Journal of Advanced Transportation

The frequent occurrence of secondary traffic accidents, characterized by vehicles losing control and straying into opposing lanes on highways, has emerged as a pressing concern. To address this issue, attention has been focused on the pivotal role of median guardrails as safety barriers. While conventional guardrails have effectively hindered vehicles from veering off course, mitigating accident severity, they are now inadequate in meeting the heightened protective standards necessitated by the surge in truck traffic and advancements in vehicle capabilities. To evaluate and enhance the protective capabilities of guardrails, this research employs a vehicle finite element (FE) model in conjunction with a W-beam guardrail system. Collision trajectories, acceleration, and displacement metrics were analyzed to compare the effectiveness of three improved guardrail designs in preventing crossing in the event of a runaway truck. Furthermore, based on the design of the retrofitted guardrail, the optimization of the structural parameters was carried out by a multiobjective optimization method using radial basis function (RBF) and NSGA-II algorithms with the size of the guardrail as the design variable. The collision simulation comparisons reveal that the double W-beam arch-reinforced guardrail surpasses both the double W-beam and the arch-reinforced guardrail regarding protective performance. Notably, the double W-beam design offers a viable option for disposing of obsolete guardrails postdemolition. The optimized design underscores that optimal structural protection is achieved when meticulously adjusting the thickness of the upper girder plate and the arch to precise dimensions. This refined guardrail system enhances safety and achieves material efficiency, utilizing less steel in its construction. By elucidating effective design modifications and the determination of optimal structural dimensions, this study provides its ideas for safer roads and more efficient infrastructure development.

Current airport ground operations, relying on single and fixed aircraft taxiing rules, struggle to handle dynamic traffic flow changes during peak flight times at large airports. This leads to inefficient taxiing routes, prolonged taxiing times, and high fuel consumption. This paper addresses these issues by proposing a new adaptive method for dynamic taxiway routing in airport ground operations. This method aims to reduce ground taxiing time and fuel consumption while ensuring the safety of aircraft taxiing. This study proposes a multiobjective A∗ algorithm with time windows which takes into account the allocation of resources on airport taxiways and introduces factors such as turning angles, dynamic turning speeds, and dynamic characteristics of the ground operations. Experiments conducted over the 10 busiest days in the history of Tianjin Binhai International Airport demonstrate that the algorithm excels in minimizing total taxiing time, differing only by 0.5% from the optimal solution. It also optimizes multiple objectives such as fuel consumption and operates at a solving speed approximately three orders of magnitude faster than the optimal solution algorithm, enabling real-time calculation of aircraft taxiing paths. The results of the study indicate that the proposed multiobjective A∗ algorithm with time windows can effectively provide decision support for dynamic routing in airport ground operations.

To ensure the safety of operations in the airfield area, it is crucial to address the increased conflict risks resulting from the growing number of vehicles and aircraft. Based on the complex network theory, this study takes aircraft and vehicles in the airfield area as nodes and selects five different indicators (average degree, average node weight, average weighted clustering coefficient, network density, and network efficiency) to characterize the operation state of the airfield area, so as to identify conflict risks. Building on this framework, an ATT-Bi-LSTM innovation prediction model based on LSTM network architecture is established to forecast the evolution of network indicators over time. By leveraging the algorithm to predict the temporal evolution of indicators, valuable insights into the future evolution of conflict risk can be gleaned from the prediction results. Real operational data from Xi’an Xianyang Airport are utilized as a demonstrative example in this study. The results of the experiments illustrate that the analytical approach proposed in this study achieves a precise identification of the indicators. The experimental results are then compared with data from other predictive models that operate on the same data set. Compared to alternative prediction models, the accuracy is increased by nearly 10%, reaching 89.78%. The results of the study help to accurately identify conflict risks in the airfield area in advance and provide strategic conflict avoidance strategies for relevant staff. This is essential to ensure the security of airfield area.

The traffic environment of mountainous highways is more complex than that of nonmountainous highways, with higher driving loads, which increases the risk in overtaking. The changes in the driver’s pupils, eye gaze behavior, and heart rate can be used to evaluate the level of driving tension and safety. To analyze the driving load while overtaking on two-lane highways in mountainous areas, an actual vehicle test was conducted. Twenty-one drivers were divided into a skilled group and an unskilled group. The gaze time, gaze transfer characteristics, heart rate changes, and pupil area changes during the three stages of overtaking (intention, execution, and return) were compared and analyzed. The comprehensive evaluation of driving load during the overtaking process used the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) method and Rank Sum Ratio (RSR) method. The results show that the two groups of drivers had the highest driving load during the overtaking execution stage and the lowest driving load during the intention stage. The driving load of overtaking on sections with poor-sight distance was significantly higher than that on sections with good-sight distance, and the risk in overtaking during the execution and return stages was highest on sections with poor-sight distance. It is possible to reduce the driving load if the driver is familiar with the road conditions or has a rich driving experience. Compared to the unskilled group, the skilled group had lower driving loads at all stages of overtaking. The research results can provide a theoretical basis for optimizing traffic safety prevention and control technology on mountainous highways and for designing intelligent driving assistance.

The objective of this study is to minimize the overall transportation cost through the joint decision-making for air-cargo hub network design and fleet planning under the uncertain environment. This joint decision-making considers various factors, including hub location, node connectivity, fleet size, and flight frequency. It takes into account several uncertain parameters such as air-cargo demand and transportation cost in a realistic setting. We propose a mixed-integer programming model tailored to the characteristics of such problem, which utilizes interval numbers to address these challenges. This model aims to provide a robust scheme for the joint hub network design and the fleet planning in the uncertain environment. An improved probability-based interval ranking method is proposed to solve the model. This transformation converts the proposed model into an equivalent real-number one, simplifying the solving process. Then a hybrid heuristic algorithm, combining the advantages of Memory-Based Genetic Algorithm (MBGA) and Greedy Heuristic Procedure (GHP), is introduced to enhance the solving speed. Finally, the performance of our proposed model and algorithm is verified using real-world data from the Australian postal dataset. The results show that the proposed model reduces hub construction costs by 1.37% and fleet operational costs by 7.60%, respectively, as opposed to the use of traditional approaches. The computational time of the proposed algorithm is reduced by 28.4% and 36.5%, respectively, when compared to the use of Genetic Algorithm (GA) and Variable Neighborhood Search (VNS) algorithm.

Although Type A personality traits have been confirmed to be more frequently engaged in risky driving behaviours, the existing research on the decision-making of queue-jumping behaviours has not considered this personality trait. This study aimed to explore the decision-making processes of the subject driver’s queue-jumping and the follower vehicle driver’s yielding behaviours with Type A and Type B personalities. First, the decision-making utility variables for both players were selected, and a payoff matrix considering utility variable weights was constructed. Next, a decision-making utility evaluation questionnaire was designed, and this questionnaire and the existing Type A behaviour pattern scale were investigated simultaneously. Then, the weight coefficients of the decision-making utility variables were calculated; the replicated dynamic equations of four game combinations were constructed and the local stability principle of a dynamic system was used to determine the evolutionarily stable strategy for each game combination. Finally, the evolutionary process in which subject vehicle drivers select jumping the queue strategy and follower vehicle drivers select giving way strategy was simulated using MATLAB software based on empirical data to verify the validity of the constructed evolutionary game model. The results indicated some differences in the weight coefficients of decision utility variables between Type A and Type B personalities. The constructed game model can effectively reflect the decision-making processes of subject and follower vehicle drivers of different personality types. The dynamic evolution processes of strategy selection were different for the four game combinations. This study revealed the evolutionary game process between subject and follower vehicle drivers, laying a theoretical foundation for traffic management departments to manage queue-jumping behaviours.

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.

Advances in vehicle intelligence have ushered in the rapid development of intelligent connected vehicles and the emergence of the Internet of Vehicles (IoV), greatly improving the passenger travel experience. However, as a new mode of transport, flexible public transportation presents challenges for operators in terms of reducing costs and improving passenger experiences through complex route planning. The present study introduces B∗ as a heuristic multiobjective route planning algorithm that addresses these challenges. Using the trajectory extraction procedure (TEP) and route assignment procedure (RAP), B∗ filters out inaccessible routes and plans efficient routes on the fly to save money and enhance the passenger experience. Experimental results show that B∗ outperforms traditional methods in terms of shorter driving distances and reduced passenger waiting times, highlighting its potential to optimize bus utilization and improve travel experiences.

As a new, highly complex, and far-reaching technology, autonomous driving can be associated with various fears and uncertainties. However, recent findings show that high trait anxiety can positively contribute to the intention to use (ITU) autonomous vehicles (AVs). An explanation for this is that the possibility of handing over one’s driving control to artificial intelligence (AI) is even more relieving for more anxious people. Our study aimed to test whether this explanation can be supported by investigating to what extent this relationship can be applied to buses in which control is handed over per se–in the conventional bus to a driver, and in the autonomous bus to the AI. We also analyzed how the fear of giving up control mediates the relationship between trait anxiety and ITU. In a quasi-experimental study, 253 subjects were surveyed while riding an autonomous or conventional electric bus. The results confirmed a positive association between trait anxiety and ITU in the overall sample but not in the autonomous and conventional subsamples. Contrary to our assumptions, fear of giving up control served as a slightly suppressive but not significant mediator. The results were independent of whether control was handed over to a human driver in the conventional electric bus or to AI in the autonomous bus. Our study thus provides fundamental new insights into the acceptance of AVs and buses in general and opens the door for subsequent research based on these findings.

Understanding the factors influencing ride-sourcing trips is crucial for enhancing the quality of personalized mobility and optimizing the allocation of transportation system resources. However, the nonlinear effects of dockless bike-sharing (DBS) and the built environment (BE) across different spatiotemporal contexts have not been adequately addressed in previous research. This study aims to bridge this gap by analyzing order data and BE characteristics in Tianjin, China. Utilizing the Gradient Boosting Decision Tree (GBDT) model and Accumulated Local Effects (ALE) plots, this study explores the relative importance and nonlinear thresholds of these factors on ride-sourcing trips. The findings reveal that DBS trips during weekday AM peak hours exert significantly negative effects on ride-sourcing, whereas the impact during weekend AM peaks and daily PM peaks is positive. Furthermore, variables such as active population density, metro accessibility, and residential, entertainment, and cultural BEs have positive nonlinear impacts on ride-sourcing trips. These insights offer policy implications and resource allocation recommendations for both government bodies and operators.

In an urban road network, the ability of the traffic flow itself to alleviate congestion caused by external disruptions is overlooked. This study applied the two-fluid model to simulate mesoscopic traffic flow, focusing on the resilience of urban road networks under normal disturbances. Three resilience indices—plasticity, transition of elasticity, and elasticity—were introduced based on the failure deformation process of rigid materials. These indices were used to modify the two-fluid model’s parameters, considering the effects of bus operations and temporary roadblocks on traffic flow and service quality. A hidden Markov model (HMM) was employed to predict service quality transitions (distinction, merit, and pass), with validation using dynamic bayonet traffic data from Xuancheng and video recordings from Xi’an, China. The results confirmed that resilience varies significantly across different times and locations, with peak hours and dense urban areas exhibiting lower resilience and higher susceptibility to disruptions. Bus queuing was found to degrade service quality, and rainstorms had a more severe impact than construction zones. The study can aid in the development of management efficiency of urban road networks.

The rapid growth of cities and the increasing traffic congestion have made vehicle emissions worse, especially in developing countries. Governments worldwide are now relying on regulations and policies to manage and reduce these emissions effectively. This change in approach towards emission control is also happening in developing nations. In this study, the effects of these policy measures are quantified using a calibrated integrated traffic and emission model (AIMSUN), and two hypothetical scenarios were analyzed: one scenario is where electric vehicles (EVs) replace traditional internal combustion EVs (ICEVs) in the study area and another scenario is assuming strict implementation of Euro 4/IV emission standards. The results showed that shifting towards a higher proportion of EVs leads to significant reductions in emissions but requires increased battery consumption, highlighting the trade-off between reducing emissions and higher energy demand. Implementing Euro 4/IV standards could considerably reduce emissions, especially from motorcycles and trucks. It suggests that focusing on these categories with a phased implementation approach could bring significant environmental benefits. Policymakers in developing countries should adopt a rounded approach instead of implementing strict policies. It is crucial for them to carefully weigh the pros and cons of policy instruments before making any decisions. This study shows how traffic micro‐simulation modeling coupled with emission models can be used in evidence-based decision-making.

Enhancing hazmat truck safety through advanced driving assistance systems (ADAS) relies on both system efficacy and driver reactions. This study investigates the driving behaviors of hazmat truck drivers in response to forward collision warnings (FCWs). Traditional warning triggering methods struggle to capture diverse and immediate driver responses; therefore, our research employs a vision-based framework for driving data extraction and utilizes the K-means++ clustering method for response-based classification. Moreover, we propose an enhanced version of the intelligent driver model (IDM) based on the concept of a virtual vehicle to reproduce hazmat truck drivers’ differential behaviors during risky car-following periods, achieving results that depict improved driving simulations. This model is compared with classic benchmarks, including the IDM, optimal velocity model (OVM), and full velocity difference (FVD) model, demonstrating superior performance in terms of traffic stability and safety in extreme scenarios. Our findings highlight that preaction drivers tend to accelerate before receiving warnings, opting to overtake rather than maintain safe distances. In contrast, calm drivers decelerate in anticipation of the warning, showcasing their awareness of maintaining safety. The analysis reveals that aggressive drivers are predominantly in the 41–45 age group, indicating a higher skill level, while calm drivers are more commonly older, reflecting a trend in cautious driving behaviors. Overall, our research contributes to the development of effective ADAS by considering real-time driver responses and emphasizes the potential of our model to revolutionize commercial ADAS adoption and enhance road safety for hazmat operations.

The research on tunnel resilience has garnered increasing attention in recent years. Owing to prolonged exposure to natural or anthropogenic factors, the resilience level of many highway tunnels is continuously declining, rendering them susceptible to sudden accidents and challenging to restore postincident. Currently, although several scholars have employed diverse evaluation methods in their research on tunnel resilience, there is a lack of summarization and integration of these methods. In addition, there is also a dearth of a unified evaluation index system and framework for different types of natural or man-made disasters, which are crucial for advancing the development of tunnel resilience evaluation. This study commences with an introduction to the origin of resilience and the definition of tunnel resilience, and comprehensively summarizes commonly employed evaluation methods. Subsequently, this study centers on the resilience evaluation methods in tunnel engineering and analyzes their strengths and weaknesses. Besides, the distribution of resilience metrics in current researches is analyzed and the detailed explanations for the diverse choices are provided. According to the results and deficiencies of existing research, combined with the author’s perspectives, the index systems, evaluation frameworks, and resilience improvement strategies are proposed, which can be applied to the resilience evaluation of various operational highway tunnels under diverse disaster scenarios. Furthermore, this study also presents a case study on the evaluation of tunnel fire resilience to validate the applicability of the research findings. These findings aim to provide a guide for the operation and maintenance management of the operating tunnels and improve the scientific decision-making level of tunnel maintenance.

Multimodal public transport network (MPTN) plays an important role in relieving road traffic pressure for metropolitan area. Nevertheless, the impact of an accident happened in an individual station may not only disrupt the station itself or the single lines that go through the station but also spread over the whole network. Therefore, identifying the vulnerable stations is essential for improving the MPTN management against the systematic risk caused by accidents. In this paper, we proposed a route diversity-based approach to measure the vulnerability of stations in MPTN based on the complex network theory. The route constraint parameters were established to reflect the travel time restriction in constructing the set of passengers’ acceptable routes. In addition, an algorithm was formulated to rapidly calculate the route diversity index and meanwhile avoid the “overlapping routes” problem. A simple virtual network was used as a numerical example to compare the proposed approach with the vulnerability evaluation approaches based on degree centrality and betweenness centrality. Finally, the proposed approach was applied to the MPTN of Beijing to explain its effectiveness and potential applications. The results show that the proposed method can efficaciously estimate vulnerable nodes compared with degree centrality and betweenness centrality. Meanwhile, the acceptable routes between any OD pairs in the MPTN are 1–10 according to the constrained parameter. In addition, the average number of acceptable routes between OD pairs of Beijing MPTN is 3.649. By ranking the stations according to their vulnerability, it can be found that the top 5 vulnerable stations are all external traffic hubs or the stations around famous commercial areas. The results suggest that these stations are significant for external transport as well as crucial for internal urban transportation systems. The research output could contribute to the MPTN management in accident prevention and emergency handling.

To investigate the potential lane-changing collision risks that may arise between vehicles during lane changes and those in the original lane, a model for vehicle lane-changing collision risk is constructed specifically for this scenario, and a research analysis is conducted. First, based on vehicle trajectory data, a sample set capturing the relationships between vehicles traveling in a straight line and those changing lanes laterally is extracted and built. Interpolation methods are then applied to fill in missing values, outliers are eliminated, and data noise is smoothed during preprocessing. After preprocessing, a total of 468 vehicle pairs and 265,392 data points are obtained. Second, a real-time collision time model is established based on the preprocessed data, and collision risk probabilities are calculated accordingly. Then, the collision risks are classified into four levels based on whether the vehicle on the side actually changes lanes and the severity of the collision risks. Finally, a light gradient boosting machine (LightGBM) learning method is adopted to predict the risk levels and analyze the main factors that significantly impact the severity of collision risks. The results indicate that the longitudinal distance between the target vehicle and the preceding vehicle is the most critical influencing factor, followed by the speed of the target vehicle itself, and then the speed difference between the target vehicle and the preceding vehicle. The influence of other factors is relatively similar and does not have a significant impact.

The principle of system resilience is its ability to withstand disruptions and maintain an equilibrium state. In urban network systems, adaptive traffic signal control (ATSC) has been an effective countermeasure to mitigate traffic flow disturbance and improve resilience. This research has explored the usage of a decentralised advantage actor-critic (a2c) algorithm-based ATSC in mitigating disruptions, particularly nonrecurring congestion caused by car accidents. A reward function has also been proposed, combining deduced resilience metric, safety indicator time to collision (TTC) and system performance. A virtual simulation environment was created using simulation of urban mobility (SUMO) to facilitate the evaluation of the proposed approach. In the grid simulation environment, an overall 5.8% improvement is achieved, exceeding benchmark algorithms in three metrics, especially performance with a margin of over 5.2%. Robustness against different levels of car accidents are proven as well. Further evaluation is also implemented based on a real-world case study and demonstrates an improvement of 20.08%, highlighting the correlation of proposed method’s efficiency on the traffic flow rate and road structure.

Safety performance functions (SPFs) have become valuable tools for estimating the relationships between crashes and various causal factors when constructing crash-prediction models. However, the commonly used independent variable, the annual average daily traffic (AADT) is data on a yearly basis, which has limitations in capturing the temporal characteristics of traffic flows influenced by the passage of time. Accordingly, there have also been many studies using 15 min data to reflect real-time, which is an important time unit to understand changes in highway traffic flow. However, such a short time unit has the limitation of high instability and randomness. In light of this, this study recognizes the importance of the 15 min time interval and proposes a new approach by developing a modified hourly model that aggregates data at fine-grained 15 min intervals (00, 15, 30, and 45 min, both at the beginning and end), instead of the traditional hourly data that starts and ends at the peak of each hour to compensate for the existing limitations. The analysis focused on South Korea’s nationwide highways, and models were developed based on both statistical and machine-learning approaches to compare their performances for selecting the final model. Additionally, a modified temporal SPF is introduced to predict crashes by assigning weights based on a Dirichlet distribution to models with overlapping time intervals aggregated in 15 min increments. This innovative approach overcomes the limitations of existing 15 min models, where the number of crashes is too small for effective training if the model is simply developed by dividing the time. The anticipated outcome is that the proposed model will demonstrate excellent performance and serve as an effective tool for predicting highway crash risks.

Longitudinal control of autonomous vehicles (AVs) has long been a prominent subject and challenge. A hierarchical longitudinal control system that integrates deep deterministic policy gradient (DDPG) and proportional–integral–derivative (PID) control algorithms was proposed in this paper to ensure safe and efficient vehicle operation. First, a hierarchical control structure was employed to devise the longitudinal control algorithm, utilizing a Carsim-based model of the vehicle’s longitudinal dynamics. Subsequently, an upper controller algorithm was developed, combining DDPG and PID, wherein perceptual information such as leading vehicle speed and distance served as input state for the DDPG algorithm to determine PID parameters and output the desired acceleration of the vehicle. Following this, a lower controller was designed employing a PID-based driving and braking switching strategy. The disparity between the desired and actual accelerations was fed into the PID, which calculated the control acceleration to enact the driving and braking switching strategy. Finally, the effectiveness of the designed control algorithm was validated through simulation scenarios using Carsim and Simulink. Results demonstrate that the longitudinal control method proposed herein adeptly manages vehicle speed and following distance, thus satisfying the safety requirements of AVs.

Air taxis, a core service within urban air mobility (UAM), have the potential to enhance user satisfaction and address societal challenges such as traffic congestion and environmental pollution. However, the success of this service is often hindered by various concerns. To ensure successful implementation, we investigate the factors influencing public acceptance of air taxis. This study distinguishes itself from previous research in three key aspects. First, it introduces a novel classification of the factors into individual and societal dimensions. Second, it is among the first to apply a value-based adoption model to understand the intention to adopt air taxis, including UAM. Third, it uniquely considers the Korean perspective, unlike most existing studies that focus on Western cultural contexts. To identify the consumers’ perceptions, we conducted interviews with experts and surveyed a sample of 1,000 members of the general public in Korea. Our findings suggest that perceived value for society, as well as perceived value for individual users, significantly influences adoption intention. We discuss both academic insights and practical implications for policy and industry, supporting the commercialization of Korean UAM (K-UAM) promoted by the Korean government.

Deviating to the left on two-way roads can result in fatal head-on collisions. This article presents an intelligent decision-making and path-planning algorithm aimed at avoiding collision with a vehicle that has deviated from the opposing lane. The path-planning process utilizes the model predictive control (MPC) approach, employing a linear kinematic prediction model with a horizon of 2 seconds. Considering that the deviated vehicle may abruptly return to its original lane at any moment, its motion is associated with significant uncertainty. To address this challenge, the path-planning algorithm directs the ego vehicle (EV) under specific constraints to ensure that both the left and right sides of the road are symmetrically reachable in future time steps. This enables the decision-making algorithm to select the safer direction for evasive maneuver at the appropriate moment. The motion prediction of the threat vehicle (TV) is conducted until the potential collision time, taking into account its motion history, and is utilized in the decision-making process. Once the maneuver direction is determined, the collision-free path planning continues using the MPC method. To evaluate the algorithm, six simulations are conducted, modeling various distant and close encounter states of the vehicles on roads with left- and right-hand curves. The simulation results indicate the flexibility and appropriate performance of the algorithm in planning safe and maneuverable paths.

Based on the principle of direct linear transformation (DLT) in close-range photogrammetry, a method was proposed for reconstructing the motion states of the host vehicle and other vehicles based on vehicle-mounted videos. To verify the effectiveness and accuracy of the method, validation experiments were designed. Under two typical operating states, steering and straight driving, the motion states of the host vehicle and other vehicles (including trajectory, distance, speed, and acceleration) were reconstructed from the vehicle-mounted video. In the experiments, high-precision inertial navigation was installed on the other vehicle to record real-time motion data of the vehicle. Finally, in order to compare and analyze the reconstructed video results with the vehicle’s actual motion data, the recorded motion data were matched and synchronized to the same time axis as the vehicle-mounted videos through a GPS timing device. The experimental result shows that the reconstructed trajectory results based on this method can generally reflect the vehicle’s actual trajectory, with an average deviation of less than 7.4%; the reconstructed distance results have an average deviation of less than 9.3%; the reconstructed speed results have an average deviation of less than 7.3%; the reconstructed acceleration results can reflect the vehicle’s acceleration or deceleration states. The results of this study provide an effective solution for obtaining important parameters of vehicles in accident reconstruction research, such as the trajectory, speed, distance, and acceleration or deceleration, and it has significant practical value for applications.

With the development of the Shanghai International Shipping Center, the diversity of vehicle types on the highways and arterial roads near Yangshan port is continually increasing. Within such a container port corridor, large container trucks are primarily utilized for mainline transportation. Their larger size and significant inertia would increase psychological pressure on sedan drivers, and elevate their behavior risk. To investigate the effects of container trucks on drivers’ visual characteristics and driving behavior as well as to predict driving risk, firstly, this research conducted field tests in four scenarios surrounding the port. Visual characteristics and behavior data of sedan drivers were collected. Secondly, a “Visual-behavior” chain model was established. The relationship between drivers’ visual characteristics, driving behavior, and driving risk was illustrated from the perspective of time-series behavior patterns. Thirdly, three driving risk prediction models were built with Autoregressive Integrated Moving Average (ARIMA), Long Short-Term Memory (LSTM), and ARIMA-LSTM. The results indicate that the ARIMA-LSTM model shows the most effective prediction performance. This research provides a field-data comparative analysis of the driving risks influenced by a high proportion of container trucks. The findings contribute to understanding the unique mixed traffic visual environment around large-scale container ports.

In order to accurately predict the lane-changing trajectory of the vehicle and improve the driving safety of the vehicle, a lane-changing trajectory prediction model based on the combination of convolutional neural network (CNN) and bidirectional long short-term memory (BiLSTM) neural network is proposed by comprehensively considering the historical driving behavior, the spatial characteristics of surrounding vehicles and the bidirectional time sequence information of the vehicle trajectory. Firstly, the vehicle trajectory data are filtered and smoothed, and it is divided into three categories: left lane change, right lane change, and straight driving, and a lane change trajectory sample set is constructed. Secondly, CNN-BiLSTM model is constructed to identify the sample set of lane-changing trajectory. Considering the interaction between vehicles in the driving process, the information of predicted vehicle, and surrounding vehicles is taken as the input of the model. The extracted feature vector is input to the BiLSTM layer for prediction after the CNN layer feature extraction, and the horizontal and vertical coordinates of the target vehicle at the next time are output. Thirdly, the trajectory data of the US-101 dataset in NGSIM is selected to verify the performance of the CNN-BiLSTM model, and at the same time, it is compared with models such as CNN-LSTM, long short-term memory (LSTM), BiLSTM, and CNN-GRU-Att. Finally, the verification result shows that the overall fitting degree of the vehicle lane change trajectory prediction of the proposed model reaches 99.50%, and the mean square error and mean absolute error are 0.0003076 and 0.01417, which are improved compared with other models. In the meanwhile, the research on multistep trajectory prediction in different prediction time domains is carried out. It was found that the longer the prediction time domain is, the lower the prediction performance of the model decreases, but the prediction accuracy still reached more than 96%, and it was able to accurately predict the lane change trajectory.

Accurately predicting origin-destination (OD) passenger flows serves as the basis for implementing efficient plans, including line planning and timetabling. However, due to the complexity and variety of OD passenger flows types, general prediction models have difficulty in capturing the features of different OD passenger flows, which in turn leads to poor prediction performance. To address this issue, we propose an integrated framework that combines clustering and prediction methods. First, an unsupervised deep learning model is devised to automatically cluster OD flow types by capturing shape characteristics. Second, three types of features are created to enhance training efficiency, including static features, time-dependent observed features, and time-dependent known features. Based on the clustering of OD passenger flow, a weighted adaptive passenger flow prediction model is developed. The study employs a temporal fusion transformers model to enable multitype OD passenger flow prediction. In the numerical experiments, the model was applied to the urban rail transit in South China, and the model clustered 15,168 OD pairs into 4 types for prediction. The findings show that this approach enhanced the prediction accuracy by 2.0%–9.6% compared to the LSTM model and by 1.6%–4.3% compared to the Graph WaveNet. Moreover, the model can accurately assess the various features for diverse types of OD flows.