This study examines antecedents and underlying mechanisms shaping enterprises' willingness to participate in government-led emergency logistics collaboration. A questionnaire of transportation enterprises in Wuhan was conducted to assess their participation willingness. Drawing on TPB and CSR Pyramid theory, a theoretical model was developed to represent the formation mechanisms of enterprise participation willingness and empirically tested using Logit regression analysis. Results show that the government's multi-party coordination and scheduling capabilities and safety assurance capabilities have significant positive effects on enterprise participation willingness, while the compensation mechanism has a marginally significant effect. Two improvement pathways are proposed: strengthening perceived facilitating factors to reinforce enterprises' trust and confidence, and optimizing enterprises' expectations of long-term positive returns to enhance their recognition of social impact.

To explore innovative pathways for the development of emergency management discipline and address its current bottlenecks, this study systematized the theoretical logic of emergency management discipline construction based on value co-creation theory and designed a four-dimensional framework encompassing goal optimization, community building, full-process value co-creation, and technological empowerment. The results indicate that China's emergency management discipline construction currently faces challenges including fragmented value cognition, insufficient stakeholder coordination, unbalanced resource allocation, and superficial technological empowerment. An integrated theoretical framework incorporating value coupling, community construction, full-process co-creation, and technological empowerment can be constructed to guide disciplinary practice. Applying value co-creation theory can address deep-seated contradictions in discipline construction, thereby promoting its transformation from fragmentation toward systematization.

In order to explore the formation paths of farmers' non-migration intention in high-risk areas of natural disasters and formulate more effective resettlement policies, five indicators across three dimensions—behavioral attitudes, subjective norms and perceived behavioral control—were extracted as antecedent conditions based on the theory of planned behavior. Using questionnaire survey data from 155 farmers in former Caopo township, Sichuan province, a fsQCA was employed to investigate the determinants for farmers' non-migration intention. The results show that household income, agricultural dependence, risk perception, policy satisfaction and subjective norms cannot individually constitute the necessary conditions for non-migration intention. There are three main paths leading to high non-migration intention. The first is "risk indifference", characterized by low risk perception and weak subjective norms. The second is "livelihood constraint", characterized by high agricultural dependence, low household income, and weak subjective norms. The third is "economically secure but unwilling", characterized by high household income, low policy satisfaction, and weak subjective norms.

In order to enhance the identification and control of low-probability, low-consequence risk events, this study draws on the theory of acceptable risk and employs questionnaire surveys combined with repeated-measures analysis of variance to examine the applicability and scientific validity of animal metaphors in representing acceptable risks. The questionnaire was designed to assess four animals across dimensions such as perceived threat, stability, controllability, and cultural-emotional associations. Sixty participants with a background in risk management were invited to provide quantitative evaluations. The results indicate that the white rabbit received significantly higher acceptability ratings than the cat, dog, and wolf, with the lowest risk level and the most positive cultural associations, indicating its suitability as a metaphorical representation of acceptable risk. Rabbit risks are characterized by low probability, mild consequences, and cumulative effects, and may transform into unacceptable risks under the influence of multiple factors. Based on this, three prevention principles are proposed: adhering to ALARP principle, incorporating differences in individual and group risk perception, and applying high-probability thinking to low-probability events. A multi-level prevention system comprising elimination, substitution, control, and isolation is constructed to achieve comprehensive risk governance for white rabbit risks.

In order to quantify the construction risk of urban subway engineering technology interface and screen the key risk factors, a construction risk assessment method of subway engineering technology interface based on fault tree and fuzzy BN was proposed. Firstly, the construction risk factors of metro engineering technology interface were identified from five aspects: personnel, materials and equipment, technology, environment, and management. Secondly, the fault tree model was used to sort out the logical relationship between the construction risk factors, construct a Bayesian network model of the construction risk of metro engineering technology interface, and evaluate the probability of the occurrence of each construction risk factor based on the theory of fuzzy sets and the experience of the experts. Lastly, BN was used to assess the construction risk of Beijing Metro Line 17, which was the largest and most important construction risk factor in China. Finally, BN was used to simulate and evaluate the construction risk of the northern section of Beijing Metro Line 17 as an example. The results show that the probability of risk occurrence in the construction phase of subway engineering technology interface is 65%, which is in the medium risk status level and matches the actual construction of the engineering technology interface. Through reverse diagnostic reasoning, the combination of risk factors with greater impact can be quickly identified, and the efficiency of the diagnosis of the construction risk events of the technical interface can be improved. Meanwhile, through the sensitivity analysis, it can be seen that the weak safety awareness of the technical interface construction personnel and the implementation of the interface construction management system are the key risk factors leading to the occurrence of engineering and technical interface construction risks. The assessment method can provide theoretical support for the prevention and management of metro engineering technology interface construction risk. By constructing a fault tree model to clearly sort out the logical relationships between risk factors, and then converting it into a fuzzy BN model, the accurate calculation and evaluation of risk probabilities are realized. This method has been verified in practical engineering cases, enabling the precise identification of key risk factors and providing a clear direction for risk management and control.

To enhance the accuracy and robustness of miner behavior recognition in underground belt conveyor areas and reduce the occurrence of unsafe behaviors, a multimodal feature fusion-based recognition method was proposed. The method integrated RGB and skeletal modalities, where an improved Efficient Attention(EA)-SlowFast networkwas employed to extract RGB features, and an improved Mobile-Enhanced(ME)-YOLO was utilized for miner detection. The detected miners were subsequently processed by Lite-HRNet for pose estimation to extract skeletal keypoints, which were further encoded using an improved Depthwise Convolution and Channel Prior Attention Spatio-Temporal Graph Convolutional Network (DC-STGCN) to obtain skeletal features. Finally, late fusion of RGB and skeletal features was performed for unsafe behavior recognition. In the experimental design, both public datasets (Human Motion Database 51(HMDB51), University of Central Florida 101 Action Recognition Dataset(UCF101)) and a self-constructed underground coal mine unsafe behavior dataset were adopted for validation. The results demonstrate that, compared with single-modal baseline models, EA-SlowFast and DC-STGCN achieve recognition accuracies of 71.6% and 68.3% on HMDB51, and 88.3% and 85.4% on UCF101, respectively. The multimodal fusion model outperforms the unimproved fusion baseline, achieving 75.4% and 93.5% on HMDB51 and UCF101, respectively. On the self-constructed dataset, ME--YOLO attains a Mean Average Precision(mAP)@0.5 of 91.8% with an inference speed of 38 f/s, satisfying real-time requirements. The fusion model further achieves an accuracy of 90.6%, confirming the effectiveness of the proposed approach in complex underground belt conveyor environments.

In order to enhance the proficiency of general aviation pilots in walk-around inspections and ensure aviation safety, eye-tracking technology was employed to measure and analyze the completion rates and fixation behavior characteristics of pilots categorized by experience level. Based on the walk-around checklists and inspection procedures, Diamond DA-40D aircraft was divided into six Areas of Interest (AOI), including the engine, wings, fuselage, landing gear, and empennage. During the participants' walk-around inspections, the completion rates of the walk-around checks and the characteristics of fixation behavior for these six AOI were recorded by means of an eye tracker. The results show that the completion rates of inspections in five AOI, including the empennage, engine, right wing, left wing, and left landing gear, were significantly higher in the veteran group than in the novice group. With the increase in flight experience, the completion rates of general aviation pilots' walk-around inspections gradually improve. Significantly higher total fixation durations and total fixation counts were observed in the experienced group compared to the novice group across five AOI: the engine, right landing gear, right wing, left wing, and left landing gear. A significantly longer average fixation duration on the engine AOI was demonstrated by the novice group relative to the experienced group. Recommendations are provided for training to place greater emphasis on the empennage section, right wing, left wing, and left landing gear areas. Additionally, enhancement of novices' inspection capabilities regarding the engine and left landing gear areas is suggested.

To accurately detect obstacles during the operation of unmanned mining trucks in open-pit mines, an improved obstacle detection model based on YOLOv8 was proposed. Firstly, STA mechanism was introduced to address the issues of redundant computation and insufficient global modeling in traditional attention mechanisms, thereby enhancing the feature representation capability for small and occluded targets. Secondly, a neck network structure was designed to further improve the localization accuracy of small targets by fusing high-level and low-level features. Then, a compound loss function was constructed to balance the regression accuracy and robustness of bounding boxes, so as to adapt to the diverse shapes of obstacles in open-pit mines. Finally, a self-built dataset was utilized to conduct comparative experiments for verifying the effectiveness of the improved model. The results show that the accuracy, recall rate, and mean Average Precision (mAP) at an Intersection over Union(IoU) threshold of 0.5 of the improved model are 86.2%, 78.9%, and 83.1% respectively, which are 10.3%, 19.4%, and 18.0% higher than those of YOLOv8. Compared with traditional detection algorithms, the improved model exhibits advantages in detection accuracy, recall rate and computational complexity.

To improve the efficiency of pre-disaster airport snow removal resource allocation, collaborative operations between equipment and artificial snow removal were considered. Snow removal resilience was introduced to represent snow removal efficiency, and a multi-objective optimization analysis model of airport pre-disaster snow removal resource reserve was established. The NSGA-Ⅱ was used to solve the optimal solution set and Pareto frontier. The C-OWA operator was introduced to calculate the objective weights of snow removal resilience and cost, and composite weight was determined by combining the subjective weight. Based on the TOPSIS method, the best scheme of pre-disaster snow removal resource reserve was obtained, and the influence of the subjective weight assigned to snow removal resilience on the decision results was analyzed. The results show that the maximum and minimum resilience of the optimal solution set of the multi-objective optimization model considering resilience and cost are 108 600 and 93 928 m³/h, respectively, and the costs are 1.118 million yuan and 798 600 yuan, respectively. The snow removal resilience of the best solution in the optimal solution set of pre-disaster snow removal resource reserve is 97 200 m³/h, and the snow removal cost is 920 700 yuan. Compared with the optimal solution set, the resilience of the optimal solution obtained by optimization analysis is increased by 31.5%, indicating that the established pre-disaster snow removal resource allocation method is feasible. Snow removal resilience is positively correlated with the subjective weight. For example, when the subjective weight is 0, 0.2 and 0.4, the corresponding snow removal resilience are 85 496, 89 600 and 97 200 m³/h, respectively.

In order to improve the efficiency of gas extraction, numerical simulation methods were employed to analyze the characteristics of changes in gas volume fraction, gas breakthrough time, coal permeability, CH₄ extraction rate, and extraction quantity under different gas injection pressures. The results show that after flue gas injection, the volume fraction of CH₄ shows a downward trend, while the volume fractions of CO₂, N₂ and O₂ show an upward trend. The increase in injection pressure will shorten the breakthrough time of each gas in the flue gas, reduce the permeability of the coal seam and increase the gas extraction volume. At 180 days of extraction, the gas injection pressure increased from 0.5 MPa to 2.5 MPa, the coal seam permeability decreased from 8.722 7×10^-17 m² to 8.711 5×10^-17 m², and the extraction volume increased from 449.01 m³ to 715.30 m³. The gas drainage rate shows the variation characteristics of rising - falling - rising - falling. From the perspective of the influence of gas injection pressure on drainage, the gas injection pressure has a relatively small impact on the drainage in the first two stages.