Due to the perennial erosion of the substructure foundation in complex aquatic environments and the potential impact of drifting objects, wading bridges are prone to developing apparent defects. Therefore, it is essential to conduct regular inspections to assess their safety status. Owing to the limitations of underwater scenarios, conventional damage detection methods are difficult to achieve effective detection of damage. To address this issue, 3D sonar measurement technology was employed for the inspection of underwater foundations of bridges. The basic principles of 3D sonar measurement technology and its basic implementation process for underwater foundations of bridges were introduced, and an integrated processing method covering point-cloud acquisition, data processing and foundation morphology extraction was proposed, so as to generate a three-dimensional point-cloud model that could accurately reflect the apparent characteristics of the foundation. The results indicate that, under complex underwater environmental conditions, the proposed method can effectively achieve three-dimensional shape reconstruction and dimensional measurement for different types of foundations. For large-scale underwater foundations such as caissons and cofferdams, the maximum measurement error is only 7.9%, demonstrating good detection accuracy and engineering applicability. For small-scale components such as pile groups, although the average diameter measurement error reaches 18.2% due to weaker echo signals and sparse point clouds, the proposed method can still satisfy the basic requirements for geometric feature identification and safety assessment of the foundations.

To address the issues of low detection accuracy of target and poor real-time performance caused by severe environmental interference in rail track foreign object detection under conditions such as rain, fog and low light, an improved YOLOv12 railway track intrusion foreign object detection algorithm (SF-YOLO) was proposed. Firstly, to address low detection accuracy under environmental interference, an enhanced multi-dimensional collaborative attention mechanism (MCAM) was employed, which enhanced feature expression through three-dimensional collaborative modelling. Secondly, to improve detection accuracy while maintaining real-time performance, a new lightweight feature fusion module, SFNet-Rail, was established, which preserved the details of foreign object boundaries while strengthening multi-scale feature consistency. Thirdly, a FastPartial-Detect detection head network based on partial convolutions (PConv) was designed to reduce memory access redundancy and further enhance the detection performance of the model. Finally, a unified IoU loss function (UIoU) was introduced to dynamically optimize the weight distribution of prediction boxes. Experimental results demonstrate that on the self-built rail track foreign object dataset, the proposed algorithm achieves a mean accuracy (PmA@0.5) of 85.8%, with a 2.7% improvement over YOLOv12. The detection speed of the proposed algorithm is increased by 12%, and the computational load is decreased by 21.25%. The accuracy and real-time performance of multi-scale railway foreign object detection in environments such as rain, fog, and low light conditions can be effectively enhanced.

Based on Abaqus software, a simulation model for hydroplaning on airport asphalt pavements was developed by the coupled Eulerian-Lagrangian (CEL) method. The influence rules of factors such as tire load, tire pressure, pavement texture and water film thickness on hydroplaning behavior were systematically analyzed through controlled variable approach, and a prediction model for critical hydroplaning speed was constructed via multiple linear regression. The research results indicate that the established CEL model can effectively simulate the tire-pavement-water interaction. When an aircraft lands on a wet runway, if the water film thickness exceeds 8mm and the slip ratio is below 10%, braking is required to increase the slip ratio to mitigate the risk of hydroplaning. Under extremely wet conditions, the pavement gradation optimization should comprehensively consider factors such as texture depth and pavement connectivity. The established model clarifies the influence of key parameters, including water film thickness, tire load and pavement texture, providing a quantitative basis for the anti-skid design and safe operation of asphalt pavements.

The lifting of railway roadbeds via grouting is often uneven, and the effectiveness of the grouting is difficult to guarantee. To address this, the experimental study on soil pressure characteristic unit cell attached to capsular tube grouting was carried out, and the comparative analysis on the changes in additional soil pressure in the stratum caused by different grouting methods was conducted. The experimental results show that compared to direct grouting, capsular tube grouting demonstrates better linearity and controllability in the increase of additional soil pressure. In contrast to bag grouting, the additional soil pressure increase of capsular tube grouting exhibits a directionality where the radial is greater than the longitudinal. The maximum radial expansion of the capsular tube is limited by its diameter. When the grouting volume approaches full capacity, uniform compression of the surrounding soil can be achieved without longitudinal expansion, and basically no compression occurs during the grouting process. It should be noted that after the grouting material solidifies, free water may seep out. It is recommended to appropriately reduce the water content or add water-reducing agents to ensure long-term effectiveness. Based on the good directionality and controllability of capsular tube grouting, it can be used in subgrade lifting projects to replace traditional direct grouting methods, effectively avoiding grout leakage and slope deformation and cracking.

To deepen understanding of the mechanical properties of silica sand particles under impact loading, and enhance the theoretical foundation for the application of silica sand in energy dissipation in protective engineering, an improved split Hopkinson pressure bar (SHPB) apparatus was used to investigate the propagation law of stress waves, particle breakage behavior, and energy absorption effect of silica sand under medium and high strain rates. Results indicate that the velocity of stress waves in silica sand is directly proportional to the particle size, while the peak stress propagation speed exhibits a linear negative correlation with particle size. Additionally, the attenuation capacity of silica sand to stress waves increases with the increase of particle size, showing a linear relationship. Large-sized particles are more prone to fracture under impact, which can effectively increase the degree of energy dissipation. The particles after crushing exhibit a clear fractal distribution pattern, and the fractal dimension shows a good linear correlation with the crushing rate. According to the energy absorption efficiency characteristics of silica sand, under the same stress level, larger particles show higher energy absorption efficiency; under the same strain level, smaller particles demonstrate stronger energy absorption efficiency due to their lower compressibility. In practical engineering, larger particles are recommended to improve energy absorption efficiency, while smaller sizes are preferable when controlling deformation or enhancing stability is required.

To address the insufficient research on construction diversion risks of collaborative construction in multi-region serial projects, a Copula-Monte-Carlo-Mike multivariate coupling construction diversion risk analysis model was developed by integrating the Copula-Monte-Carlo flood analysis method and the Mike hydrodynamic simulation method. Taking the reach from the Qishi Hub to the Qingnian Hub in the Pinglu Canal as a case study, the characteristics of flood encounters between tributaries and the mainstream along the route were analyzed. The risk rates for different areas under various roughness conditions during the flood season were calculated, the variation patterns of the construction diversion risk rate under multi-region collaborative construction were discussed, and the formation mechanism of construction diversion risk was revealed. The research results indicate that the proposed model can simulate flood encounter scenarios involving multiple tributaries, effectively quantifying the diversion risks of the canal waterway under long-reach, multi-region collaborative construction conditions. The diversion risk exhibits significant spatiotemporal differentiation characteristics. Temporally, the risk is highly concentrated in the main flood season from June to August. Spatially, the inflow of tributaries increases the risk downstream. The risk rate shows a nonlinear growth trend of “acceleration-slowdown” with flow rate, reflecting the transition of river channel discharge capacity from normal to saturation. An increase in roughness significantly amplifies the risk at high flow rates. However, variations in bottom elevation prevent the spatial distribution of risk from evolving synchronously with the increasing flow discharge. The formation of diversion risk in the multi-region collaborative construction of the canal is a combined result of flow discharge, roughness, and bottom elevation. The non-linear driving effect of flow dominates the formation of risk. Roughness increases the equivalent flow risk level through flow accumulation effects, while bottom elevation variations regulate local risks by reshaping the flow regime.

The Fe3O4@MnFe-LDH magnetic catalyst was prepared by a two-step method to enhance the efficiency of peroxymonosulfate (PMS) oxidative degradation of Congo red. Its morphological characteristics and structural composition were analyzed, and the effects of initial pH value, catalyst dosage, PMS dosage, initial concentration of Congo red, and reaction temperature on the degradation performance of Congo red were explored. The stability and reusability of Fe3O4@MnFe-LDH were evaluated through cycling experiments and material characterization before and after reaction. In addition, radical quenching experiments, liquid chromatography-mass spectrometry (LC-MS), and the toxicity estimation software tool (T.E.S.T.) were employed to elucidate the reaction mechanism, degradation pathways, and changes in pollutant toxicity. The research results show that Fe3O4@MnFe-LDH possesses a high specific surface area and mesoporous structure, and its surface contains redox-active sites of Mn (Ⅱ)/Mn (Ⅲ) and Fe (Ⅱ)/Fe (Ⅲ), as well as functional groups such as hydroxyl and carbonate groups. Radical quenching experiment confirms that the synergistic effect of the Fe-Mn bimetallic system, involving the valence cycling of Mn(Ⅱ)/Mn(Ⅲ) and Fe(Ⅱ)/Fe(Ⅲ), activates PMS through electron transfer and promotes the generation of sulfate radicals (SO4·-), superoxide radicals (O2·-), singlet oxygen (1O2), and hydroxyl radicals (·OH). Congo red is oxidized by free radicals and gradually degraded through processes such as azo bond breakage and sulfonic acid group removal, but it is not completely mineralized. Instead, it generates seven different intermediate products, most of which have lower developmental toxicity than the original pollutant.

As a critical link in urban logistics, the “last mile” delivery faces challenges of low efficiency due to geographical constraints and traffic regulations that limited traditional truck-based distribution. The truck-drone collaborative delivery mode has the potential to overcome obstacles and improve timeliness; however, it imposes high requirements on drone takeoff and landing site configuration and truck-drone coordination. Considering potential drone takeoff and landing sites and comprehensive operational capabilities, a linear integer programming model that minimized total delivery time was constructed and an adaptive large neighborhood search (ALNS) algorithm incorporating a dynamic threshold acceptance criterion was proposed. Numerical experiments were conducted based on Solomon benchmark instances and real JD.com delivery cases to verify the validity of the proposed model and algorithm. The results show that: in small-scale benchmark instances, the improved ALNS can obtain a solution that is close to that of Gurobi (with an average of 415.00 seconds) within an average of 6.70 seconds, with an average Gap1 not exceeding 0.30%; in medium-scale and large-scale instances, Gurobi fails to find solutions within the prescribed time, while the improved ALNS achieves overall better delivery time than the traditional ALNS does, with a Gap2 of 9.40%. In two 100-node instances, the actual gap between the improved ALNS solution and the theoretical optimal solution does not exceed 15.20%. Compared with the truck-only mode, the collaborative delivery mode reduces delivery time by up to 19.20%, with an average saving of 13.80%. In practical examples, the collaborative delivery model can save 23.90% of the overall delivery time compared to the truck-only model. The convergence results of the algorithm indicate that the improved ALNS can accelerate convergence by dynamically adjusting the acceptance threshold. Sensitivity analysis indicates that when the drone endurance exceeds approximately 3,960 s and the payload capacity exceeds approximately 12,000 g, the improvement in delivery time diminishes significantly. In medium-scale instances, collaborative delivery efficiency is optimized when the number of potential takeoff and landing sites is approximately 7 and the spatial spacing is approximately 4,600 m. The improved ALNS based on site-feature operators reduces average delivery time by 0.40%, 1.30%, and 3.00% in small-scale, medium-scale, and large-scale instances, respectively.

An evaluation method for the optimization of microcirculation traffic organization based on the macroscopic fundamental diagram (MFD) was proposed, so as to analyze the impact of microcirculation traffic organization on regional traffic states from a macroscopic perspective. Firstly, a micro-circulation traffic organization scheme was constructed based on stochastic network loading (SNL) and the actual road network. Secondly, the MFD model was established using route length, traffic volume, road travel probability, and the impact of non-motor vehicles on the micro-circulation traffic system as variables. Based on this, a micro-circulation traffic organization evaluation model was obtained. Finally, the actual road network of Kunming was selected as the test object for traffic simulation analysis. By dividing the area into large and small sections, MFD evaluation was conducted on three kinds of micro-circulation traffic organization schemes with setting the shortest path, the maximum flow path and the maximum travel probability path as one-way streets. When the shortest path was designated as a one-way path, the maximum service flow in small road network areas and large road network areas increased by 22.6% and 15.8%, respectively. When the path with the largest traffic volume was designated as a one-way road, the maximum service flow in small road network areas and large road network areas increased by 11.1% and 10.6%, respectively. When the path with the highest travel probability was designated as a one-way path, the maximum service flow in small road network areas and large road network areas increased by 31.5% and 27.6%, respectively. The proposed model can effectively evaluate micro-circulation traffic organization schemes of different road network scales and provide a scientific evaluation basis for the formulation of micro-circulation road network control schemes.

Traffic flow prediction is one of the core tasks of intelligent transportation systems, with the primary challenge lying in the high coupling of non-Euclidean spatial dependencies and dynamic temporal characteristics within complex road networks. To address the limitations of existing methods in modeling spatial-temporal dependencies and fusing multi-source information, a novel model, STFformer, was proposed. The proposed model designed a global spatiotemporal interaction module, enabling joint modeling of spatial dependencies and temporal dynamic evolution of global nodes. Meanwhile, a weight-sharing dynamic graph convolution module was incorporated to refine node features through a time-varying adjacency matrix, capturing latent non-Euclidean structures and enhancing the model’s perception ability towards dynamic changes in road network topology. Furthermore, a gated enhancement fusion module was designed to realize adaptive integration of multi-source spatial-temporal information and stable representation, thereby improving the model’s generalization and robustness. The validity verification was conducted on a real traffic flow benchmark dataset. The research results show that the proposed model outperforms baseline models across three prediction metrics, confirming its predictive accuracy and stability in complex spatial-temporal scenarios.

In complex and dynamic traffic environments, accurately predicting pedestrian trajectories is crucial for fields such as autonomous driving, service robots, and smart cities. However, the high uncertainty of trajectory distribution, especially the spatial distribution differences between short-term and long-term predictions, poses challenges for modeling. To address this, a three-stage pedestrian trajectory prediction framework incorporating the idea of ITransformer, ITER (inverse trajectory transformer), was proposed. Through a staged strategy, the motion patterns at different time scales were captured in a hierarchical manner. Firstly, through short-term prediction, local dynamic patterns were learned. Secondly, the potential distribution of endpoints was modeled to grasp the global trend of long-term trajectories. Finally, fusing the information from the first two stages, the fine modeling of the complete future trajectory was achieved. Experimental results show that, in five scenes of the ETH/UCY dataset, ITER reduces the average displacement error by 13.04% and the final displacement error by 20.51% compared to the Agentformer model. In the SDD dataset, the proposed method reduces the average displacement error by 8.28% and the final displacement error by 8.02% compared to the Y-Net model. The research indicates that the proposed model achieves significant performance improvement, which can generate accurate and stable future trajectories, achieving the precise prediction for pedestrian trajectories.

Addressing the issue of excessive crowding and exceeding capacity limits of scenic spots during the holiday travel peak season, sudden large tourist flow scenarios in specific areas of the scenic spots were focused and the impact mechanism of tourist flow warning information dissemination on downstream tourist evacuation psychology was explored, aiming to optimize tourist flow management in scenic spots and prevent crowded and stampede accidents. Firstly, based on the protective action decision model (PADM), the evacuation decision-making mechanism was analyzed. Secondly, the questionnaire survey was designed and conducted to obtain perceptual data of tourists. Finally, a structural equation model (SEM) was constructed to reveal the impact of tourist flow early warning information (risk level, presentation format), tourists’ trust in the information, evacuation knowledge and experience on tourists’ risk perception and panic emotion. The moderating effect of information presentation formats on the psychological impact pathway of tourist evacuation was compared by multi-group analysis. Results indicate that risk level, trust, and evacuation experience are key antecedents of risk perception, while risk level, risk perception, trust, and evacuation knowledge collectively influence panic emotion. Crucially, the information presentation format exhibits a significant moderating effect. “Emphasizing action instructions” can effectively convey risk perception while inducing lower panic levels. Therefore, scenic area management should adopt this as the core strategy for issuing early warning information, in order to enhance emergency management effectiveness and safety.

In response to the dynamic and uncertain characteristics of traffic risk propagation of mountain highways, a traffic risk propagation modeling method based on the integration of graph theory and hidden Markov model (HMM) was proposed, and a theoretical analysis framework for the dynamic evolution of traffic risks on mountainous highway was conducted. Specifically, by firstly calculating the change rate of traffic risk index sequence data and then using a 60-minute time window, the sequence was segmented, and the correlation of the indicators was analyzed. Subsequently, based on minimum spanning tree (MST) graph theory, the indicator correlation matrix was mapped into an undirected graph structure, where risk indicators served as graph vertices and risk propagation paths served as graph edges. By introducing correlation distance measurement methods to quantify edge weights, the core indicator identification method and the time-varying evolution patterns of the traffic risk network were systematically revealed. Finally, combining the characteristic of no aftereffects of Markov chains, the risk states were categorized into three levels, including high, medium and low. A Markov state transition probability model was constructed to achieve quantitative estimation and dynamic prediction of risk state transition. Empirical results demonstrate that the proposed method can effectively analyze the changing patterns of traffic risk states, with a mean absolute percentage error (MAPE) of 5% for the prediction of observed variables, verifying the effectiveness of the proposed model. The research results provide theoretical basis and decision-making support for the dynamic management and control of traffic risks on mountainous highway.

Concrete cracks widely exist on the surfaces of bridges, roads, and buildings, affecting engineering safety. Therefore, crack damage detection of concrete structures is an important indicator for evaluating infrastructure safety. To address the problems such as low accuracy in traditional manual inspection, large parameter volumes and blurry crack details in existing crack detection models, a high-precision crack detection method for wall-climbing robots based on the YOLO-DSD network was proposed. The proposed method was based on the instance segmentation framework of the YOLO11-seg model, integrating the C3K-DWR module, the attention M-SEAM module and the Dyhead segmentation head. The YOLO-DSD algorithm reduced computational load while maintaining high-precision crack segmentation, and its accuracy surpassed that of mainstream segmentation models in the YOLO series. The research results show that compared with the baseline model, the precision of the proposed method (Precision) reaches 80.5%, with an improvement of 7.2%, and mAP50 reaches 71.6%, with an increase of 5.4%. Finally, the YOLO-DSD algorithm is deployed on a wall-climbing robot for inspection, enabling accurate extraction of crack edge information in complex backgrounds, which provides an efficient, robust and practical crack detection solution for real engineering applications.

To address the lane change planning problem for autonomous vehicles in scenarios with obstacles during changing lanes, an improved whale optimization algorithm (IWOA) was proposed and applied to lane change path planning. The proposed IWOA incorporated a three-layer adaptive mechanism: in the initialization layer, Tent mapping was used to generate a population with high fractal dimension, enhancing early exploration diversity; in the iterative control layer, the dynamic mapping function for the nonlinear convergence factor was constructed to extend the global search cycle; in the mutation enhancement layer, the adaptive Cauchy perturbation strategy based on information entropy was introduced, significantly improving the algorithm’s ability to escape local optima. During the path generation phase, cubic spline curves were used to model control points, and the comprehensive cost function was constructed, which included path smoothness, obstacle safety distance, lane boundary constraints. Simulation and real-vehicle test results show that compared to the traditional whale algorithm, average convergence speed and path smoothness of the proposed IWOA is improved by 18.5% and 18.7%, respectively. The research results indicate the proposed algorithm demonstrates superior convergence accuracy, escape capability and path feasibility in multiple typical lane-changing path planning scenarios.