The number of elderly drivers in China continues to grow, and the changes in the driver structure pose challenges to traffic safety. Compared to drivers in other age groups, elderly drivers’ psychological function gradually declines and they are more prone to traffic accidents. Cognitive function is significantly correlated with driving safety performance. Based on the driving characteristics of elderly people, this article started from three cognitive functional areas of attention response ability, executive processing ability, and spatial perception ability, and designed driving simulation experiments to obtain cognitive driving behavior data. It analyzed the differences in driving behavior characteristics among young people, middle-aged people, and elderly people. By combining subjective and objective methods to determine indicator, weights, a method for calculating the cognitive driving behavior index was proposed. A generalized linear mixed model was established with driver attributes and cognitive function as independent variables and cognitive driving behavior index as dependent variable to explore the impact of different factors on cognitive driving ability. The results showed that age, weekly driving frequency, self-regulation, and TMT-B (Trail Making Test-B) were significantly correlated with cognitive driving behavior index, with MMSE (Mini-mental State Examination) showing marginal significant correlation. The cognitive driving behavior index of elderly drivers was greatly influenced by individual traits. Compared to the elderly, the cognitive driving behavior index of young people was worse, while that of middle-aged people was better. People with lower weekly driving frequency had better cognitive driving behavior index than those with higher weekly driving frequency. Drivers with low and medium self-regulation frequencies have better cognitive driving behavior indices than those with high self-regulation frequencies. TMT-B measurement showed that the cognitive driving behavior index of drivers with normal cognition was better than those with cognitive impairment. Starting from the perspective of human factors in traffic accidents, this study explored the cognitive challenges faced by elderly drivers, proposed a calculation method for the cognitive driving behavior index of elderly people, and analyzed the influencing factors, providing reference for simplifying the evaluation process of elderly driving suitability and formulating driving safety intervention strategies.

This paper aims to solve the problem of slow convergence rate and large error of existing intelligent algorithms in the process of optimizing support vector machine to identify risky driving behavior. Firstly, Tent mapping was used to replace the random setting of population initialization of ASO algorithm to increase the diversity and quality of atomic population. Secondly, the hybrid mechanism of dimension-by-dimension pinhole imaging reverse learning and Cauchy mutation was used to improve the diversity of preferred positions of atomic individuals and overcome the problem that ASO algorithm is easy to fall into local optimum and premature convergence. Finally, the adaptive variable spiral search strategy was introduced to improve the atomic individual position update process, so as to improve the global search ability of ASO algorithm, realize the effective balance between global search and local development, and alleviate the problem that ASO algorithm is easy to fall into local optimum and lack of convergence accuracy. Taking the vehicle trajectory data of the exit ramp of Shanghai North Cross Channel as the input, the study used the hybrid strategy to improve the ASO algorithm so as to optimize the LSSVM parameters. And it constructed the classification and identification model of the risk driving behavior of the expressway exit ramp based on IASO-LSSVM. Numerical simulation results show that the average value, standard deviation, best fitness and worst fitness of the numerical simulation results of the IASO algorithm in 12 benchmark test functions are closer to the best optimization value. Compared with ASO-LSSVM and LSSVM, the accuracy, precision, recall and F₁ value of risk driving behavior classification and identification results of IASO-LSSVM model increased by 11.5~24.5, 14.1~29.0,15.1~28.6, 14.7~31.2 percentage points respectively, and the error range was the smallest in different types of risky driving behavior identification results. The accuracy and convergence rate of IASO algorithm are better than those of ASO algorithm, and the IASO-LSSVM model can be used for accurate identification of different types of risk driving behavior, which can provide data support and theoretical basis for reasonable discrimination of vehicle driving trajectory state and formulation of early warning and prevention measures of risk driving behavior.

Battery electric buses are increasingly applied and promoted in public transportation systems due to their advantages, such as low emissions and low noise levels. However, their limited driving range and long charging time necessitate frequent charging during daily operations, thus leading to a new charging scheduling problem. A reasonable charging schedule is of great significance in reducing the construction cost of charging facilities and charging costs. However, current research on optimizing electric bus charging schedules typically assumes a linear relationship between charging time and state of charge (SOC), and often neglects the comprehensive optimization of charging schedules and charging station operations, resulting in poor scenario reproduction and resource inefficiencies. Therefore, this paper further studied the optimization of charging schedules based on a predefined set of bus trip schedules. A mixed-integer programming model was developed to minimize the total system cost by optimizing the occurrence periods, the start and end times of charging, and the schedules of the charging piles synchronously. The model also fully considers time-of-use electricity pricing policy, partial charging strategies, and the nonlinear characteristics of battery charging. To solve the problem, this paper first linearized the nonlinear charging function of the battery into a piecewise linear one and then used the commercial solver Gurobi to obtain the optimal solution. Additionally, a tailored algorithm was designed based on the minimum-cost-maximum-flow theory and the deficit function. Multiple sets of experiments were conducted to validate the effectiveness of the proposed algorithm based on five bus routes in Beijing. The results, obtained through both Gurobi and the proposed optimized algorithm, demonstrate that the proposed algorithm can achieve a significant reduction in total system costs, ranging from 28.34% to 56.1% across various scenarios. These findings confirm the efficiency of the algorithm and its potential to optimize charging schedules effectively.

The phenomenon and mechanism of sound generation and transmission of the car door sealing system with wind excitation are complex, and its research is much less.Taking the door primary sealing cavities of the B-pillar and C-pillar and the sealing cavity of the backdoor of a SUV as the research objects, this paper extracted their structures which are equivalent to regular cavity structures. It established a small acoustic wind tunnel with a testing platform and the testing method for the cavity sound generation and transmission, and conducted experiment on the sound characteristics and influencing factors, as well as the sound transmission characteristics of different compression of sealing strips and sealing gaps for the cavities with wind excitation. The results indicate that the mechanism of whistling from the door sealing cavity (backdoor cavity) is different. At low wind speeds, it is generated by coupling resonance between self-sustained oscillation and Helmholtz resonance cavity, while at high wind speeds, it is generated by coupling resonance between self-sustained oscillation and cavity mode, and self-sustained oscillation and cavity resonance with broadband excitation are the essential reasons for other peaks in the spectrum. There are significant sound characteristic differences between B-pillar and C-pillar the cavities and the backdoor cavity due to the sealing strip. The changes in yaw angle, pitch angle, and turbulence intensity of the flow have a significant impact on the amplitude and frequency of self-excited oscillation excitation, which is one of the sources of peak noise with a bandwidth below 1 000 Hz in the car. For small sealing strips compression (such as 0~4 mm) the sound transmission may increase with the increase of compression. The research on the sound phenomenon and mechanism of small cavity in car door gaps based on cavity noise theory is missing from previous studies. The sound transmission conclusion of different compression of the sealing strip has guiding significance for the design of car door seals.

The event wherein vehicles from the adjacent lane execute a lane-change maneuver, cutting into the lane occupied by autonomous vehicles, epitomizes a typical high-risk scenario on expressways within the domain of autonomous driving. Replicating such scenarios for testing on actual expressways involves significant safety risks. Virtual simulation test is one of the best approaches to addressing this issue. In order to automatically generate mass high-fidelity expressway lane-change cut-in test scenarios, this paper presents an automatic generation algorithm of lane-change cut-in test scenarios for autonomous vehicles based on TsGAN (Time-Series Generative Adversarial Network). In this algorithm, the time headway and lateral gap at the cut-in moment are taken as the evaluation metrics for scenario risk assessment, and 2 853 instances of lane-change cut-in scenarios with four different risk levels are extracted from the real expressway trajectory dataset highD. A model for generating lane-change cut-in test scenarios is established based on TsGAN, and is trained with the extracted real trajectory data. Then, the model is employed to generate the trajectories for the lane-change vehicle and the tested vehicle before the cut-in moment. To authenticate the generated trajectories, the distribution similarity and spectral error between the generated and the real trajectories are compared. Furthermore, an in-depth analysis of the kinematic interplay between the two vehicles at the cut-in moment and the distribution of trajectory parameters in the generated scenarios is performed to validate the coverage of the generated scenarios within naturalistic settings. The findings can be summarized as follows: (1) as illustrated by the distribution of key trajectory parameters, the average similarity between the generated and the real lane-change trajectories is 79.7%, with an average spectral error less than 8%, and more than 83.2% of the generated trajectories are within the buffer of the most analogous real trajectories, indicating a notable fidelity of the generated trajectories; (2) as compared with the collected real scenarios, the generated instances exhibit a more expansive coverage and a more even distribution within parameter intervals, and the time headway and lateral gap between the lane-change vehicle and the tested vehicle at the cut-in moment decrease by 17.83% and 16.37% in average, respectively, the distribution range of trajectory parameters expands by 19.44%, signifying a heightened coverage of the generated scenarios; and (3) the proposed TsGAN-based generation model has the capability of emulating lane-change cut-in test scenarios with four different risk levels, exhibiting pronounced specificity.

Considering the current maturity level of autonomous technology and the general public’s acceptance of this technology, the connected mixed platoon mode may be a practical transitional strategy prior to the widespread deployment of fully autonomous platoons. To study the influence of connected condition and platoon mode on the operational characteristics and eco-safe influence of connected mixed platoon, this paper constructed a connected mixed platoon mode consisting of five vehicles: a human-driven leader vehicle equipped with an L2 connected driving assistance system, followed by connected autonomous vehicles. Thirty-six participants were invited to participate in driving simulation experiments using the developed test platform. Given that time-domain analysis focuses on the variation of indicators in the time dimension, while frequency-domain analysis can explore the frequency distribution characteristics of indicators, this paper analyzed vehicle operation characteristics from the time and frequency domain dimensions with indicators of speed, acceleration, fuel consumption and speed safety entropy. The results show that both the connected condition and the platoon mode contribute to stable vehicle operation and significantly enhance their eco-safety. Compared with the traditional single-vehicle mode, the connected mixed platoon has the best comprehensive efficiency and can reduce the fuel consumption by 10.67% and the speed safety entropy by 73.25%. The research results can provide scheme reference and platform support for autonomous driving enterprises and industries in executing connected HMI design, navigate on autopilot (NOA) development, and the feasibility and effectiveness test of connected mixed platoon.

The S-curve pavement formation is susceptible to waterlogging due to its distinctive linear geometric characteristics. Therefore, the mastery of the waterlogging pattern is beneficial for the prevention and management of the hazards associated with waterlogged road sections, particularly in the design and construction phases. Using computational fluid dynamics software Fluent and the Eulerian Wall Film (EWF) which is based on the Eulerian-Lagrange method in computational fluid dynamics analysis, this paper carried out simulation calculations on the the typical S-curve bidirectional multilane road model with refinement treatment. The simulation included different variables, such as geometric features and rainfall, and obtained the film thickness and flow velocity distribution of the ponded water. The simulation results indicate that during the rainfall stage, 0% to 1% of the cross slope of the S-curve represents a waterlogged road section, 1% to 2% of the cross slope represents a waterlogged road section, and the waterlogging situation of the road section with a cross slope exceeding 2% is influenced by the intensity of rainfall. During the drainage stage, a cross-slope of 0%~1% on the S-curve represents a challenging road section, while a cross-slope of 1%~2% indicates a poorly drained road section. A cross-slope of more than 2% indicates a road section that is smoothly drained. The drainage of the road section at the end of the drainage time can remove the water film to less than 2 cm under different geometric conditions and rainfall conditions. The dimensions and gradient of the cross slope influence the lateral distribution of water as well as the longitudinal distribution. As the drainage stage progresses, the water film thickness distribution law changes to a longitudinal high in the middle and low on both sides with the overall distribution of S pattern, while the water film flow rate distribution law becomes longitudinal low in the middle and high on both sides. The width of the roadway exerts a significant influence on the total amount of water that falls onto the roadway from rainfall. This, in turn, affects the water film thickness. The ultra-high retardation rate primarily increases the number of road sections that are waterlogged or prone to water damage, thereby influencing the range of water distribution.

Aiming at the problem that the current object detection algorithm has many parameters and larger computation amount, resulting in slower response and difficulties of application in intelligent vehicle systems, this paper proposed an improved CenterNet object detection algorithm. That is, the lightweight MobileNetV3 network was applied to replace the original ResNet-50 network to reduce the amount of computation; the depth-separable PANet was applied to replace the feature enhancement network to obtain the features after the fusion of the multi-scale feature information, and the SimAM attention mechanism was introduced to strengthen the attention of the target features before the fusion of the features, then the SiLU activation function was used to replace the original object detection network in the ReLU activation function to enhance the network learning. Finally, the CPAN-ASFF module was proposed to fuse the depth-separable PANet output multi-scale feature maps to improve the object detection accuracy. The optimized KITTI dataset was applied to train and detect the improved CenterNet object detection algorithm for validation. And the results show that its mean average precision is 80.7%, which is 12 percentage points higher than that of the original CenterNet object detection algorithm; its detection speed is 65 f/s, and its number of parameters is 8.91 M, which is 72.73% less than the original algorithm. The improved algorithm performs better in the detection of occluded objects, overlapped objects and objects similar to the background. In the SODA10M dataset, the detection accuracy and speed of the algorithm are better than the current mainstream algorithms. The optimization of the algorithm and the experiments in the paper laid the technical support for the application of intelligent vehicles in practical engineering.

Large scale activities can cause a sudden increase and dissipation of traffic flow in the area around the venue, resulting in occasional and uncertain fluctuations of the road network operation in the surrounding area. The existing methods are insufficient to capture the evolution mechanism of traffic flow under the influence of multi-dimensional factors in special events at the prediction scale. In order to fully exploit the information of time series and influencing factor features of road section speed and effectively deal with the coupling mechanism between different influencing features in speed prediction, this paper proposed a speed prediction model (MC-LSTM) combining Interpretable Machine Learning and Long Short-Term Memory network. Firstly, the study combined the characteristics of large scale activities to construct the set of influencing factors. Then it used the XGBoost algorithm to evaluate the relative importance of the impact of activity scale, nature and other factors characteristics on the speed of road sections around the venue. It quantified the synergistic utility of multiple factors on the operation state of the road network around the venue, fused LSTM networks, considered the time-dependent relationship of traffic state, captureed the temporal correlation of different historical periods, and accurately predicted the speed of road sections around the venue during the activity. MC-LSTM was validated by taking the road network around large scale activities venues in Beijing for six consecutive months. The results indicate that the prediction accuracy of the MC-LSTM model can reach more than 94.5%, which is better than that of XGBoost model considering multiple factors synergism, LSTM model considering only single factor features and the LSTM model not considering external features. It proved that the model proposed in this paper has better validity and stability. This study can provide a decision basis for optimizing the traffic organization of the road network around the large scale activities venues and formulating traffic control and security measures.

It is an important basis for precise formulation of transportation emission reduction measures to accurately analyze the importance of factors influencing residents’ travel mode and the sensitivity of carbon emissions. According to the comprehensive analysis of the influencing factors such as family attributes, personal attributes, travel attributes and environmental attributes of the residents’ travel survey, the prediction model of residents’ travel mode was constructed based on LightGBM (Light Gradient Boosting Machine) and verified. Combined with the travel activity level, the carbon emission coefficient of various energy types, the standard coal coefficient and other parameters, the travel carbon emission prediction model based on the resident attribute data was constructed. Finally, taking Guangzhou as an example, the carbon emission intensity and total amount of residents’ travel mode were predicted, and importance of factors influencing travel mode and sensitivity was analyzed.The results indicate that the carbon emission prediction model constructed based on the attribute data of residents can more accurately predict the carbon emission of various modes of travel, better analyze the importance and sensitivity of the influencing factors of carbon emission, and comprehensively reveal the relationship between travel behavior, travel mode and travel carbon emission. Among them, the distance between the start and the end and the nearest bus station or the distance from the nearest subway station, the cost of self-driving and travel distance are important factors affecting the choice of residents’ travel mode. The competitiveness of subway travel increases significantly with the decrease of distance when the distance between the starting and the end point and the nearest subway station drops by 55%. In the area with high density of bus stops, the distance between the start and the nearest bus station is not sensitive to residents’ travel mode choice. It is the turning point of the residents’ travel mode and carbon emission when the carbon emission cost increases by 400%. After passing the turning point, the car travel mode is difficult to transfer. The carbon emissions fell the fastest, with a maximum reduction of 90.4% when the reduction in travel distance was within 90%.

Differentiated highway toll strategy can promote carbon emission reduction in the road field and boost the realization of the “double-carbon” goal. In this study, the carbon reduction benefits are considered as a part of the highway’s benefits. With the goal of simultaneously considering highway benefits and logistics enterprise costs, the study constructed a differentiated toll pricing model based on dual-layer programming. Firstly, the factors influencing truck route choices were analyzed; then Logit model was used to construct transfer traffic flow measurement method based on differentiated highway charging and carbon emission measurement method based on transferred traffic flow, and it proposed carbon trading pricing model under shadow price; finally, considering the partial maximization of carbon emission reduction benefits and the cost minimization of logistics enterprises, it proposed the dual-layer programming model of differentiated charging, and at the same time, the traffic flow data of highway A and parallel national and provincial roads in Shandong province were used for example analysis. The results indicate that, through the adjustment of differentiated highway toll strategy, the dual-layer programming model ensures optimal benefits for highway management while minimizing logistics enterprise costs. Furthermore, adjusting toll standards around the optimal toll rate has minimal impact on the benefits of Highway A, with benefit fluctuations remaining below 1.5%, demonstrating good revenue stability. The dual-layer programming model for differentiated toll pricing model proposed in this study can maximize the increase in highway carbon reduction benefits and minimize logistics enterprise costs.

In order to overcome the difficulty in estimating the indirect carbon emissions from new energy vehicles due to the energy consumption in a large-scale expressway network, this paper proposes a new carbon emission calculation method of new energy vehicles based on the toll data in expressway ETC network. Firstly, the traffic flow data of new energy vehicles on expressways were cleansed and processed. Then, according to the distribution of toll gantries, road sections were defined and segmented. On this basis, the energy consumption of different types of new energy vehicles was analyzed. Furthermore, a carbon emission quantification model of new energy vehicles in the operational stage was established. Finally, by taking Guangdong Province as an example, the carbon emissions of new energy vehicles on expressways were calculated, with its spatial distribution characteristics being also analyzed. The results show that (1) Category 1 pure electric passenger cars are the main source of carbon emissions from new energy vehicles on expressways, accounting for 74.20% of the total carbon emissions of new energy vehicles, followed by Category 1 pure electric trucks, whose carbon emissions account for 14.18% of the total, and Category 1 plug-in hybrid passenger cars contribute 11.62% of the total; (2) the cities in the Pearl River Delta urban agglomeration are the main contributors to high carbon emissions from new energy vehicles, and the less developed regions in eastern, western and northern Guangdong account for less than 13% of the total carbon emissions from new energy vehicles; (3) the road segments with high carbon emissions from new energy vehicles mainly locate in the transportation hubs and inner-city expressway rings of developed cities, and Guangzhou metropolitan area as well as Shenzhen metropolitan area as the center has a radiating effect on the expressway network of surrounding cities; and (4) in less developed areas, due to the low density of expressway network and relatively lagging infrastructure, the penetration rate of new energy vehicles is low, the corresponding carbon emissions from new energy vehicles are generally low.

In view of the influence of free surface in tank on floating state and stability, this paper proposed a new floating state algorithm of multi-convex composite structure. The floating body and the liquid tank were decomposed into multiple tetrahedral elements, and then the buoyancy and center of buoyancy of the floating body and the center of gravity of the liquid tank were determined by analyzing the relative position of each tetrahedron and the water surface/liquid surface. An improved double iteration method was also proposed to determine the floating state of multi-convex composite structures: the inner iteration method was used to simulate the heaving motion of the floating body to obtain the water/liquid surface equations outside the floating body and inside the tank; the rotation motion of floating body was simulated by outer iteration. The two iterative methods were carried out alternately until the distance between the gravity action line and the buoyancy action line of the floating body meets the accuracy requirement, and the gravity is equal to the buoyancy force. The study calculated the floating state of a semi-submersible vessel and a semi-submersible offshore platform under various typical working conditions using self-compiled program, and the influence of free surface was also considered. In addition, MAXSURF modeling was used for comparative analysis, and AUTOCAD was used to verify the accuracy. The results show that: (1) the algorithm can take into account the influence of free surface on floating state with clear principle and good convergence; (2) the algorithm is easier to obtain accurate solutions for both traditional single-convex structures (such as single hull vessel) and multi-convex structures (such as multi-hull vessel, floating offshore platform, etc.), and its computational efficiency is higher; (3) the algorithm can realize the “building-block” modeling mode, which can greatly reduce the workload in multi-condition modeling; (4) the algorithm does not need to rely on foreign third-party graphics platform (such as AUTOCAD, CATIA) , which lays a good foundation for the localization of related industrial software in our country.

An adaptive feedback tracking control scheme with state and input quantization was designed for the track tracking control problem of unmanned unmanned surface vehicle under the restricted communication bandwidth at sea. While ensuring effective tracking, it reduces the burden of maritime communication signal transmission, decreases the actuator execution frequency and reduces the control amplitude. Firstly, the system control law was designed based on the adaptive backstepping method, which combined with the dynamic surface technology to effectively reduce the computational inflation problem of the virtual control law. For the uncertain terms existing in the control system, a fuzzy logic system was used for approximation. Next, the state variables and input variables in the control system were quantized separately using a uniform quantizer, and the quantized state feedback information was used in the design of the unmanned surface vehicle track tracking controller. Based on the obtained quantization information, a control law for tracking the trajectory of an unmanned surface vehicle was proposed under the conditions of simultaneous consideration of state and input quantization. The boundedness of the errors between quantized and unquantized variables in the closed-loop control system was demonstrated by a recursive approach. The stability of the designed fuzzy adaptive feedback tracking control system with state quantization and input quantization was demonstrated based on Lyapunov stability theory when both state quantization and input quantization were considered. Finally, the effectiveness of the proposed scheme is verified by two sets of simulation experiments. That is, under the simultaneous consideration of state quantification and input quantification, the unmanned surface vehicle can still maintain a good tracking performance of the ideal trajectory, and effectively reduce the execution frequency of the actuator, which is more in line with the practice of navigation engineering.

The impact of land use on passenger flow within the influence scope of urban rail transit stations has different spatiotemporal differentiation characteristics. To explore the complex nonlinear relationship between land use and passenger flow at different stations, this paper proposed a differentiation identification method based on the spatial distribution of land use variables, and through time-phased multiscale geographic weighted regression, station clustering indicators that can characterize the spatiotemporal changing characteristics of land use impact on passenger flow were obtained. K-means++ algorithm was used to divide the stations into four categories, and the complex nonlinear relationship between land use and railway passenger flow under different categories was explored based on the improved gradient boosting decision tree model. Research shows that the accuracy of nonlinear model can be effectively improved by capturing the spatiotemporal heterogeneity of the relationship between the land use and passenger flow and classifying the stations properly. According to the output results, the key factors are different for each category. For the first category, the bus station number and sidewalk density have top relative importance value of 61.35% and 30.08% respectively; the key factors are the same for the fourth, but with an importance value decreasing from 61.35% to 30.31% for the bus station number. For the second category, the building densityhas the greatest impact with a relative ratio of 66.57%, and on the contrary, which only accounts for 5.59% for the third one. Meanwhile, there are significant and varying threshold effects on the relationship between land use and rail transit passenger flow. The result shows that different types of stations should put different emphasis on land use development, and land use design indicators should be controlled within a reasonable range. This research will provide theoretical support and quantitative guidance for the formulation of differentiated land use development strategies around stations.

Urban spatial heat map reflects population aggregation and street vitality. In order to explore the interactive relationship between urban rail transit and spatial heat map, this study used Baidu heat map and rail transit station ridership data to analyze the coupling characteristics between network centrality, ridership and nearby spatial heat index of rail transit stations on a micro level, taking Shanghai as a case study. Firstly, it investigated the overall coupling relationship between two categories of station attributes and spatial heat through Pearson bivariate correlation analysis. Then, bivariate spatial autocorrelation and geographically weighted regression analysis methods were introduced to explore the spatial association patterns between network centrality and spatial heat, as well as between spatial heat and ridership, followed by a spatial differentiation comparison between the two coupling types. The results show that the coupling relationship between rail network centrality and spatial heat is obviously better than that between ridership and spatial heat at station level, since traffic location advantage can usually develop higher spatial heat, while ridership may be affected by more complex factors. Spatial heat map is more suitable for quantifying the interaction between rail transit and urban space in areas outside the urban core, where increasing rail network centrality has a multiplier effect on spatial heat improvement, but improving spatial heat in areas with low-density development is more conducive to stimulating ridership. It is feasible to evaluate the ridership potential of new stations outside the urban core area by using spatial heat map, but this data alone is not enough to predict ridership. The urban renewal around rail transit stations can be optimized by referring to the differences between the two types of coupling at different spatial locations. This study explored the analytical framework for improving the layout of rail transit network based on urban spatial heat map, and optimizing TOD (Transit-Oriented Development) stations for factors negatively affecting their coupling. It provides a new perspective for measuring the man-land relationship of urban rail transit on the micro level.

By adding bilateral elevated layers, the renovation and expansion of the expressway have improved the traffic capacity, but a subsequent rise in problems, such as insufficient spacing, excessive traffic volume, and multi-diversion, has made the existing traffic signs fail to meet the drivers’ guidance needs. In response to this issue, this study took the Ji-He Expressway in Shenzhen as an example and designed four optimized traffic sign schemes of interchange based on the conventional standard scheme, and developed the corresponding stereoscopic compound expressway scenes by the Unity3D engine. The vehicles on the expressway will drive out of the ground-level road at the first diversion, enter the elevated-level road at the second diversion, and connect to other expressways at the third diversion. 28 drivers simulated driving in different scenes using external interactive devices, and their driving behavior data such as speed, acceleration, lane position, etc., were obtained. By calculating the point-by-point significance of indicators within each affected section in different guidance sign design schemes, differences in spatial dimensions of each indicator were presented, and further analysis shows that 3 km warning signs can help drivers perceive the situation of the ramp ahead as early as possible. Lane guidance signs can improve the longitudinal and lateral stability of vehicles at the ramp, and drivers can take better lane changing strategies. Navigation voices assistance can improve the driver’s speed control proficiency, lane changing awareness of expressway exits, and psychological comfort at the interchange of stereoscopic compound expressway. The research shows that Unity3D has high fidelity, good interactivity, and low cost, and it can provide technical support for relevant research on driving behavior in complex road scenes in the future. At the same time, this study also provides a theoretical reference for the establishment of traffic guidance signs in expressway renovation and expansion projects.

In order to effectively improve the efficiency of highway vehicle toll evasion inspection, based on ETC （Electronic Toll Collection） toll data, this paper proposed a highway vehicle evasion recognition model by combining KNN (K-Nearest Neighbor), adaptive boosting (Adaboost) algorithm and cost-sensitive learning mechanism. Firstly, in view of the large volume and redundancy of the original ETC toll flow data, data discretization and standardization processing rules were developed to repair and standardize the data form, and then two types of toll evasion features were extracted. Secondly, seven types of toll evasion, such as large vehicles with small tags, were selected as the main research objects by analyzing the ETC data set. Thirdly, to address the problem of inefficient model classification due to the “high-dimensional” characteristics of the evasion data, the best subset of features showing the evasion characteristics was selected by Pearson and Spearman correlation analysis and ReliefF importance analysis. Fourthly, to address the model overfitting problem caused by the class “imbalance” between toll evasion vehicles and normal vehicles, KNN was used as the base classifier in the Adaboost algorithm, and the boundary ambiguity of different categories was alleviated through TomekLinks undersampling, then a cost-sensitive learning mechanism was introduced to improve the model’s emphasis on the minority class (toll evasion vehicles) to alleviate the tendency to discriminate the majority class (normal vehicles). Finally, the performance of the KNN-Adaboost model incorporating cost-sensitive learning mechanisms was verified by comparing the recognition effects of different classification models for various types of evasion events. The results show that the precision of the proposed model is 0.98, Recall is 0.96, F1-score is 0.97, and Kappa coefficient is 0.95, indicating that the proposed model can better solve the sample class imbalance problem than other models and has higher recognition accuracy for minority class，and it can be a reference for improving the efficiency of highway toll inspection.

To accurately evaluate the decay process, mechanism and influencing factors of anti-sliding performance of asphalt ultra-thin wear layer after opening to traffic, this study first used arc mold and its supporting wheel grinding mechanism to prepare arc-shaped ultra-thin wear layer specimens. Then, it carried out the accelerated loading test of asphalt ultra-thin wear layer specimens with four gradation types and three asphalt binder grades by driving wheel accelerated loading system. Finally, it scanned the specimens after a certain number of wheel loads by laser texture scanner. Combined with the test data, the macro and micro structure depth decay models of asphalt ultra-thin wear layer based on S-type function were established respectively, and the independent and coupling effects of each porosity and asphalt binder on skid resistance were analyzed based on grey correlation analysis. The experimental results show that the driving wheel accelerated loading system combined with the S-type function model can quantitatively evaluate and predict the evolution process of the performance of the asphalt ultra-thin wear layer. The porosity affects the early decay rate of the macro and micro structural depth of the ultra-thin wear layer. The larger the porosity, the greater the early decay rate of the macro and micro structural depth of the ultra-thin wear layer. The middle and late decay rate of macro-structure depth is more affected by the grade of asphalt binder. The higher the performance grade of asphalt binder, the slower the decay rate of macrostructure depth in the middle and late stages. Void ratio and asphalt binder grade have similar influence on the decay rate of the micro-structure depth of the ultra-thin wear layer in the middle and late stage. Therefore, when designing the anti-sliding performance of the ultra-thin wear layer, it is beneficial to select the open-graded ultra-thin wear layer mixture with large porosity for the anti-sliding durability of the pavement, and the performance grade of asphalt binder can be reasonably selected according to the traffic volume.

To improve the traffic safety of highway bridge sections in foggy environments, this paper considered the influence of variable speed limit signs on driving behavior. To obtain the best effect, it put forward a quantitative evaluation method from the perspective of the obedience effect. Taking E’dong Yangtze River Bridge as the prototype, this paper selected the lowest visibility （100 m） and free flow service level of the bridge in recent years as the test environment. Three speed limit strategies were designed, namely, the control group S_Ⅰ (no speed limit strategy）, the experimental group S_Ⅱ (90~70 km/h speed limit strategy), and the experimental group S_Ⅲ (90~70~50 km/h step-by-step speed limit strategy）. Relying on the driving simulator, the micro-driving behavior data of foggy bridge scenes with different speed limit conditions were realized. The action mechanism of variable speed limit signs and driver characteristics were analyzed from the rapidity, stability, and accuracy of driver response by repeated measure analysis of variance, and the effectiveness of different speed limit strategies was evaluated using the fuzzy comprehensive evaluation method. The results show that the variable speed limit sign can make the driver take deceleration measures earlier, and the vehicle’s stability in the fog area is better. When the visibility is 100 m in fog, the steady-state frequency of 90~70~50 km/h step-by-step speed limiting strategy is more significant, the spatial stability is better, the speed overshoot and the following ratio are more minor, and the response accuracy is higher. The results of the fuzzy comprehensive evaluation show that the 90~70~50 km/h step-by-step speed limit strategy, as the optimal scheme, can effectively improve the adaptability of driving behavior in fog areas, reduce driving risk, and improve the stability of vehicle operation. The research results provide a solution for setting variable speed limit signs for bridges on foggy days and can provide adequate support for the active safety prevention and control of bridges on foggy days.

The existing research on identifying black spots in traffic accidents is mostly based on accident frequency or accident rate, without considering the impact characteristics of traffic accidents on different locations. In order to comprehensively consider the differential effects of traffic accidents in different traffic environments and road network characteristics and to solve the zero inflation problem of zero values far exceeding the classical discrete distribution in traffic accident data, this paper proposed an improved network kernel density estimation method that considers the comprehensive importance of nodes, and identified urban traffic accident black spots based on the zero inflation negative binomial regression model. Firstly, in the topological road network, a comprehensive impact index of accidents was constructed by comprehensively considering the traffic environment and road conditions at the location of the accident, and the accident severity index was embedded into the traditional network kernel density estimation. By generating a smooth density surface on the road network, the spatial aggregation of point events was qualitatively reflected. On this basis, a discrimination model based on zero-inflated negative binomial regression was constructed to clarify the boundary range of accident-prone areas and quantitatively depict the spatial distribution characteristics of accident black spots at different severity levels. Finally, an example analysis was carried out for Huaqiangbei street in Shenzhen. The results show that the search efficiency indexes of the proposed method are all larger than those of the planar kernel density estimation method at the threshold levels of 70%, 80% and 90%. Furthermore, some non-road areas are no longer mistaken, and the accuracy of the model is 3.60%, 5.31% and 7.20% larger than those of the traditional network kernel density method respectively after considering the comprehensive importance of nodes.