorithm with a Kullback–Leibler penalty is then used to learn multi-objective strategies balancing safety, efficiency, comfort, and fuel economy in freeway and signalized-intersection scenarios. The results show that the behavior-aware strategies exhibit different strengths across traffic environments. On the freeway, the defensive-threshold strategy maintains a larger time headway, reduces positive acceleration, and lowers system-level fuel consumption, whereas the default, aggressive, and general strategies preserve higher traffic efficiency. At the intersection, signal control narrows the differences among strategies and limits the influence of longitudinal threshold settings on most evaluated indicators. These findings provide a quantitative basis for selecting behavior-aware control thresholds and designing robust AV strategies for mixed-autonomy traffic containing aggressive human drivers.

To evaluate resilience strategies for food cold chains facing climate-intensified floods, we develop a predict-and-prescribe approach that integrates vulnerability prediction with policy simulation. The predictive component utilizes a probabilistic vulnerability graph attention network to rank facility vulnerability, outperforming benchmarks by 33.4% in accuracy. The prescription embeds these rankings within a flood-coupled cold chain agent-based model to simulate thermodynamic spoilage driven by flood-induced operational failures.Across flood intensities, prescription reveals an objective mismatch: protecting structurally central nodes yields 15% higher system benefit than targeting predicted high-loss sites. This gap is driven by an ‘inventory bomb’, in which 90.8% of losses concentrate at downstream retail nodes. In response, we identify a ‘layered defense’ hybrid strategy combining physical robustness, resource redundancy, and network flexibility as optimal.These findings establish a dual-tool policy framework: use AI for ex-ante risk provisioning, but prioritize structural leverage over predicted loss magnitude for emergency triage in catastrophic events.

Existing flood resilience assessments face a fundamental tradeoff: correlational analyses of real-world data suffer from confounding bias, while simulations of hypothetical scenarios lack empirical grounding. To address this identification gap, we propose a physically grounded causal inference framework that derives unbiased causal parameters directly from physical observation. First, we generate high-fidelity flood maps from synthetic aperture radar (SAR) imagery using a fine-tuned U-Net model adapted for dense urban environments. Second, we integrate these empirical maps into a microscopic traffic simulation model to quantify dynamic network disruptions. Third, we apply double machine learning (DML) to isolate the causal effect of flooding on transport resilience while rigorously controlling for high-dimensional urban form confounders. Applied to a flood event in a dense coastal megacity, our analysis reveals two key discoveries: a nonlinear tipping point around 25% flood intensity, distinct from recent studies assuming linear degradation, beyond which system performance collapses chaotically; and a "dual nature of vulnerability" where low-density suburbs suffer a fragility of isolation ( 2.9 & times; higher road closure rate), while dense urban cores exhibit a fragility of congestion (1.47 & times; greater waiting time increase). Validated across 130 sensitivity tests, this framework provides a robust, data-driven blueprint for spatially targeted resilience investments, demonstrating that causal identification in disaster contexts requires physically grounded observation, not synthetic assumptions.

Extreme rainfall and flood events increasingly threaten urban transportation networks, yet conventional resilience evaluations fail to capture the dynamic interplay between flooding, traffic congestion, and emergency service accessibility. To address the above gap, this study provides an integrated framework for transportation resilience evaluation with the proposed spatiotemporal accessibility analysis. Our approach combines real-time traffic simulation, flood modeling, an enhanced two-step floating catchment area (E2SFCA) method with a novel congestion-sensitive impedance function, and grid-based percolation theory to quantify network fragmentation via accessibility thresholds and largest connected component (LCC) metrics. A case study in Shanghai’s Huangpu District demonstrates that peak-hour floods induce greater accessibility losses and faster network fragmentation than off-peak events. Sensitivity analysis highlights accessibility’s higher responsiveness to congestion dynamics over facility siting, highlighting the importance of adaptive traffic management in disaster planning. The workflow provides a robust tool for planners to anticipate vulnerabilities, prioritize interventions, and enhance time-critical disaster preparedness.

Traffic network resilience refers to the ability of a traffic network to maintain a certain capacity and service level even when disturbed by external factors, as well as its capacity to recover following a disruptive event. This paper integrates traffic simulation with resilience analysis of urban road traffic networks and proposes a framework for identifying critical roads in urban road traffic networks from a resilience perspective. This framework is both theoretical and applicable to any unanticipated disruptive event. By incorporating four attribute indicators—traffic, topology, urban function, and socio-economic factors—the framework assesses the importance ranking of each road segment in an urban road traffic network both before and after an unanticipated disruptive event. A case study is conducted using a real urban road traffic network in Shanghai. From the perspective of policymakers, corresponding policy recommendations are made to enhance the resilience of urban road traffic networks against unanticipated disruptive events and to mitigate socio-economic losses.

The vehicle-infrastructure cooperative control system (VICCS) leverages autonomous driving technology and interactive communication between vehicles and infrastructure to maximize system-wide benefits. As this technology emerges, a thorough socio-economic evaluation is essential to substantiate its utility. Analyzing comparisons with traditional systems will assist in adopting this innovative technology. This paper quantifies the potential benefits of the VICCS through several steps: defines the application scenarios of VICCS, models the behavioral control of vehicles and traffic signals, simulates the system in mixed-autonomy traffic environments at signalized intersections, analyzes the operational performance and service levels of VICCS, and evaluates the costs and benefits for the private and public sectors. This study employs a technical framework for VICCS that integrates deep reinforcement learning (DRL) methods to optimize vehicle speed and dynamic traffic signal control. The DRL approach is crafted to forecast the system’s performance and level of intelligence in prospective settings more accurately. The findings reveal that the anticipated VICCS will confer considerable benefits, including enhanced safety, operational efficiency, and environmental sustainability, at a cost to be incurred compared to existing systems. This will result in an annual economic gain of at least CNY10,000 (the difference between the expenditure and the gain) for the private and public sectors. This paper provides policy recommendations to support the strategic deployment of VICCS, informing stakeholders of the practical implications and facilitating the traffic system’s integration into advanced mechanisms.

The hit-and-run caused a delay in medical assistance to the victim and posed a significant threat to the safety of drivers in road tunnels. This study investigates the potential factors contributing to drivers’ hit-and-run violations in river-crossing tunnels. This paper built three models (the logit model, the random parameter logit model, and the random parameter logit model with heterogeneity in means) based on a dataset consisting of crashes reported in thirteen river-crossing tunnels in Shanghai, China. Potential contributors from five aspects (offending drivers, vehicle conditions, tunnel characteristics, environmental conditions, and crash information) were explored. Results showed that the random parameter logit model with heterogeneity in means produced the highest fitting accuracy among the three models. Eight important variables (nighttime, single-vehicle, multi-vehicle, two-wheeled vehicle, passenger car, heavy goods vehicle, rear-end, and short tunnel) were found to affect hit-and-run violations significantly. The research has highlighted that nighttime and short tunnel increase the likelihood of hit-and-run and other variables are the opposite. The results of this study could provide useful information for the development of interventions to improve the level of safety in tunnels and reduce the rate of hit-and-run offenses.

IntroductionAs urbanization progresses and vulnerable populations increase, equitable accessibility remains a critical issue. This study evaluates the accessibility of transit-oriented development (TOD) in Shanghai, focusing on barrier-free facilities in metro stations.MethodsA comprehensive evaluation framework combining the Analytic Hierarchy Process (AHP) and the System Usability Scale (SUS) was developed to assess metro station accessibility. Thirteen evaluation factors formed a composite accessibility index. A case study of two Shanghai metro stations, Xinzhuang and Xujiahui, was conducted using quantitative metrics, surveys, and interviews.ResultsA strong correlation between AHP scores and SUS ratings validated the framework’s reliability. The study provides recommendations for enhancing metro accessibility.DiscussionThe proposed framework offers a robust tool for evaluating and improving urban transit accessibility, with implications for inclusive mobility policy and design.

Transit-oriented development (TOD) is a leading urban planning paradigm whose success largely depends on the effectiveness of its guidance systems. Although guidance systems have been extensively researched in various contexts, there is a significant gap in evaluating these systems within TOD scenarios, particularly regarding quantitative assessment methods. This study introduces a novel quantitative evaluation methodology to assess the various dimensions of guidance systems in TOD contexts. It identifies four primary criteria: normativity, effectiveness, continuity, and individualization, alongside 13 specific secondary evaluation indicators. The contributions of these indicators are determined through expert scoring using the analytic hierarchy process (AHP) and are subsequently validated through on-site user assessments. Data collection for TOD guidance systems involved scoring using a system usability scale (SUS) questionnaire and objective data measurements across five routes within two representative TODs in Shanghai. The results reveal a strong linear regression relationship between AHP quantitative scores and SUS questionnaire scores, validating the effectiveness of the proposed indicator framework. In summary, this study addresses the gap in evaluating TOD guidance systems and provides a comprehensive evaluation system to guide future design and development efforts in TOD scenarios.

Transit-oriented development (TOD) strategies on subway stations have been implemented in many high-density cities globally to enhance public transportation system efficiency and promote public transportation mobility. Focusing on the developments of intricate metropolitan systems, researchers attempted to elicit “latent rules” by proposing a generic TOD performance evaluation system. This study suggests a multi-indicator TOD performance evaluation method based on a multi-indicator approach grounded in the analysis of multisource urban big data, revealing the role of rail transit TOD station characteristics on critical indicators of station operation through an interpretable machine learning approach. Using Shanghai, China, as a case study, the methodology employed 26 widely used indicators related to TOD development and utilized a BP neural network model trained in a sample space of 77 rail transit TOD stations, aiming to predict the four critical station performance indicators. The robustness of the explanatory variables in the model has been verified by various methods, affirming their consistencies with the development characteristics of the city and the stations. The performance assessment methodology achieves significant predictive results and is computationally feasible, with potential values in applications in other high-density cities worldwide.

Numerous researchers have endeavored to amalgamate critical transit-oriented development (TOD) indicators, such as development density, walkability, and diversity, into a single TOD index to assess TOD performance. However, implementing TOD in megacities necessitates a more comprehensive selection of indicators, an objective calculation methodology, and accessible calculation data for the TOD index. This study introduces a method based on multi-indicator TOD performance assessment using multi-source urban big data; it uses Shanghai as a case study to evaluate and analyze the impact of site characteristics on performance. The method constructs the Comprehensive Socio-Economic Development Index (CSEDI) based on four indicators of TOD site operations. It establishes a multivariate regression model utilizing principal component analysis to extract 22 leading component indicators as independent variables from 71 indicators associated with TOD. Within the sample space of 77 rail transit TOD sites in Shanghai, the CSEDI exhibited a robust correlation with the independent variables. The evaluation results of the case study demonstrate consistency with the development characteristics of the city and the sites, indicating that the evaluation method can guide the renovation of existing sites and the development of new sites.

Rainstorms and flooding are among the most common natural disasters, which have a number of impacts on the transport system. This reality highlights the importance of understanding resilience—the ability of a system to resist disruptions and quickly recover to operational status after damage. However, current resilience assessments often overlook transport network functions and lack dynamic spatiotemporal analysis, posing challenges for comprehensive disaster impact evaluations. This study proposes an SR-PR-FR comprehensive resilience evaluation model from three dimensions: structure resilience (SR), performance resilience (PR), and function resilience (FR). Moreover, a simulation model based on Geographic Information System (GIS) and Simulation of Urban MObility (SUMO) is developed to analyze the dynamic spatial–temporal effects of a rainstorm on traffic during Xi’an’s evening rush hour. The results reveal that the southwest part of Xi’an is most prone to being congested and slower to recover, while downtown flooding is the deepest, severely affecting emergency services’ efficiency. In addition, the road network resilience returns to 70% of the normal values only before the morning rush the next day. These research results are presented across both temporal and spatial dimensions, which can help managers propose more targeted recommendations for strengthening urban risk management.

One of the most important goals of cooperative driving is to control connected automated vehicles (CAVs) passing through conflict areas safely and efficiently without traffic signals. As a typical application scenario, allocating right-of-way reasonably at unsignalized intersections can effectively avoid collisions and reduce traffic delays. Proposed here is a new cooperative driving strategy for CAVs at unsignalized intersections based on distributed Monte Carlo tree search (MCTS). A task-area partition framework is also proposed to decompose the mission of cooperative driving into three main tasks: vehicle information sharing, passing order optimization, and trajectory control. Based on the schedule tree of the vehicle passing order, the root parallelization of MCTS combined with the majority voting rule is used to explore as many feasible passing orders (leaf nodes) as possible in a distributed way and find a nearly global-optimal passing order within the limited planning time. The aim is for CAVs to perform proper trajectory adjustments based on the obtained passing order to minimize traffic delays while making the slightest acceleration adjustments. A coupled simulation platform integrating SUMO and Python is developed to construct the unsignalized intersection scenarios and generate the proposed distributed cooperative driving strategy. Comparative analysis with conventional driving strategies demonstrates that the proposed strategy significantly enhances efficiency, safety, comfort, and emission, aligning well with innovative and environmentally friendly urban mobility aspirations.

Statistical analysis reveals that the unique environment of horizontal curve roads significantly contributes to the severity and fatality rates of traffic accidents. This study leveraged accident data from the Florida Department of Transportation (FDOT) to explore the severity of traffic accidents on horizontal curves and its influencing factors. Bayesian network was combined with information theory for the analysis of the severity and determinants of accidents on horizontal curves from the perspectives of network topology, the strength of the relationship between influencing factors, and the pathways of influencing factors. Results show that, (1) Traffic accident causation is complex, with a hierarchical network structure of factors rather than direct impacts from individual variables. (2) The strength of the relationship and dynamic change correlation between each variable are obtained. Results demonstrate that accidents are rarely caused by a single factor, and the severity of traffic accidents can be prevented and reduced by controlling variables states.(3) The analysis of the influence pathways of uncontrollable variables, like weather, revealed specific state combinations (e.g., Fog+Slippery, Rain+Slippery, Fog+Wet) that significantly escalate accident severity. This study presents an advanced model for predicting and diagnosing traffic accidents on horizontal curves, offering insights into the causative factors and their quantitative relationships and influence pathways. Keywords:Traffic safety, Horizontal curve, Bayesian network, Information theory, Accident prediction and diagnosis

This paper proposes a systematic Driver-Pressure-State-Impact-Response (DPSIR) method for traffic performance assessment under urban floods caused by extreme rainfall events. A coupled rainfall-flood-traffic model has been developed as the prerequisite for evaluation with publicly available data. The DPSIR method integrates the natural and transportation systems into cause-effect-response impact chains and facilitates dynamic traffic performance measurement on the road and network levels. We apply this approach to the urban networks' motorways, trunk, and arterial roads in Shanghai, China. The results help identify critical roads and verify the effectiveness of demand-control strategies in improving traffic resilience. By offering decision-makers a comprehensive view of traffic operation under disruptions, this method can help them build a more resilient transport system in flood risk management.

Pedestrian movements constitute a complex self-organizing system in which various potential risks and conflicts are introduced by the accumulation of randomness and inconsistency over time. Due to the lack of relevant data, knowledge of low-risk behavior identification and risk formation mechanisms in pedestrian movement is insufficient. In this study, we present probable risk indicators reflecting the consistency of a crowd’s state and establish a pedestrian risk identification model that provides an early warning method for safety and security management. A complete pedestrian movement risk identification process is summarized by converting real-world video files into calculable data via video recognition technology, making it possible to identify and obtain crowd risk information in real-time and locate abnormal phenomena. The process makes extensive use of hierarchical and machine learning methods for identifying pedestrian states, proposes quantified conflict-prone and congestion-prone indices to classify different risk types, and defines an improved “crowd risk” index for locating potential crowd dangers. The combination of video recognition technology with machine learning has apparent advantages in solving potential risks in pedestrian movement, particularly in terms of the risk identification rate, the effectiveness of the risk classification, and the accuracy of risk positioning. Results showed that the three-stage risk identification process could accurately determine congestion and conflicting risks and indicate the potential risk-prone in the crowd. The proposed method has significantly improved the output and description accuracy compared with other methods in terms of risk positioning.

Parking lots have many complex structures, diverse functions, and plentiful elements. The frequent flow of vehicles with narrow and dim spaces increases the probability of various traffic accidents. Due to the low severity and lack of relevant data, there is limited understanding of safety analyses for parking lot accidents. This study integrates multisource data to establish a Bayesian diagnostic model for parking lot accidents. The mutual information method is used to screen the possible influencing factors before modeling to reduce the subjectivity of Bayesian networks. Studying the cause and effect analysis of accidents provides diagnosis and prediction for property damage and event causes. This provides valuable correlation information between factors and accident characteristics, as well as consequences under the influence of multiple factor chains. As the developed model has good accuracy, this study proposes a parking lot safety evaluation system with a library of countermeasures based on the model results to ensure rigorous conclusions. The combination with ITS technology gives the system high scalability and adaptability in multiple scenarios.