Complex road infrastructure and traffic operation environments on ground roads can significantly affect the lane line perception of automated vehicles. Focusing on ground roads, a test vehicle equipped with a multi-sensor system from the Tongji University vehicle-mounted holographic information collection system was used to conduct LiDAR-based lane detection tests on six accident-prone roads in Shanghai. A hybrid modeling approach (combining machine learning and binary logistic regression models) was used to analyze the key factors that influence autonomous driving lane detection failure and the impacts of their feature changes on three dimensions: road characteristics, lane marking properties, and traffic operation. First, a feature importance analysis was conducted based on the LightGBM model, and the SHapley Additive exPlanations (SHAP) method was used to analyze the impacts of individual feature changes on lane detection failure. Next, a binary logistic regression model was used to determine the significant factors and their interaction effects on the important influential factors. The feature importance results indicate that the 10 critical factors influencing lane line perception failure, in decreasing order of importance, are the type of marking combination, test vehicle operating speed, absence of markings, vehicle occlusion, road width, lane width, marking wear, type of leading vehicle, type of guiding marking, and number of leading vehicles. A narrower lane width (2.6 m) and the presence of large trucks in front of the vehicle increased the probability of lane detection failure. Missing or worn lane markings, double-dashed-line combinations, and pedestrian crosswalk lines increased the probability of detection failure. The results of the binary logistic model indicate that, except for road width, all other factors significantly affect lane line detection failure. Moreover, there are interaction effects between the lead vehicle type and lane width as well as between the lead vehicle type and operating speed. The research findings provide guidance for lane design and optimization in mixed traffic environments, and they offer directions for sensor manufacturers and automakers to optimize LiDAR perception algorithms.

Connected and automated vehicles facilitate rapid, collaborative, and shared travel by integrating information flows in the cyber-communication layer with physical entities in the vehicular layer. However, challenges such as channel fading and resource scarcity persist in the cyber communication domain. These issues contribute to system nonlinearity, random disturbances, and behavioral uncertainty during vehicle operation, which thereby undermines platoon stability. Hence, the development of platoon control methods has become a critical area of research for ensuring vehicle safety and mitigating the adverse effects of communication delays on vehicle traveling states. Safety platoon control focuses on considering the spatiotemporal value of the traveling state under communication delays from the perspectives of spatial computing loss and information timeliness. Significant research value and practical implications for enhancing the intrinsic safety of vehicles and the robustness of communication networks are offered. First, by acknowledging the spatial correlation of information transmission, a Spatial Adjustment intelligent driving model (SA-IDM) was developed to normalize the communication structures among multiple preceding vehicles while adjusting the state information based on signal attenuation. Second, by recognizing the temporal correlation of information transmission, the age of information (AoI) was introduced to quantify the timeliness of state information. A frequency domain analysis was conducted to assess the string stability of the SA-IDM and derive an AoI threshold that indicates when information is outdated. This threshold was converted into a time value for the entire transmission process. In addition, a retransmission strategy using an information update roadside unit was proposed. Finally, the model parameters were calibrated using real vehicle tests and a genetic algorithm. Numerical simulations confirm the effectiveness of the proposed model. Results show that SA-IDM can suppress the impacts of disturbances from the leading vehicle of a platoon, and when the AoI threshold is 0.2 s, it can effectively measure the value of time. The state information retransmission strategy ensures the timely acquisition of information by the platoon under a high delay occurrence rate (DOR). When the DOR=20%, the reduction in the average spacing deviation of the SA-IDM compared with that of the IDM is 33.5%. When the DOR=70%, the average spacing deviation reduction achieved by the information update-based roadside unit state information retransmission strategy, compared with the maximum AoI retransmission strategy, is 21%. In addition, vehicles located at the front and rear of the SA-IDM platoon exhibit weaker safety, thus necessitating more frequent information transmission.

At complex intersections, such as large and skewed intersections, autonomous vehicles (AVs) face the challenge of multitarget collision avoidance when performing right-turn driving tasks that require the real-time perception of dynamic changes in a road traffic environment. To evaluate the readiness of right turns at intersections from the perspective of areas from where AVs must accurately perceive the right turn (i.e., the safe sight zone), field tests were conducted at six smart intersections in Shanghai, China. This study aims to reveal the relationship between the road traffic environment and autonomous driving perception and to clarify intersection boundaries. A safe sight zone evaluation method based on driving tasks and perceived targets is proposed. Through driving task analysis, perceptual subtasks for “intersection entry” and “execute turn” were identified. The static road infrastructure and dynamic traffic participants that needed to be detected were identified to form dynamic safe sight zones. The perception capabilities of safe sight zones were evaluated based on whether the AVs accurately detected the perceptual targets. The intersection design and traffic environment were considered as the input features, and a perception prediction model was constructed using the CatBoost ensemble learning model. The relationship between perception capability and contributing factors were revealed using the SHapley Additive exPlanations (SHAP) post hoc interpretation technique. The results indicate the following. ① Dashed-dashed types of lane lines have a great effect on detection failure. The failure probability increases for right-turn radii greater than or equal to 8 m. Large intersections, such as those with wider lanes (15.5 m), and more lanes (four lanes) at the entrance are more likely to lead to failure. ② Failure probability increases in dusk conditions. The failure probability increases when there are more than two non-motorized vehicles or more than one motorized vehicle in the front. Large trucks in the front affect failure to a greater extent than do small vehicles. The results support readiness evaluations for intersections with similar road traffic environments from the perspective of sight-zone perception, guide the management of smart intersection reconstruction and expansion, and the management of AV-testing open roads. These results also provide the contributing factors and thresholds for the optimization of autonomous driving perception algorithms.

The field-of-view of a roadside sensing unit (RSU) is limited. In addition, static and dynamic occlusions in traffic scenarios may adversely affect the perception accuracy of an RSU. In such cases, even if multiple RSUs cooperate, the RSU-captured trajectories may deviate from the ground truth and vary with the RSU placement, thereby impairing the performance of the RSU in conflict risk monitoring. Considering the need to optimize the RSU configuration prior to real-world deployment, this study proposes a framework for virtually assessing the RSU risk monitoring performance at intersections using trajectory data. To emulate the real-time risk surveillance process of an RSU, an agent-based method was developed to identify and analyze multimodal conflict risks at intersections. The trajectory data were integrated with a three-dimensional intersection model to reconstruct dynamic traffic scenarios. After the parameterization of the RSU model, a virtual procedure was established to simulate cooperative trajectory perception. To address the time efficiency issue of collecting cooperatively perceived trajectory data using continuous simulations, a deep neural network (DNN) was introduced to generate co-perceived trajectories using prior knowledge (raw trajectory data and probabilistic occupancy maps) and simulated trajectory data as the input and response, respectively. The raw and RSU-captured trajectory data were used separately to compute the time-to-collision (TTC) indices of conflict events, and TTC-based cross-entropy was introduced to quantify the risk monitoring performance of the RSU. The proposed framework was tested on trajectory data collected from Intersections A and E in the open-source CitySim dataset. The results indicate that RSU risk monitoring performance is closely associated with RSU placement. The average root mean square error of the trained DNN model on the test dataset was 0.353, which implies that the DNN model can generate RSU-captured trajectories. The TTC-based cross-entropy obtained by the DNN-based method was significantly and positively correlated with that obtained by time-consuming simulations and outperformed the simplified RSU model that did not consider perceptual constraints. The simulation approach takes approximately 10⁰ -seconds to obtain a single trajectory, whereas the DNN model achieves this in just 0.01 s, thus highlighting its significant speed advantage (16 GB of RAM, Intel^® Core^TM i7-12700H@2.30 GHz CPU, and RTX-3060 GPU). Compared with existing methods, the proposed framework considers both the perceptual constraints of the RSU and multimodal conflict risks at intersections. The proposed framework can be used to pre-estimate RSU risk monitoring at specific intersections and will be valuable in the future construction of digital transportation infrastructure.

To analyze the relationship between drivers' cognitive control and driving, clarify the differences in driving performance among drivers in different cognitive control groups, and explore the mechanism of the influential relationship of “cognitive control-vehicle control ability-driving performance,” this study recruited 49 young subjects and carried out a psychological experiment that simultaneously considered two typical cognitive control abilities: inhibitory control and working memory. Based on the test results for inhibitory control and working memory, drivers were divided into four groups: high inhibitory control and high working memory (Group A), high inhibitory control and low working memory (Group B), low inhibitory control and high working memory (Group C), and low inhibitory control and low working memory (Group D). Based on the driving simulator, a driving experiment was conducted by setting up road risk scenarios, including expected and unexpected event types. The risk event response assessment of the entire road was used to represent the overall driving performance. The average speed on a normal road section without risk events was used to represent the longitudinal control ability of the vehicle, and the standard deviation of the lateral lane position was used to represent the lateral control ability of the vehicle. The results show that the risk driving tendency of unexpected events is 15.1% higher than that of expected events. In terms of risk response assessment, Group A significantly reduces the risk driving tendency by 16.8% compared with Group D. Moreover there is no significant interaction between event type and cognitive control on the risk response. In terms of vehicle control ability, the average speed of Group D is a significant 16.2% higher than that of Group A, whereas there is no significant difference in the standard deviation of the lateral lane position among each group. The vehicle longitudinal control ability plays an intermediary role between cognitive control and the risk response, thus forming a simple relationship framework of “cognitive control-speed control-risk response.” Therefore, the significant differences in risk response and average speed are both caused by simultaneous changes in the two cognitive control abilities, and drivers in different cognitive control groups mainly cause differences in risk response through longitudinal control. These findings provide references for early interventions in driving safety, cognitive training, and the design of vehicle-mounted safety systems.

Under mixed traffic of cars and trucks, differences in physical performance and driving behavior between cars and trucks can easily lead to dangerous driving behaviors, such as the sudden acceleration, deceleration, or overtaking of cars, which can affect the stability of traffic flow and increase the risk of traffic accidents. Therefore, this study focuses on car-truck mixed traffic scenarios, innovatively proposes the “Oppression of Truck” concept, and explores the lane changing behaviors and driving risks of cars under the influences of car-truck interactions. First, the oppression measurement of trucks (OM_T), considering driving style, was constructed by introducing molecular interaction forces to quantify the oppression of trucks on cars. Then, using the truck oppression measurement to optimize the car lane change intention identification method, a two-stage lane change crash risk prediction model was developed by integrating the OM_T, which comprised a lane-change intention identification model based on a CNN-LSTM (Convolutional Neural Network-Long Short-Term Memory) hybrid neural network and a crash risk prediction model for car lane changing behavior based on LightGBM. A vehicle trajectory data set consisting of data from real traffic scenarios was used to verify the validity of the model. The results indicate that lane changing cars are generally more strongly oppressed by trucks. Moreover, skilled drivers can withstand high truck pressure, whereas cautious drivers are more sensitive to truck oppression and tend to keep driving under low pressure. Additionally, there is a time-lag correlation between truck oppression and driving risk, where stronger oppression can affect vehicle driving behavior and lead to changes in driving risk. Models that incorporate the truck oppression indicator show higher accuracy in lane change intention identification and crash risk prediction. Truck oppression has a higher feature contribution in the crash risk prediction model, which provides a new perspective and effective theoretical support for the microscopic modeling of complex interaction scenarios as well as active safety management and control.

Abnormal driver behavior (ADB) is a common and unsafe driver behavior that can easily lead to traffic accidents and poses a serious threat to road safety. To comprehensively, accurately, and swiftly detect ADB, this study proposes an abnormal driver behavior identification and quantification model based on hybrid contrastive learning (ADB-HCL). The proposed model innovatively incorporates a Mix-up data augmentation strategy into the contrastive learning framework and thereby generates mixed samples of normal and abnormal driver behaviors. This expands the coverage of the driver behavior feature space, thus enhancing the ability of the model to identify unknown or rare abnormal behaviors. In addition, by generating a template feature representing normal driver behavior and calculating the distance between the test driver behavior and the template feature, the model quantifies the degree of abnormality and hence overcomes the limitations of traditional methods that provide only discrete outputs for known abnormal driver behaviors. Experimental results based on the DAD(Driver Anomaly Detection) and AIDE(AssIstive Driving pErception) datasets show that the ADB-HCL method excels at identifying unknown abnormal driver behaviors and achieves an accuracy of 86.73% with an inference time of only 10.75 ms, which represents a 6% to 15% improvement over existing methods. The quantification results of driver behavior abnormalities indicate that this method enables the fine-grained quantification of abnormal driver behaviors. The findings demonstrate that ADB-HCL has significant advantages in terms of the comprehensiveness, accuracy, granularity, and speed of detecting abnormal driver behaviors, this showcasing the potential applicability of ADB-HCL in vehicle active safety technologies.

Compensatory behavior is crucial for drivers to balance their workloads and driving tasks, maintain driving performance, and enhance safety in complex traffic scenarios. To investigate the characteristics of drivers' compensatory behaviors under high-workload scenarios and quantify the effectiveness of compensatory behaviors in enhancing driving safety, a driving simulation experiment with a combination of 3 secondary tasks and 3 follow-up scenarios was designed, during which 42 subjects were recruited to perform driving trials. First, six indicators, including the braking deceleration, time headway, and subtask completion rate, were selected to construct an index mechanism for measuring compensatory behavior, and a multiway analysis of variance was performed. Then, a safety margin index was extracted, and a generalized linear mixed model for predicting the collision risk was constructed using the logit connection function. Finally, the drivers' compensatory behavior positivity was ranked using a technique for order preference based on similarity to an ideal solution, and the drivers were categorized into two groups using a clustering method. Subsequently, a driving safety evaluation model was constructed based on the extreme gradient boosting model, which used the optimization algorithm to find the optimal hyperparameters, including the maximum depth, learning rate, and number of trees. The results show that subjects adopt compensatory strategies such as decreasing speed, increasing headway, elevating effort, and interrupting subtasks in high-load scenarios. The model results verify the effectiveness of the compensatory behavior. By increasing the headway by 46.2% and 49.5% under the general and emergency braking scenarios of the front vehicle, respectively, drivers can maintain the same safety status. The model recall and accuracy increased by 7.13% and 3.08%, respectively, after incorporating driver compensatory behavioral characteristic variables. These findings provide a reference for driving behavior regulation and traffic safety evaluation in complex environments.

As typical weaving sections, freeway diverge areas often experience hazardous behaviors, such as sudden braking and lane changes, that can easily lead to severe traffic conflicts. To effectively evaluate the safety levels of freeway diverge areas, this study addresses the optimization of conflict data and proposes a conflict dataset screening method as well as an extreme value modeling approach that integrates both conflict probability and severity. For conflict probability, the time difference to collision (TDTC) was utilized as an indicator to analyze three collision scenarios in freeway diverge areas and calculate the time thresholds for conflict events. Regarding conflict severity, delta-V was introduced to filter out conflict events with the potential to result in fatalities. To validate the effectiveness of the proposed fusion conflict data in enhancing the accuracy of safety assessments, the proposed dataset was compared with a traditional dataset that considered only the conflict probability. Block maximum extreme value models were constructed for both datasets to conduct safety assessments. The results show that the extreme value model constructed using the fused dataset achieves extreme value reproducibility, exhibiting a mean absolute error of 0.046 and a root mean square error of 0.058. Its estimation accuracy and fit to actual conflicting data outperform those of traditional datasets. When applied to predict the collision frequency in each diverge areas, the proposed model demonstrates enhanced evaluation reliability and more accurate accident predictions, aligns better with actual conflict situations, and improves prediction reliability.

This study investigates the wind field characteristics above a submerged rectangular bluff body subjected to the combined effects of wind and wave action. The study uses the moving measuring point method to assess these characteristics. Three heights of the bluff body above the water surface were measured, defined as dimensionless heights based on the ratio of the height of the bluff body exposed to the water surface to the wave height, where the wave height was 3 cm. The heights in ascending order were 1.67, 2.50, and 3.33. Wind field measurements were taken at the same height as the bluff body at three downstream positions, with dimensionless lengths defined as the ratio of the distance from the leading edge of the bluff body to the wavelength of 76.50 cm, in descending order of 0.05, 0.10, and 0.15. The results indicate the following: ① The wind speed above the bluff body exhibits an obvious acceleration effect, which becomes more pronounced as the height of the bluff body increases at the same downstream position and measurement height; ② The wind speed acceleration effect is more significant when the measurement point is closer to the leading edge of the bluff body, at the same height above the water; ③ A comparison of the wind speeds at three different heights of the bluff body with those without the bluff body reveals that the acceleration effect of the bluff body on wind speed reaches its maximum when the ratio of the length from the leading edge of the bluff body to the wave height is 0.05, with acceleration ratios of 23.37%, 27.53%, and 38.14%, respectively; ④ Turbulence intensity is found to be only weakly influenced by the wind speed, where a higher turbulence intensity is observed when the bluff body is exposed to the water surface, particularly near the wall at the same downstream position.

Floating bridges have good application prospects in deep-water areas with weak soil foundations, and the dynamic characteristics of their floating foundations are significantly influenced by the mooring system. A 1∶70 scaled hydrodynamic model test was conducted for three types of floating foundations: inclined cables, tension legs, and combination systems. The dynamic characteristics of different floating foundations were investigated and compared based on the natural frequency, amplitude response operator (RAO), and power spectral density, and the applicable wave environments of the mooring systems were specified. Based on a hydrodynamic numerical simulation, a parametric analysis of the combination mooring system was conducted, considering different inclined cable-to-tension leg stiffness ratios and the overall mooring stiffness. Furthermore, the influence of the mooring stiffness setting on this type of floating foundation was analyzed. The results showed that the inclined system had a broad frequency range of high response, and the motion coupling of each degree of freedom was significant, suitable only for weak-wave areas. For the tension leg system, the low-frequency horizontal motion was significant, which could reach 2.3 times of those of the other two types of systems under the same irregular wave. This type of system is unsuitable for water with low-frequency swells. The peak RAO of the horizontal displacement for the combined system was significant; however, a high response occurred in a narrow frequency range. By adjusting the cable stiffness, the horizontal resonance frequency could be tuned far from the wave peak frequency, which could effectively reduce the structural dynamic response. This system is suitable for a wide range of applications. This study provides a reference for future research and applications of floating foundations of deep-water bridges.

To explore the key issues and future directions in the study of asphalt pavement textures, this paper provides a comprehensive review of the research progress in three-dimensional (3D) reconstruction and evaluation methods of asphalt pavement textures, both domestically and internationally. First, 3D reconstruction methods of asphalt pavement textures are systematically reviewed. The 3D reconstruction techniques for asphalt pavement surface tomographic images are summarized, and active laser scanning and passive image-based methods for the 3D reconstruction of pavement surface textures are introduced. The algorithm principles and characteristics of the monocular, binocular, and multi-view image depth estimations are compared, and the application of pavement texture 3D generation technology is analyzed. Subsequently, based on computational principles, the pavement surface texture evaluation methods are categorized into the geometric statistical index method, spectral index method, fractal index method, and image feature method, and the corresponding evaluation indices are further classified into 2D evaluation indices based on the pavement surface profile and 3D evaluation indices based on the pavement surface texture. Multidimensional and multiscale analyses are conducted on the applicability conditions, advantages, and limitations of the different evaluation indices. Finally, future research directions concerning the evaluation and reconstruction techniques of pavement surface textures are discussed, and the development trends of intelligence, digitization, and informatization in the evaluation and 3D reconstruction of asphalt pavement textures are anticipated. This study provides a reference and source of inspiration for academic research on the functionality of pavement surfaces and development of modern high-quality pavements.

Soil arching is a primary load-transfer mechanism in pile-supported embankments. For embankments supported by a composite foundation with rigid friction piles, cement soil piles, and gravel piles, the settlement of the pile head under embankment fill and localized vehicle loads significantly impacts the soil arching of the embankment. However, traditional trapdoor models usually use fixed supports, which do not account for the influence of pile head (or arch foot) settlement on the soil arch shape and load transfer of the embankments. To address this, a multispan spring-based trapdoor test apparatus with movable arch feet was designed in this study. The vehicle loads were simulated with localized surface loads, and movable blocks with springs of varying stiffnesses were used to model the piles and the soil between the piles. Fifteen orthogonal spring-based trapdoor tests and one reference test with a fixed trapdoor were conducted. Under the self-weight of the fill and localized loads, the earth pressure in the affected area of the soil arching and the displacement of the arch foot and trapdoor were monitored using earth pressure cells and laser displacement meters, respectively. The evolution process of the soil arch morphology was reproduced using digital image correlation (DIC) technology. The test results show that the arch foot settlement weakens soil arching. The weakening effect varies across different arching states and becomes more prominent with an increase in trapdoor stiffness. Additionally, the arch foot settlement under the fill weight significantly decreased the soil arching height, restricted the upward propagation of the differential settlement within the embankment, and expanded the affected width of the soil arch, causing a transformation from a steep to a flat soil arch. Under a localized load, arch foot settlement significantly reduces the stability of the mid-span soil arch, promotes load transfer to the adjacent span, and mobilizes the soil arching of the adjacent span to carry the load.

Prefabricated bridge deck panels (PBDPs) are easily constructed, are associated with lower pollution levels, and cause minimum disruptions in existing traffic and the surrounding environment. Owing to these advantages, they are gradually becoming the research focus of bridge engineering. Configurations and materials of joints have significant effects on the speed of construction, integrity, and durability of PBDPs. Wet joints, with their small amounts of grout, high tolerance for construction errors, and superior mechanical properties, have gradually become commonly used joint types for PBDPs. Herein, previously published research studies on wet joints for PBDPs are systematically summarized. The configurations and performance of common wet joints are introduced first. The designs and behaviors of wet joints with high-performance materials are then clarified. The prediction models for the capacity of bar connections and post-tensioned connections subjected to bending moments, shear, and tension are discussed. The practical engineering values of the wet joints are demonstrated based on engineering practices. Finally, the development trends and future research directions of wet joints for PBDPs are pointed out. Based on high-performance materials, the novel configurations of the wet joints associated with superior performance, simple and convenient construction approaches, and economic efficiency, constitute the premise of future studies. Working mechanisms, failure modes, and performance prediction models of novel wet joints subjected to bending moment, shear, and fatigue loading are the basis for future works. In addition, a practical engineering-oriented design method of wet joints for PBDPs is the focus to be investigated. This paper can provide a broad view and technical support for the application of PBDPs.

Under the backdrop of rapid intelligentization and informatization development in bus systems, bus en-route operation control has gradually become an important research highlight in bus operation optimization. Such research helps to enhance the efficiency of bus operation and passenger travel experiences, and it provides strong support for the development of the public transportation industry. To analyze the problems and challenges in bus en-route operation control, this study systematically reviewed and analyzed the status and development achievements of relevant research in terms of research direction, scope, and practical application to provide a clear research context, theoretical framework, and methodological guidance for subsequent researchers. This study followed the research approach of “state diagnosis-control optimization”, considered a bus bunching incident as an example of a poor en-route operation state, provided a focused summary of the influencing factors that lead to such incidents, and addressed other possible poor en-route operation states as well as related identification and diagnosis technologies. Hence, a comprehensive examination of the methodologies employed in bus en-route operation control is undertaken. First, the methodologies and limitations of existing research on the optimization objectives, control objects, and constraint conditions are analyzed. This is followed by a comprehensive analysis of the control strategies from disparate dimensions, including stop control and interval control, as well as an examination of the characteristics and applicable scenarios of the various control strategies. Furthermore, this paper presents a summary of commonly used optimization model-solving algorithms and discusses their applicable conditions and optimization effects. Finally, this paper offers a prospective outlook on future research directions and development trends to provide new ideas and system optimization directions for future work to improve bus operation efficiency and service quality.

To address the issues of high testing costs and poor accuracy caused by individual differences in subjective perceptions and objective physiological information of passengers in traditional motion sickness comfort evaluations for intelligent electric vehicles, a motion sickness evaluation model based on vehicle motion parameters is proposed. First, based on the mechanism of motion sickness, real-vehicle tests were conducted to collect passengers' subjective perceptions of motion sickness along with corresponding physiological data, such as galvanic skin response (GSR), respiratory rate, and pupil diameter. Simultaneously, vehicle motion parameters, including longitudinal, lateral, and vertical accelerations, were collected. Second, passengers' subjective assessments of motion sickness comfort combined with objective physiological data to analyze significant differences and degrees of association using the Kruskal-Wallis (K-W) non-parametric test and partial effect equivalence methods. A motion sickness evaluation model was established based on multiple objective physiological signals such as GSR and pupil diameter variations. Furthermore, Pearson correlation analysis method was employed to construct a correlation matrix linking subjective motion sickness perceptions, objective physiological signals, and vehicle motion parameters. The relationships and sensitivities of vehicle motion parameters, including longitudinal, lateral, and vertical accelerations, and their rates of change with passenger motion sickness on real-world open roads were explored. Using a ridge regression analysis, weights for the impacts of different vehicle motion parameters and cumulative time on motion sickness comfort were determined, facilitating the development of a motion sickness evaluation model based on vehicle motion parameters. Finally, experimental validation was conducted to compare the motion sickness evaluation model developed based on multiple physiological signals with that based on vehicle motion parameters. The results showed that the overall prediction accuracy of the motion sickness evaluation model developed based on the vehicle motion parameters was 88.7%. The proposed evaluation model effectively mitigates the impact of individual physiological differences in traditional testing and achieves accurate comfort evaluations for intelligent electric vehicles. This study provides theoretical support and a practical foundation for the design and optimization of decision-making algorithms and control execution strategies for future intelligent electric vehicles.

With the continuous promotion of strong transportation strategies, the demand for high-quality sand and gravel materials in China's transportation infrastructure remains high. Environmental protection requirements continue to grow, and the shortage of natural resources continues to exacerbate. Therefore, alternatives to sand and gravel aggregates have become the main direction of development in the field of road engineering. China's western region has long been plagued by wind and sand problems, and successive exploratory studies have been conducted on the application of aeolian sand in road engineering, which have confirmed the significance of aeolian sand resource utilization for promoting green and low-carbon transportation in the sustainable development of road engineering industry. Therefore, this study focuses on the problems of large regional differences in aeolian sand and the lack of corresponding standardized research. The physicochemical properties of aeolian sand and its engineering characteristics in different regions are systematically discussed, analyzing its potential activity, excitation mode, and mechanism. The progress of research on the application of aeolian sand in roadbed and pavement engineering is summarized, thereby illustrating the influence and mechanism of aeolian sand on the performance of pavement concrete and semi-rigid bases, comprehensive utilization and treatment technology of aeolian sand in roadbed engineering, prevention and control of wind-blown sand of aeolian sand roadbeds, and performance of aeolian sand roadbeds in resisting scouring and water damage. Finally, development trends in the application of aeolian sand and research focus directions for the future of road engineering are presented.

Transportation sector is one of the major sources of global carbon emissions, and as a major transport country, China is facing a huge challenge to reduce transportation carbon emissions. Researchers have conducted extensive research on carbon emission assessment and emission reduction measures in recent years. This paper focuses on the highway construction stage, reviewing previous studies on three aspects: carbon emission assessment, carbon reduction, and carbon sequestration compensation. The results show that: most carbon emission assessment methods use the Life Cycle Assessment(LCA) model throughout the highway construction stage, in which the basic data accounting focuses on the emission factor method, as well as software, platforms, and other tools combined with artificial intelligence; emission side of the carbon reduction is mainly centered on green geo-technology, solid waste resource utilization technology, and green-efficient project management, to completely accomplish the carbon reduction goal; carbon compensation research includes compensation measures and effect evaluation of the carbon sequestration, and its carbon compensation measures are mainly in using slope vegetation photosynthesis to sequester carbon and new materials to increase carbon sinks. Moreover, carbon sequestration effects are evaluated using the subjective-objective combination analysis method. The analysis reveals that unsolved problems still exist: lack of unified standards for emission factors, leading to considerable errors in carbon emission assessment; lack of comprehensive assessment methods for the specific emission reduction ratio of emission reduction measures and the poor combination with artificial intelligence; the carbon compensation measures are largely insufficient and the carbon sequestration effect assessment method is subjective. Future research should focus on exploring the accuracy of the assessment method; proposing more effective carbon emission reduction measures and enhancing the research on the assessment of emission reduction effect; adopting more diversified carbon compensation measures, and establishing a universally applicable compensation effect assessment method. This review paper provides a comprehensive reference for carbon emission reduction research in transportation geotechnical field and assists in realizing the goal of the “carbon peaking and carbon neutrality” strategy.

Load-carrying capacity is a key index reflecting the service performance of highway bridges. Its reliable evaluation is key for determining the safety risk of bridges in advance, which can provide an important basis for decision-making, such as that pertaining to bridge reinforcement and reconstruction. In this study, the connotations of bridge load-carrying capacity evaluation were analyzed from the aspects of evaluation targets and evaluation methods, the current development of technical standards for load-carrying capacity evaluation in the United States and China was introduced, and the implementation method and evaluation theory of proof and diagnostic load-testing methods were demonstrated. Proof load testing can directly determine whether the bridge load-carrying capacity satisfies the requirements, whereas diagnostic load testing should be further developed in the future as it offers more bridge information and ease of implementation. Additionally, in this study, the verification theory used for the load-carrying capacity in the codes of various countries was systematically compared; three types of resistance correction methods, including direct structural testing, state-parameter mapping, and a time-varying deterioration-prediction model, were summarized; the correction methods for the bridge load effect were described comprehensively in terms of load and bridge models; and the reliability targets of bridge evaluation in various countries were analyzed. The British and American codes have established a multilevel evaluation reliability target by discounting the design load and resistance factor, which is worthy of reference for the Chinese bridge evaluation code. Finally, an outlook into the future development of load-carrying capacity evaluation methods in terms of testing methods, evaluation theories, and technological equipment is presented.

The reinforcement effect of prestressed anchor cables with small diameter on the surrounding rock of soft rock tunnels was systematically investigated. First, the bearing arch effects of the surrounding rock of a tunnel strengthened using small-diameter prestressed anchor cables were simulated and analyzed based on a stratum-structure model. A generalized “load-structure” mechanical analysis model of the anchored surrounding rock was established. A formula for calculating the bearing capacity of the anchored surrounding rock with the combined support of long and short anchor cables was derived. Subsequently, numerical analysis of the simulated loading of the anchored surrounding rock was performed. The development process of a plastic zone in the anchored surrounding rock and the main factors influencing the ultimate bearing capacity were studied. The effectiveness of anchor cable support schemes was also explored. Finally, the reinforcement effect of small- diameter prestressed anchor cables on the surrounding rock of a soft rock tunnel was verified and summarized through on-site testing. The results indicate that a superimposed arch composed of a shallow and a deep bearing arch is formed in the surrounding rock under the combined support of long and short anchor cables. In this case, the diffusion range of the pre-tensioning force is higher than that obtained with the short anchor cable scheme, and the engineering economy can be considered comparable to that of the long anchor cable scheme. After the anchored surrounding rock is loaded, its inner surface first enters a plastic state. Considering the corresponding load when the inner surface enters the plastic state as the bearing capacity of the anchored surrounding rock tends to be conservative. The ultimate bearing capacity of the anchored surrounding rock can be obtained using numerical calculation methods, which mainly depend on the strength of the surrounding rock and the anchoring force of the anchor cable. The active support obtained with small- diameter anchor cables and high pre-tension can significantly increase the overall stiffness of the anchored surrounding rock. Under the conditions of relatively soft rock and soft rock strata, small- diameter (Φ21.8) prestressed combined long-short anchor cables (5 m+10 m, 19 per ring of upper and middle benches, spacing of 80 cm, design anchoring force of 450 kN, design pre-tensioning force of 350 kN) were used, and the maximum deformation of the tunnel was basically controlled within 30 cm according to actual measurements. For extremely soft rock formations, ensuring that small-diameter anchor cables have a sufficient anchoring force is a key technical problem that must be solved urgently.

To systematically analyze and summarize the current research status and development trends in the operational management of emergency medical service (EMS) vehicles, this study organizes the research framework of EMS vehicle operational management into three levels: strategic, tactical, and operational, based on 1 502 articles indexed from the Web of Science database. The findings reveal that at the strategic level, research on EMS vehicle location focuses on continuous improvement of coverage definition and accurate characterization of inherent uncertainties within the system. Key research methods include stochastic planning and robust optimization as uncertainty modeling and optimization approaches. At the tactical level, EMS vehicle relocation is categorized into multiperiod and dynamic relocation based on the triggering of relocation decisions. Given the complexity of relocation with respect to location, the research emphasizes the application of heuristic and reinforcement learning algorithms in addressing real-world large-scale problems. Decisive issues at the operational level include EMS vehicle dispatch, destination selection, and route planning. Research on EMS vehicle dispatch has evolved from rule-to model-based and from independent to joint optimization in relocation. Destination selection involves coordinated optimization with hospital workload, and route planning primarily addresses special scenarios such as disaster response. In future research, optimization in EMS vehicle operational management should focus on the dual research threads of dynamics and uncertainty. This entails accurately characterizing the sources of system uncertainty while leveraging finer-grained data to assist real-time decision-making. In terms of specific modeling and solving techniques, joint optimization of multiple decision problems across different levels should be conducted to progress from local to system optimum EMS vehicle location and dispatch schemes. However, efficient algorithms for handling real-world large-scale scenarios continue to pose a challenging research direction.

Recently, autonomous vehicles (AVs) have gradually reached the stage of real-world testing and demonstration on public roads. AVs interact frequently with humans, and their driving performance requirements have shifted from a functional level of “safe and stable driving” to an interactive level of “driving like a human.” The construction of evaluation indices and methods that can accurately characterize the human cognition of driving behavioral ability is necessary to guide the continuous development and improvement of automated driving technology to demonstrate human-like or superhuman driving ability. This study focused on the construction of an evaluation indicator system for the driving behavioral ability of autonomous vehicles. This paper first elaborates on the definition and boundaries of the driving behavioral ability of autonomous vehicles. Then, the current status and existing problems of the evaluation index system for driving behavioral ability are presented. Moreover, an STCER-H index system is proposed, which includes five dimensions of instantaneous indicators: Driving Safety, Travel Efficiency, Driving Comfort, Energy Efficiency, and Regulatory Compliance, as well as a comprehensive “Humanoid Level of Driving Behavior” indicator. Consequently, the existing modeling methods for individual indicators are reviewed, the connotations of each dimension indicator are clarified, and the current status and problems of each dimension indicator are summarized. At the level of statistical evaluation in multidimensional aggregation, the definition and modeling suggestions of “Humanoid Level of Driving Behavior” indicator are primarily discussed. Finally, a summary of the challenges and future research prospects of the existing evaluation index system for driving behavioral ability is presented to provide a reference for academic and industry research.

Driver state monitoring technology, as a key means for improving vehicle intelligence and safety, aims to accurately identify and deeply understand the driver's actions, emotions, and attention states. Although significant progress has been made in this field, a systematic summary of the principles of deep learning algorithms is lacking. In view of this, this paper systematically reviews driver state monitoring algorithms based on images and deep learning to meet the needs of the continuous development of intelligent vehicle technology. First, the methodology in the literature is elaborated upon. The existing publicly available datasets are then organized and described. Subsequently, in-depth exploration is conducted from the aspects of data selection and processing, model architecture, model training and evaluation, and optimization. Finally, the shortcomings of the current research are summarized, and the main future development directions are outlined. The results show that: ① the research on driver state monitoring based on image and deep learning has progressed to a certain depth; ② data selection and processing techniques show diversity; ③ model architectures continue to evolve in the direction of multi-modal, multitasking, lightweight, and high robustness, gradually beginning to adopt training strategies for incomplete supervision and multi-objective optimization. However, most research methods lack systematic testing of actual driving scenarios neither fully considering the behavioral characteristics of drivers under natural driving conditions nor the changes in the human-computer interaction patterns of intelligent vehicles, making it difficult to construct an all-around monitoring function for various driving scenarios and driver personalities. The further development of driver state monitoring algorithms is mainly limited by two factors. First, the current deep learning methods still have deficiencies in their domain adaptation, interpretability, and operational efficiency. Second, large-scale high-quality datasets under natural driving environments are lacking. This review is dedicated to providing effective guidance and important references for further development of high cognitive driver state monitoring systems.

In recent years, the pace of bridge construction in China has been steadily accelerating, with the scale of projects and technical level reaching world-leading standards. The high-quality and innovative development of bridges is an important starting point and basic prerequisite for building China into a country with strong transportation network. To further enhance the strength of the bridge engineering discipline in China, promote the high-quality development of green low-carbon, sustainable and intelligent bridge engineering in China, and support the construction of a transportation powerhouse, this review, based on the analysis of the current industry development status and trends, systematically summarizes the latest scientific and technological innovation achievements in the field of bridge engineering in China in recent years and comprehensively sorts out the future key development directions covering four major themes: bridge engineering structural design and system innovation, disaster prevention and mitigation and structural safety, green construction and intelligent construction, and healthy operation and maintenance and longevity assurance. Specifically, it covers 21 hot research directions, including bridge function and analysis, high-performance materials, steel bridges and composite structure bridges, long-span bridge structures, innovative bridge foundation structures, new progress in bridge seismic resilience research, bridge wind, fire, and blast resistance, bridge engineering collision and protection, water resistance and resilience, multi-hazard coupling of bridges, high-quality bridge construction, green construction technologies and construction technologies, bridge monitoring and assessment, intelligent detection, on-bridge traffic safety, bridge life extension technologies, and integrated construction and maintenance platforms. The review provides guidance and reference for the development of the bridge engineering discipline in China and offers new perspectives and basic materials for researchers and technicians in the field of bridge engineering.

In the process of rapid development of road engineering, the problems caused by subgrade permanent deformation (PD) haven't been completely solved. Clarifying the evaluation and control methods for subgrade PD under long-term cyclic loading can ensure the durable and stable operation of road engineering. Firstly, this paper explored the main conditions and setting methods for PD test of subgrade soil. Subsequently, the constitutive models based on classical soil mechanics and empirical models based on experimental phenomena were sorted out. Next, the calculation process, verification methods, and evolvement rules of subgrade PD were summarized. Then, three methods for controlling subgrade PD were discussed, including critical dynamic stress, structural measures, and failure probability. Through analysis of research progress, it is found that there are four main problems with subgrade PD, namely inaccurate test methods, incomplete prediction models, unreasonable calculation theories, and unclear control standards. The specific problems and potential challenges in each aspect are elaborated in detail. Four prospects for future research are also given. Firstly, it is necessary to establish a static earth pressure coefficients database of subgrade soil to form a unified test method for PD of subgrade soil. Secondly, the influence rules and internal mechanisms of loading action duration and intermittent duration on PD of subgrade soil should be clarified, and the mechanical model for PD of subgrade soil should be derived under the theoretical system of element model and fractional-order calculus. Thirdly, the calculation method of subgrade humidity field considering the influence of dynamic loading should be innovated, and then the fully coupled calculation method of subgrade PD under humidification action based on the mechanical model for PD of subgrade soil should be established, and a comprehensive verification platform of subgrade that can scientifically simulate the climate environment and stress state should be developed. Fourthly, a control standard for subgrade PD based on pavement performance requirements should be determined with reliability as the goal, and then the corresponding relationship between structural failure and material deformation should be quantified, and the granular materials improvement layer structure design method for subgrade performance control should be optimized.

Uncertainty characterization of surrounding rock parameters is the fundamental cornerstone of tunnel long-life design, and the key is to obtain sufficient accuracy under limited data samples. To address this, a novel method for uncertainty characterization of surrounding rock parameters has been proposed by combining the Bootstrap method and the Akechi Information Criterion (AIC), studying the minimum sample size required to obtain sufficient accuracy. Firstly, the mean and standard deviation of surrounding rock parameters was obtained by the Bootstrap method. Secondly, the probability distributions of the sample under this resampling size were identified by the AIC. Thirdly, the confidence intervals for the mean and standard deviation of the parameters with a confidence level of 95% were calculated. Subsequently, the minimum numbers of samples required for an accuracy of 90% were determined. By this way, the curacy of the uncertainty characterization of surrounding rock parameters was ensured. The proposed method was illustrated through Hoek's classical weak rock parameters. Results indicated that the minimum sample sizes for the mean and standard deviation of weak rock parameters are 12 and 22, respectively. These minimum sample sizes derived from the proposed method were validated by real data of weak rocks from two different places, and the results agreed well with the real data. Furthermore, by incorporating the triple standard deviation criterion, this proposed method was applied to conduct uncertainty characterization of surrounding rocks for the third and fourth level rock mass in the rock mass classification standards. The minimum number of samples for weight, deformation modulus, cohesion, internal friction angle, and Poisson's ratio, were obtained. These could provide valuable insights for the uncertainty characterization of surrounding rock parameters in engineering practices, which in turn would aid in tunnel reliability assessments and long-term design considerations.

Existing deep-learning-based methods for structural damage identification rely heavily on massive amounts of labeled data. Therefore, a meta-learning-based approach is proposed for structural damage localization and quantification. First, a structural damage localization and quantification model was established using an artificial neural network. This model was used to learn the nonlinear mapping relationship between structural modal data (frequency and mode shape) and substructure stiffness parameters. Second, a model-agnostic meta-learning strategy was used to train the damage localization and quantification model. The generalizability of the damage localization and quantification models can be improved by optimizing the initial weight parameters of the artificial neural network (ANN). The proposed method utilizes a model-agnostic meta-learning training strategy to acquire prior knowledge, thereby accelerating the learning process for new structural damage localization and quantification tasks with limited training data. The method was verified on a numerical three-span bridge and benchmark project of the Z24 bridge. The results demonstrate that the proposed approach provides efficient and accurate localization and quantification of potential structural damage using limited data. Compared with conventional ANN and transfer learning methods, the method exhibited faster convergence and higher identification accuracy.

Rock freeze-thaw damage is a crucial issue in cold-region tunnel-engineering research. To better understand the mechanical properties of cold-region rocks and the microdamage caused by the freeze-thaw action, compression, acoustic-wave, and computed tomography (CT) scanning tests were conducted on granite under the freeze-thaw action. The physical and mechanical parameters of the rock and the microdamage characteristics were obtained. Based on three-dimensional reconstruction, a quantitative analysis of pore evolution was performed, which reveals the mechanism of frost damage in the cold-region tunnel surrounding rock. Based on continuum damage theory and microelement statistical theory, a mechanical damage constitutive model considering the initial freeze-thaw damage and residual deformation was derived. The results show that after 50 freeze-thaw cycles, the longitudinal wave velocity of the specimen decreases by 16.60% and the total porosity increases from 7.98% to 10.01%. The linear elastic modulus, peak stress, and residual strength decrease as the peak strain increases. The freeze-thaw action can enhance the development of connectivity between pores, intensify seepage effects, and increase the probability of rock ductile failure, thereby exhibiting clear softening characteristics. The parameters of the new constitutive model can be determined easily and present clear physical significance, high accuracy, and practicability. The model is suitable for describing the stress-strain relationship of the frost-rock damage process and for reflecting the residual-strength characteristics of rocks. The results of this study provide theoretical guidance for the service-performance analysis of cold-region tunnels.

Scientific and reasonable suspender maintenance policies play a major role in ensuring the safe operation of cable-supported bridges. This study addressed the decision-making challenges associated with the maintenance and replacement of vulnerable suspenders by considering their appearance and structural damage state. Accordingly, a preventive maintenance decision-making method that minimizes the combined costs of maintenance and risk throughout the bridge's lifecycle was proposed. First, an optimization objective function was constructed based on the maintenance decision-making problem. The suspender service context was defined as the environment, and the bridge operation and maintenance management system acted as the agent. In addition, the state space, action space, state transition probability matrix, and reward function were established. The expectation of the cumulative discount reward replaced the objective function of the maintenance optimization problem, and state prediction and maintenance decision models based on the Markov decision process were constructed. Then, a preventive maintenance decision method for the suspender system was established based on the suspender system maintenance decision model and dueling double deep Q-network (D3QN) algorithm, which incorporates both a target network and an experience replay mechanism. Finally, a maintenance decision-making framework for the suspender system was constructed using the state prediction model and preventive maintenance decision-making method. With a suspension bridge used as a case study, the state prediction model enabled continuous interaction between the agent and environment, simulating the degradation and maintenance processes of the suspenders while generating the necessary data for training the neural network. Based on the interaction data, the D3QN algorithm network model was trained to obtain the optimal maintenance policy, which was then compared with traditional policies. The results show that the proposed method comprehensively considers the maintenance cost and structural risk and dynamically and adaptively adjusts the maintenance policy. Compared with the traditional policy, the maintenance cost of the policy obtained under the proposed method can be reduced by more than 12%.

Under out-of-plane deformation or high fatigue stresses, conventional crack-stop hole repair of fatigue cracks in steel structures is susceptible to crack perforation (i.e., secondary crack initiation), thereby resulting in unsatisfactory fatigue strengthening. This study proposes cold-expanded crack-stop hole technology for repairing fatigue cracks in steel structures. This principle involves a cold-expanded crack-stop hole using a mandrel to induce residual compressive stress around the hole, thereby reducing the fatigue stress level and extending the fatigue life. Cold expansion tests of crack-stop holes and fatigue tests were conducted on steel plates with type Ⅰ cracks. The distributions of residual strain around the hole after cold expansion were obtained, and the evolution of the residual strains during fatigue loading was clarified. The effects of cold expansion rates (0%, 1%, and 2%) and hole-to-crack tip distances (0, 5, and 10 mm) on the fatigue performance of steel plates with type Ⅰ cracks were investigated, and the life-extending mechanisms for the cold-expanded crack-stop holes were revealed. The results indicate that, for a specified hole-to-crack tip distance, increasing the cold expansion rate can enhance the fatigue life. Increasing the cold expansion rate (not exceeding 2%) can extend the distribution range and value of residual compressive stress around the cold-expanded crack-stop holes. Increasing the hole-to-crack tip distance reduces the improvement of fatigue life by cold-expanded crack-stop holes. The maximum fatigue life can be obtained when the cold expansion rate and the hole-to-crack tip distance are 2% and 0 mm, respectively; Compared with the case of conventional crack-stop hole specimens with a hole-to-crack tip distance of 0 mm, the total fatigue life of the specimens increased by 50.82%. Further increasing the cold expansion rate may potentially enhance the fatigue life, which needs more studies. Finally, a model for predicting the initiation life of fatigue cracks around cold expanded crack-stop holes based on nominal S-N curves and fatigue notch factors was proposed. Analysis results show that the predicted values agree well with experimental results in the log-log coordinate system, particularly when the hole-to-crack tip distance is 0 mm, with an error of ±5%; for hole-to-crack tip distances of 5 and 10 mm, the error is ±20%.

This research investigates the viscoelastic properties of subgrade soil, which cause the subgrade to show significant differences in dynamic resilient modulus (M_R) when subjected to loads of different duration. To accurately predict the M_R of subgrade soil, this study employed an improved dynamic triaxial test method to investigate the relationship between M_R and factors such as load duration, confining pressure, and cyclic deviator stress. Two typical subgrade soils with high liquid limit silt and low liquid limit clay were selected for this study, and specimens with different working conditions were prepared for M_R testing. Subsequently, the influence of different factors on the M_R was analyzed. Analysis of the test results shows that as load duration increases, the M_R of both soil samples gradually decreases. Moreover, the M_R under different load durations displays different sensitivities to cyclic deviator stress. Grey relational analysis was then applied to assess the impact of factors, such as load duration, cyclic deviator stress, and confining pressure on the M_R. Subsequently, combined with the Kelvin model, a comprehensive viscoelastic M_R prediction model was established considering the compaction degree, moisture content, stress state, and load duration. Finally, the test results of other subgrade soils were used to verify the established prediction model and compared with conventional models that did not consider viscoelasticity. The validation results show that the newly established M_R prediction model, which considers the viscoelastic properties of subgrade soil, has high accuracy and applicability. The research results provide valuable references for subgrade design and engineering practices.

The purpose of this study was to develop a new application of energy-absorbing materials in the road maintenance field and to produce a preventive maintenance seal that can improve the road surface function and enhance the structural bearing capacity of existing roads. A new road energy-absorbing material was used as the matrix, and a “sandwich” structure was used as the framework. A road maintenance energy-absorbing seal was designed and prepared, considering texture reconstruction. Image-processing analysis and accelerated loading tests were performed to analyze the decay law of the surface texture characteristics and road surface function of the road maintenance energy-absorbing seal and evaluate the durability of the seal. The effect of the road maintenance energy-absorption seal on decreasing the strain at the bottom of an asphalt concrete plate was evaluated using a continuous loading test of wheel rolling, and its load-bearing and buffering effects were investigated. Based on dynamic thermomechanical analysis, the microenergy-absorbing characteristics and damping behavior of the road maintenance energy-absorption seal were described, and its buffering mechanism was revealed. Finally, this study lays a solid foundation for the extensive investigation and promotion of the road maintenance energy-absorbing seal. The results show that the aggregate coverage rate is 40%, based on the seal surface texture and surface functions (wear resistance and sliding resistance). The ratio between the 2.36-4.75 mm and 1.18-2.36 mm aggregates is 25∶75. The spraying plans for the energy-absorbing material are 1.0 and 2.0 kg·m^－2 for the upper and lower layers, respectively. After 40 000 cycles of loading and wear cycles, the surface texture of the road maintenance energy-absorbing seal attenuated slightly, and the decline in the durability of its surface was evident. A road maintenance energy-absorbing seal can effectively reduce longitudinal and transverse strains at the bottom of an asphalt concrete plate. Moreover, it can convert the original tensile strain into compressive strain or decrease the value of the original tensile/compressive strain by over 30%-50%. The loss factor [tan(δ)] of the energy-absorbing seal is 0.1-0.3, and the seal can exhibit excellent damping performance within wide ranges of temperature (－50 ℃-200 ℃) and frequency (10^－4-10⁸ Hz).

A long short-term memory (LSTM) neural network model for predicting slope displacements based on maximum mutual information coefficients (MICs) and the XGBoost algorithm (MIC-XGBoost LSTM) was established to accurately predict slope displacements. First, the effects of different rainfall conditions on the slope were investigated. The maximum MIC was used to analyze the correlation between different rainfall conditions and the cumulative displacement of the slope, and the rainfall-influencing factors with significant correlations were determined. Next, based on the XGBoost algorithm, feature construction was performed on the influencing factors with high correlation using the cumulative displacement data of the slope, and the construction features were normalized with the original features. The normalized data were divided into training and validation sets. LSTM was used to predict the displacement of the Shangluo rock slope on the G312 National Highway. The XGBoost, LSTM, and MIC-XGBoost-LSTM prediction models were used to train and predict the cumulative displacement value of the slope, and the prediction accuracy was evaluated based on the root-mean-square error (RMSE), mean absolute error (MAE), and mean absolute percentage error (MAPE) indicators. In addition, the RMSE was used to determine the longest prediction cycle and minimum training sample size for the MIC-XGBoost LSTM model. Finally, the displacement data of the Baishui River landslide were used to further validate the model. The results show that the correlations between daily displacement increment, evapotranspiration, net rainfall, cumulative seven-day rainfall, and cumulative displacement at the monitoring point are higher than those of other factors, and the MIC of the feature values constructed using four related factors and the output feature values is 0.97. The RMSE, MAE, and (MAPE) of the predicted results obtained using the MIC-XGBoost-LSTM model are 0.25%, 0.185%, and 0.024%, respectively, which are lower than those of XGBoost and LSTM. Based on the RMSE, the longest prediction cycle and minimum training sample size of the MIC-XGBoost-LSTM model are 56 and 675, respectively. Finally, the displacement data of the Baishui River landslide were used for verification. The evaluation indicators are lower than those of the XGBoost and LSTM models, demonstrating that the MIC-XGBoost-LSTM slope displacement prediction model has high reliability.

A system reliability analysis model is developed using an improved vector projection response surface method (IVPRSM) to enhance the efficiency and accuracy of calculating the static system dependability of long-span continuous steel-truss bridges. A novel approach was developed to enhance the efficiency of computing the component dependability index by optimizing the sampling process of the vector projection response surface technique. The limit state functions of steel truss bridge components can be effectively rebuilt, the design points can be promptly searched, and the failure probability of the components can be evaluated using the IVPRSM. The candidate failure components can be screened using the β-unzipping method. Considering the structural topology model was modified by assuming failure in the potential failure elements, the reliability indices of the remaining components were calculated using IVPRSM. Therefore, identifying the primary failure modes of steel-truss bridges and constructing a fault tree is possible. The structural system reliability index was determined using the differential equivalent recursion technique, which relied on the equivalent linear functions of the failure modes and the correlation coefficients between the failure modes. The efficacy and precision of the IVPRSM were confirmed through a reliability study of three numerical arithmetic cases. Considering a double-deck continuous steel truss bridge with a main span of 300 m as an engineering example, the proposed system reliability analysis method was used to calculate the reliability indices of the key components of the steel truss bridge in each failure stage. The study findings demonstrate that IVPRSM exhibits superior computational efficiency and accuracy compared with conventional approaches. At the ultimate limit condition of the load-carrying capacity, the reliability index for all types of critical members ranges from 4.1 to 4.8. The lower chord at the pivot of the main span of the truss girder, the upper chord and web member at approximately 1/4 of the main span (300 m), and the strengthened vertical bar in the center of the main span pose a significant danger of failure. Consequently, twenty primary failure modes affecting the load-carrying capability of the steel truss bridge were identified, resulting in a system dependability index of 4.6. This study proposes an IVPRSM-based reliability method for the static systems of continuous steel truss bridges. This algorithm may assist in designing continuous steel truss bridges by considering system reliability.

Joint bolts reduce the longitudinal and transverse bending stiffnesses of shield tunnels. Based on the classical longitudinal equivalent continuous beam theory for shield tunnels, as well as considering the reduction effect of the transverse bending stiffness of shield segments and the combined action of axial force and bending moment, this study proposes a model for calculating the longitudinal bending stiffness of shield tunnels based on an elliptical cross-section and a strict elliptical-integration derivation. The governing equation of the proposed model is a transcendental equation, and a numerical method is used to solve the governing equation. The proposed model is compared with existing models derived from circular and elliptical cross-sections, but not strictly elliptical integration. Additionally, the effects of the transverse bending stiffness of shield segments and the material stiffness of shield segments and joint bolts on the longitudinal equivalent bending stiffness of shield tunnels are analyzed. The findings show that for a circular cross-section, the results of the proposed model are consistent with those of existing models. Existing models derived based on a circular cross-section are specific cases of the proposed model; when the geometric and material parameters are provided, the longitudinal equivalent bending stiffness of shield tunnels and the ratio between the axial force and bending moment present nonlinear and positive correlations within the range of positive and negative critical ratios between the axial force and bending moment. The longitudinal equivalent bending stiffness is constant outside the range of positive and negative critical ratios between the axial force and bending moment. Meanwhile, it decreases with the transverse bending stiffness of the shield segments or the material stiffnesses of the shield segments and joint bolts. These findings provide a theoretical reference for the longitudinal-deformation analysis of shield tunnels.

To investigate the impact of the behavioral preference and dynamic spatio-temporal relationship of pedestrian-vehicle interaction (PVI) on potential traffic conflict, a safety analysis approach incorporating machine learning and ordered logistic modeling is proposed. The pedestrian-vehicle crossing data at unsignalized crosswalks was collected through unmanned aerial vehicle (UAV). The 876 pairs of PVI were extracted from of trajectories road users. Considering the dynamics of interactions, the interaction indicators based on competition for right-of-way and collision relationships were proposed. A representation learning model was trained to capture time-series data of PVI, transformed these into potential representations, and clustered them to identify and analyze typical interaction patterns and features. Utilizing these patterns and varying conflict severities, multiple ordered logistic models were developed to investigate the factors influencing conflict risk and to explore differences in risk causation across distinct interaction patterns. The findings reveal that PVI can be categorized into three distinct patterns: near-interaction, far-soft-interaction and far-hard-interaction; Reductions in relative pedestrian-vehicle distances, increases in lower vehicle speed limits and upper pedestrian speed limits are common factors that reduce conflict severity; For near-interactions, rapid pedestrian deceleration, right-of-way competition, and high-speed vehicles reduce conflict risk, and both extreme vehicle deceleration and rapid pedestrian acceleration increase the danger; For far-interactions, an increase in the pedestrian speed lower limit leads to an increase in conflict risk; Rapid vehicle deceleration and right-of-way competition reduce the conflict risk for far-soft-interactions; Pedestrian sharp deceleration and high-speed vehicles are the main factors that reduce the conflict risk of far-hard-interactions, while pedestrian sharp acceleration raises the danger. The application of the combined approach reduces the influence of the heterogeneity of PVI behavior on the analysis results. The conclusions of the study provide a theoretical basis for enhancing safety in pedestrian crossing on urban roads.

End-to-end autonomous driving methodologies eliminate the need for manually defined rules and explicit module interfaces. Instead, these approaches directly map trajectory points or control signals from raw sensor data, thereby addressing the inherent shortcomings associated with traditional modular methods, such as information loss and cascading errors, and overcoming the performance limitations imposed by rule-driven frameworks. Recent advancements in self-supervised-learning-based generative artificial intelligence have exhibited substantial emergent intelligence capabilities, significantly promoting the evolution of end-to-end methodologies. However, the existing literature lacks a comprehensive synthesis of the advancements in generative end-to-end autonomous driving. Consequently, this paper systematically reviews the research progress, technical challenges, and developmental trends in end-to-end autonomous driving. Initially, the input and output modalities of the end-to-end models are delineated. Based on the historical progression of end-to-end autonomous driving, this paper provides an overview and comparative analysis of the foundational concepts, current research status, and technical challenges of traditional, modular, and generative end-to-end methods. Subsequently, the evaluation methodologies and training datasets utilized for end-to-end models are summarized. Furthermore, this paper explores the challenges currently faced by end-to-end autonomous driving technologies in relation to generalization, interpretability, causality, safety, and comfort. Finally, predictions are made for the future trends of end-to-end autonomous driving, emphasizing the fact that edge scenarios provide critical support for the training of end-to-end models, which can enhance the generalization capabilities. In addition, self-supervised learning can effectively improve training efficiency, personalized driving can optimize user experience, and world models represent a pivotal direction for the further advancement of end-to-end autonomous driving. The findings of this research serve as a significant reference for refining the theoretical framework and enhancing the performance of end-to-end autonomous driving systems.

Fault diagnosis of power battery systems is key to ensuring the safe and reliable operation of electric vehicles, in which the avoidance of false alarms not only reduces the driver's anxiety regarding vehicle safety but is also necessary for the practical application of the diagnostic method. Therefore, it is crucial to improve the reliability of the method. Abnormal voltage fluctuations in a power battery system are critical signals released by the deterioration of battery performance; hence, entropy methods, which can satisfactorily assess the degree of data dispersion, have been widely studied in battery fault diagnosis. However, when the classical Shannon entropy method based on interval probability was validated in engineering practice, many primary and secondary false-alarm single cells were found in the results. Vehicle voltage data with thermal runaway accidents were first used to analyze the fault diagnosis principle of the model to improve the accuracy of the method. Furthermore, based on normal vehicle operation data, the false-alarm mechanisms of the model in two typical scenarios were investigated. Under the above conditions, two measures were proposed to mitigate the false and missing alarm problems of the original method: the data optimization method and the kernel density estimation and entropy fusion method. Finally, real fault samples with different fault characteristics were selected to test the generalization ability of the algorithms, and their validity and reliability were verified separately. Based on a large amount of normal in-service vehicle data, a comparative analysis of the performance before and after model optimization was conducted. The results show that the relative false alarm rates of the two methods on normal vehicles decrease by 90% and 98%, respectively. Thus, this study significantly improves the reliability of the diagnostic strategy, promotes the online real-vehicle application of the methods, and provides ideas for analyzing and optimizing the accuracy of other fault diagnostic strategies.

To evaluate the fire resistance of hanger systems in suspension bridges, vehicle fires are classified into five levels. Levels 1 and 2 represent passenger vehicle fires, levels 3 and 4 correspond to truck fires, and level 5 represents tanker fires. These vehicle fires are characterized by distinct maximum heat release rates and burning durations. The proposed hierarchy was validated using existing vehicle fire experiments. Geometric features of flames are established for Levels 3, 4, and 5 vehicle fires based on previous vehicle fire incidents. For passenger vehicle fires, a cylindrical flame radiation model was employed to compute spatial radiative heat flux, validated through three full-scale car fire tests. In the case of truck fires, a prismatic flame radiation model was used to calculate spatial radiative heat flux. For tanker fires with crosswinds, a computational fluid dynamics method validated by a liquefied natural gas trench fire test was employed to calculate the heat flux envelope on the hanger surface. An incremental temperature calculation formula for hanger cross-sections with radiative heat flux boundary conditions was derived, and validation was performed using finite element models. Using mechanical property tests of high-strength steel wires at high temperatures, a quantitative relationship between critical temperature and design safety factor of hangers is developed based on the ultimate load-carrying capacity at high temperatures. Finally, integrating the above outcomes, a five-step theoretical framework is proposed to evaluate the fire resistance of hanger systems under graded vehicle fires. This algorithm can serve as a reference for the assessment and fire-resistance design of hanger systems in suspension bridges.

Oil-tanker explosion fire has enormous power and poses a severe threat to the safety performance of crossing sea bridge. In order to study the structural response of suspension bridges exposed to complex extreme fire environments caused by oil-tanker explosions, and to clarify the safety of suspension bridges under complex extreme fire loads, a large-span suspension bridge was selected as the research object. The prediction process of suspension bridges safety performance (fire resistance) during oil-tanker explosion fires was provided. Firstly, the computational fluid dynamics-finite element method (CFD-FEM) coupling method was used to reconstruct the oil-tanker explosion fire environment. A three-dimensional multi-scale numerical prediction model for local girder segment and the entire bridge structure were established. The heat transfer mode of bridge segment and performance evolution of the entire bridge structure during oil-tanker explosion were revealed in depth. Subsequently, the high-temperature response and failure mode of steel box girder (stiffening girder) under oil-tanker explosion were studied, and the effects of different fire positions, distance from the fire surface to bottom plate of steel box girder, and wind speed on the fire response behavior of suspension bridge were analyzed. A fire resistance limit warning method for suspension bridges exposed to oil-tanker explosion environment was proposed. The research results indicate that the deformation of local suspension bridge segment under oil-tanker explosion continues to increase. And the fire affected bridge segment shows a failure mode of overall downward deflection followed by upward bowing in middle area, resulting in a development trend of first increasing and then decreasing for suspension cable force in the middle area. The fire position has a significant impact on the overall structural performance of the suspension bridge. As fire position approaches the mid span area, the deflection of girder segment in the middle area increases by 62% compared to the girder segment adjacent to the tower. When the distance from fire surface to steel box girder is reduced from 50 m to 20 m, the peak deflection and total bowing amplitude (the difference between peak values of deflection and bowing) of local girder segment increase by more than 19%, and the structural failure time is advanced by 10 minutes. Wind speed would change the shape of deflagration flame, significantly affecting the distribution of heating surfaces and high temperature response characteristics on both sides of box girders. When wind speed is 8 m·s^－1, fire intensity of the windward side box girder is significantly reduced, and the total bowing amplitude of bottom plate is reduced by 17% compared to 2 m·s^－1. The critical temperature during the bending deformation of steel box girder bottom plate is between 510 ℃-550 ℃, and limit temperature during the buckling instability of steel box girder bottom plate is between 685 ℃-715 ℃. The critical temperature and limit temperature can be used as two-stage warning temperatures for safety performance, thereby achieving two real-time warnings before steel box girder failure. The research conclusion can provide theoretical bases for the safety performance monitoring and early warning of cable supported steel bridges in complex fire environments, and further guide the safe operation and maintenance of similar bridges.