As a crucial component in the evolution of modern warfare toward digitization and intelligentization, unmanned aerial vehicle (UAV) equipment requires a more precise and efficient operation and maintenance (O&M) system. Based on the Department of Defense Architecture Framework (DoDAF) 2.0, the integration of Multi-Agent Systems (MAS) and military simulation technology provides a comprehensive, rational, and feasible theoretical foundation for the construction and validation of an intelligent O&M system for UAV equipment. Firstly, starting from the O&M tasks of UAV equipment in intelligent warfare, this study analyzes the capability requirements for intelligent UAV O&M by following the generation path of scenarios, activities, and capabilities. Three core capabilities are proposed: situational awareness, decision support, and mission execution. Secondly, various O&M tasks are decomposed into behaviors of multiple types of agents, and based on this, an intelligent O&M system for UAV equipment is designed using a “cloud-edge-terminal” distributed architecture. Finally, simulations are conducted to model and validate UAV equipment maintenance tasks. Experimental results demonstrate that the MAS-based UAV O&M system significantly enhances support efficiency, accuracy, and response speed, offering a novel solution for O&M in future UAV operations.

The rapid growth of the aviation industry highlights the need for strong safety management. Analyzing data on unsafe aviation events is crucial for preventing risks. This paper presents a new method that integrates the Transformer network model, clustering analysis, and feature network modeling to analyze Chinese text data on unsafe aviation events. Initially, the Transformer model is used to generate summaries of event texts, and the performance of three pre-trained Chinese models is evaluated and compared. Next, the Jieba tool is applied to segment both summarized and original texts to extract key features of unsafe events and prove the effectiveness of the pre-trained Transformer model in simplifying lengthy and redundant original texts. Then, cluster analysis based on text similarity categorizes the extracted features. By solving the correlation matrix of these features, this paper constructs a feature network for unsafe aviation events. The network’s global and individual metrics are calculated and then used to identify key feature nodes, which alert aviation professionals to focus more on the decision-making process for safety management. Based on the established network and these metrics, a data-driven hidden danger warning strategy is proposed and illustrated. Overall, the proposed method can effectively analyze Chinese texts of unsafe aviation events and provide a basis for improving aviation safety management.

To enhance path planning efficiency in unmanned aerial vehicle (UAV) search missions in complex environments, this paper proposes a coverage path planning (CPP) algorithm for a UAV that integrates the deep Q-network (DQN) with the A* algorithm (DQN-A*). In the proposed DQN-A* algorithm, a dual-driven reward mechanism is established, comprising a probability-weighted reward and a step-dependent reward, steering the UAV toward high-probability regions. Furthermore, to handle previously unknown obstacles in real time, the algorithm employs a multi-stage obstacle-identification strategy, enabling the UAV to improve coverage of traversable cells by dynamically adjusting its local path when newly detected obstacles are encountered. A theoretical analysis derives a principled recommended range for the UAV positional identifier based on statistical feature analysis; this range is then validated through extensive simulations. Additionally, Hamiltonian path pre-training is introduced to accelerate convergence. Comparative simulations demonstrate that the proposed DQN-A* algorithm achieves higher area-coverage and target-detection probabilities than benchmark algorithms in environments with unknown obstacles, offering valuable insights for positional encoding in deep reinforcement learning (DRL)-based robotic coverage problems.

Due to the information non-independence of attributes, combined with a complex and changeable environment, the analysis of risks faces great difficulties. In view of this problem, this paper proposes a new three-way decision-making (3WD) method, combined with prospect theory and a non-additive measure, to cope with multi-source and incomplete risk information systems. Prospect theory improves the loss function of the original 3WD model, and the combination of non-additive measurement and probability measurement provides a new perspective to understand the meaning of decision-making, which could measure the relative degree by considering expert knowledge and objective data. The theoretical basis and framework of this model are illustrated, and this model is applied to a real in-service aviation equipment structures risk evaluation problem involving multiple incomplete risk information sources. When the simulation analysis is carried out, the results show that the availability of this method is verified. This method can also evaluate and rank key risk factors in equipment structures, which provides a reliable basis for decisions in aviation safety management.

Composite materials are critical components in advanced equipment such as aerospace, however, delamination defects that readily arise during manufacturing and in service present serious risks to equipment safety. Terahertz non-destructive testing is highly effective for analyzing the internal structure of composite materials, making it an effective approach for precise identification of delamination defects. Current terahertz detection approaches mainly depend on single domain features, making it difficult to capture complementary information from both the time and frequency domains. To address this, a Time-Frequency Feature-fusion Network (TFFN) is proposed. In this network, a three-branch architecture is employed: local transient patterns and pulse-related structural features are extracted by the local time-frequency branch; damage-sensitive frequency bands are focused on by the frequency-domain branch through a channel-space-frequency band attention mechanism; and deep integration of time-frequency features is achieved by the time-frequency fusion branch using Manifold Mixup. Finally, the features extracted from the three branches are adaptively fused via a cross-branch attention mechanism, and defect identification is accomplished by the classifier. Experimental results show that this method achieves accuracies of 98.40% on the glass fiber reinforced polymer (GFRP) dataset and 98.63% on the quartz fiber reinforced polymer (QFRP) dataset, surpassing the best existing method by 2% and 1.25%, respectively. A substantial improvement in both defect identification accuracy and the model’s generalization ability for layered structures is thereby achieved.