Journal of Advanced Transportation最新文献

This study proposes a label-setting shortest-path algorithm with timetable and label constraints to address the path-planning problem in multimodal urban agglomeration transportation networks. The proposed algorithm addresses the limitations of traditional shortest-path methods, which are often challenged by complex conditions, including transfer constraints, timetable dependencies, and restrictions on the number of transfers. By analyzing the characteristics of urban agglomeration trip chains, this study constructed a multimodal network model incorporating six transportation modes: walking, buses, intercity coaches, metro, intercity railways, and private vehicles. A deterministic finite automaton was introduced to constrain feasible mode sequences, ensuring that path planning aligns with real-world travel patterns. The algorithm incorporates time-window constraints to simulate the effects of static timetables in scheduled transportation modes, such as railways and coaches. By improving the label-correcting algorithm via topological sorting and applying state-dominance rules to reduce redundant computations, it achieves optimal path planning under multiple constraints. Case studies demonstrate that the algorithm effectively balances transfer frequency and travel cost, reducing total cost by approximately 5% while ensuring feasibility, thereby validating the synergistic advantages of multimodal transportation networks. Therefore, the proposed algorithm can theoretically support multimodal traffic selection behavior or mixed traffic network modeling.

In loitering activity scenarios, vessels frequently execute course changes within localized maritime spaces, often exhibiting extreme turning maneuvers that generate ultralong, dense, and highly nonlinear spatiotemporal trajectories. Traditional prediction models demonstrate limitations in processing dynamically changing trajectory features, leading to insufficient prediction accuracy under such loitering conditions. To address this challenge, this study proposes a GL-LoiterDNet, a hybrid deep learning–based vessel trajectory prediction model. The model incorporates multidimensional trajectory characterization features including speed fluctuations, navigational positions, and course differentials. It integrates 1D convolutional neural networks (1D-CNN), bidirectional gated recurrent units (BiGRU), and bidirectional long short-term memory (BiLSTM) networks to capture both localized abrupt variations and long-term evolutionary patterns in loitering trajectories, thereby mitigating feature degradation phenomena. Experimental validation using trajectory data from vessels in Sagami Bay, Japan, demonstrates that the GL-LoiterDNet model outperforms 14 baseline models in prediction accuracy and robustness. The model exhibits rolling multistep trajectory prediction capability for loitering scenarios, achieving an average positioning error below 0.7 km within 10-min prediction windows. This research can provide reliable theoretical and data-driven support for continuous vessel positioning and monitoring in complex maritime operation scenarios.

Public health crises profoundly impact work activities and commuting behaviors. The multiple waves of COVID-19 outbreaks across cities worldwide have demonstrated that people voluntarily or involuntarily adapt their work patterns, such as shifting to remote work, and adjust their commuting choices. This study investigates changes in commuting behaviors among urban residents during three pandemic wave stages: prewave, outbreak, and postwave, focusing on work patterns, commuting frequency, and key influencing factors. A retrospective longitudinal survey was conducted in Shanghai after the first wave of COVID-19 outbreak to collect information on respondents’ work, commuting, pandemic-related, and sociodemographic characteristics. Descriptive analysis and statistical tests revealed a 35% increase in telecommuting and a 50% decrease in commuting trips during the outbreak, with near-normal levels postwave. Multinomial logit models identified key factors influencing commuting frequency, such as telecommuting, weekly working hours, and direct commutes to workplaces. Telecommuting increased commuting frequency prewave, decreased it during the outbreak, and continued to reduce it postwave. Work intensity consistently increased commuting frequency, with the most significant impact prewave and the least during the outbreak. The findings provide insights for policymakers to better understand and enhance strategies in response to unforeseen public events, including potential future pandemics like Disease X.

Anticipating traffic patterns is a vital component in efficiently managing traffic within smart cities. This proactive approach contributes to alleviating traffic congestion, promoting environmental preservation, reducing commute times, and strengthening overall safety measures. Numerous studies have verified that external factors, such as weather conditions, exert an influence on traffic patterns. Therefore, utilizing these factors as supplementary variables can improve traffic prediction accuracy. This paper presents a novel prediction model called the multi-input sequential multihead attention (MI-SMHA) model. This model integrates weather condition information into traffic prediction tasks, aiming to enhance prediction accuracy and computational efficiency. It utilizes advanced techniques from sequential modeling and attention mechanisms, specifically tailored to handle traffic and weather data such as temperature, wind speed, precipitation, visibility, and humidity. This integration aims to leverage the complementary nature of weather conditions in forecasting traffic patterns, yet it remains flexible enough to be generalized to support a wide range of multivariate time series prediction tasks. Data from real-life traffic detectors are utilized to perform experiments and comparisons with two baseline models and three state-of-the-art models to validate and assess the efficiency of the proposed model. The MI-SMHA model was efficient and reliable in forecasting future traffic flow, significantly reducing errors compared to the other models.

This study investigates key factors influencing the immersive experience of passengers in autonomous vehicles (AVs) and proposes a novel theoretical model. In this model, four core dimensions were identified: (1) Emotional and Sensory Experience, (2) Interaction and Engagement Experience, (3) Trust and Safety Experience and (4) Dispositional Experience. It is found that Emotional and Sensory factors, such as lighting and sound, primarily affect passenger comfort and mood. In contrast, the Interaction and Engagement factors that focus on human–machine interaction (HMI) and AR/VR devices enhance passengers’ engagement. As for trust and safety factors, passengers’ confidence towards AVs is addressed through clear communication during driving processes. Dispositional factors, including technology acceptance and personalisation, contribute to passengers’ overall satisfaction in AVs. In addition, external factors such as intelligent transportation systems (ITSs), intelligent connected vehicles (ICVs) and smart city infrastructure further impact passengers’ experiences in safety and efficiency. The study highlights several emerging research areas requiring further investigation, such as multisensory feedback, dynamic personalisation and cultural inclusivity differences in AV experience. The proposed theoretical model serves as a foundation for future work aimed at enabling the design of AV systems that are more attentive and accommodating to passengers by sourcing control both domestically and externally, ultimately enhancing the passengers’ experience.

Safety anomalies are the early warning and precursors to major accidents. Preventing such incidents requires robust accident models to identify and mitigate risk factors. This study enhances the system hazard identification, prediction, and prevention (SHIPP) model to develop a safety barrier–based accident analysis framework tailored to high-speed train operations. The proposed model employs a fault tree to represent the causal relationships among various safety barriers and an event tree to depict the progression from safe operation to catastrophic outcomes. To validate the approach, the study collects a total of 60 cases of operational safety accidents and reconstructs the accident process comprehensively. The causal relationships among contributing factors are visualized clearly, providing a foundation and technical support for accident process analysis and the formulation of preventive measures.

High-speed railway (HSR) operations heavily rely on visual inputs, yet there is a notable gap in examining how HSR drivers adjust their eye movements in response to different lighting conditions, despite the pivotal role visual cues play in such environments. This investigation employed a Tobii Nano eye-tracker to capture the visual behaviors of HSR drivers during simulated driving exercises. It centered on 4 areas of interest (AOIs): the front window, prompt area, dashboard, and speed dial. By using Train Sim World 3, we created 3 scenes (open section, short tunnel, and long tunnel) and utilized 4 key metrics (average pupil diameter, APD; time to first fixation, TFF; duration of first fixation, DFF; fixation duration, FD) to evaluate the variations in visual attention patterns of HSR drivers. The results reveal a significant relationship between these indicators and driving scenes. Drivers in tunnel settings tend to have a longer time to form fixations, reflected by longer TFF, duration of first fixation (DFF), and FD. Pupil dilation is most pronounced in tunnels with weaker light (long tunnels), while stronger light (short tunnels) leads to increased TFF, DFF, and FD. At the outset of the driving task, the front window and speed dial are the earliest fixated (earlier TFF). Throughout the driving, speed dial continues to be a central fixation, manifested by extended DFF and FD. Gaining insight into HSR drivers’ visual behaviors is essential for enhancing driving safety.

With the continuous growth of high-speed railway passenger transportation demand, how to improve the capacity has become an urgent problem to be solved. The signal system based on moving block can effectively improve the utilization of line capacity. From the perspective of signal system, this paper studies the line capacity benefits brought by CTCS-3 combined with moving block. First, in response to the challenges of implementing moving block under CTCS-4 based on existing technologies and considering the need for line interconnection, this paper proposes a CTCS-3 solution that combined moving block. Secondly, this paper proposes a multiagent-based high-speed railway network train tracking simulation modeling method and establishes infrastructure and train simulation models under two signal system scenarios: CTCS-3 and CTCS-3 combined with moving block. Finally, this paper selects the Beijing-Shanghai High-Speed Railway as a research case and verifies the railway capacity indicators. The results show that the application of CTCS-3 combined with moving block is expected to further tap the transportation capacity potential of the existing high-speed railway network.

Confronted with the disturbances arising from various risk events, it is crucial to accurately measure the severity of risks in the dispatching section for efficient train operation and transportation management of a high-speed railway (HSR). This paper proposes a risk mapping method for daily HSR disturbances based on a self-formulated operation loss model, aiming to assist in identifying the spatiotemporal transportation bottlenecks and mitigating the propagation of risks. The calculation models for operation loss under risk disturbances are first established, with a focus on the instantaneous operation loss (IOL) of affected trains and the cumulative operation loss (COL) of the dispatching section, giving specific considerations on delay status, train importance, and operation scheme. Based on the delay characteristics observed in various risk scenarios, the variation curves of IOL for affected trains and dispatching sections are categorized into triangular and trapezoidal patterns. Combining the historical data statistics, the spatiotemporal risk distribution matrix is then established by occurrence probability calculation, event probability decomposition, and grid operation loss calculation, using well-designed algorithms. Meanwhile, the importance of risk scenario features is analyzed through LightGBM classification to identify key attributes. To validate the feasibility of the proposed approach, a case study has been conducted on weekday risk disturbances in a dispatching section administrated by the Shanghai Railway Bureau. The results demonstrate that this approach can accurately depict the distribution of risk severity by considering both operation losses and decomposed probabilities, where the average COL of station risks ranges from 0.14 to 0.64, while the average COL of section risks ranges from 0.09 to 0.49. Furthermore, the attributes contributing to the risk severity can be effectively extracted for various scenarios, such as the primary delay, risk position, and train speed heterogeneity. Finally, a discussion on the generalizability and challenges of applying this method provides further verification and detailed explanations for HSR risk mapping.

This study examined underground roads to evaluate the effects of traffic congestion prevention strategies. A specific framework, called the traffic congestion judgment criteria and process (TJCAP), was developed for underground road application. Using this framework, the study analyzed congestion relief effects by applying traffic strategies commonly used on surface roads. A real underground road in Seoul was used as a testbed. Microscopic traffic simulation was conducted using the VISSIM to create a realistic simulation network. The model was calibrated using observed traffic volume and speed data, both on the underground and adjacent surface roads. This approach enabled the analysis of traffic strategies aimed at reducing congestion. Results showed that the effectiveness of the strategies depends on the type of surface road (interrupted or uninterrupted flow) and its traffic conditions. In particular, the strategies were effective when the connected surface road had a level of service (LOS) of D or better.