Linfeng Zhong

Automatic Dependent Surveillance–Broadcast (ADS-B) is a key source of air traffic surveillance data, but its open broadcast mechanism and possible trajectory-data inconsistencies may reduce surveillance reliability. This study proposes a Short-window Temporal Latent-Consistency Network (STLC-Net) for window-level anomaly detection in real-world ADS-B surveillance data. Raw ADS-B messages are transformed into physically interpretable kinematic features and segmented into short sliding windows to capture local motion inconsistencies. STLC-Net combines LSTM-based temporal encoding with an encoder–decoder–encoder latent-consistency mechanism, and fuses reconstruction and latent-consistency discrepancies to produce anomaly scores. Experimental results show that STLC-Net achieves a PR-AUC of 0.952540, precision of 1.0, recall of 0.882682, and F1-score of 0.937685 under highly imbalanced conditions. Ablation studies further confirm the effectiveness of short-window representation, temporal modeling, and latent-consistency learning. These results indicate that STLC-Net provides an effective and physically interpretable framework for detecting rare kinematic anomalies in ADS-B trajectory surveillance data.

Fatigue in air traffic controllers (ATCOs), particularly within remote tower operations, poses a substantial risk to aviation safety due to its detrimental effects on vigilance, decision-making, and situational awareness. While electroencephalography (EEG) provides a promising avenue for objective fatigue monitoring, existing models often fail to adequately capture both the spatial dependencies across brain regions and the temporal dynamics of cognitive states. To address this challenge, we propose a novel EEG-based fatigue detection framework, Spatio-Temporal Graph with Center Loss Network (STG-CLNet), which jointly models topological brain connectivity and temporal EEG evolution. The model leverages a multi-stage graph convolutional network to encode spatial dependencies and a triple-layer LSTM module to capture temporal progression, while incorporating center loss to enhance feature discriminability in the embedding space. We constructed a domain-specific EEG dataset involving 34 ATCO participants operating in high- and low-traffic remote tower simulations, with fatigue labels derived from three validated subjective metrics. Experimental results demonstrate that STG-CLNet achieves superior classification performance (accuracy = 96.73%, recall = 92.01%, F1-score = 87.15%), outperforming several strong baselines, including LSTM and EEGNet. These findings underscore the potential of STG-CLNet for integration into real-time cognitive monitoring systems in air traffic control, contributing to both theoretical advancement and operational safety enhancement.

A swarm of Unmanned Aerial Vehicles (UAVs) comprises multiple UAVs that are capable of completing tasks beyond the capabilities of a single UAV. Due to the unique challenges of UAV missions, these vehicles often operate far from base stations, making network connectivity crucial for the successful completion of UAV swarm missions. However, existing methods do not account for strategies to maintain the overall connectivity of the UAV network when nodes are under attack. To address this issue, we propose a method named Adaptive Communication Power Control (ACPC) that dynamically adjusts UAV communication power to mitigate potential connectivity losses caused by node failures or malicious attacks. This adjustment ensures that the network can maintain information exchange among the remaining UAVs even in the event of disruptions. Additionally, we introduce a novel evaluation method to assess the overall connectivity of the network and validate ACPC through simulations of UAV swarm missions where some UAVs experience failures. Using this evaluation method, we measured the network’s state before and after attacks and recovery and calculated the additional energy consumption required by the UAVs. The results indicate that our method can increase the resilience of the UAV network by up to 5.36 times, while only raising the total communication energy consumption to 1.53 times.

The identification of influential nodes in complex networks is a hot topic among scholars. Classical methods commonly analyze single node information or static structures but seldom emphasize dynamic properties and the interactive influence of nodes. Here, we proposed an improved Dynamic-Sensitive centrality (IDS) method by considering the interactive influence of both the self and neighbor nodes. Based on six real aviation networks and the Susceptible Infected Recovered (SIR) spreading disease model, we simulated the actual spreading process within these networks. Relevant experiments were conducted through Kendall’s correlation coefficient, the imprecision function, and the complementary cumulative distribution function. The experimental results demonstrated that the IDS can more accurately identify the influential node and effectively differentiate the node influence in the network compared with other benchmark methods. Especially in the EU air-2 network, the IDS results in Kendall’s correlation coefficient are improved by 105% compared to the DS centrality.

The Automatic Dependent Surveillance-Broadcast (ADS-B) system is vital for modern air traffic control, enhancing both safety and efficiency. However, its lack of encryption and authentication introduces significant security risks. One major threat is the ADS-B spoofing attack, where attackers manipulate the ICAO address in ADS-B messages to create fake aircraft or impersonate trusted ones, potentially misleading pilots and air traffic controllers and leading to dangerous maneuvers. To address these challenges, we propose the Revin-Discrete Cosine Patch Attention (RDPA)—a novel method leveraging the RevIn mechanism, Discrete Cosine Transform, and patch mechanism within a Transformer architecture. This approach is specifically designed to detect spoofing attacks within ADS-B data that traditional analyses might overlook, including deceptive maneuvers mimicking normal flight behavior, sudden deviations in flight paths, and inconsistencies in transmitted data, all of which pose substantial risks to airspace security. Our RDPA model has been rigorously tested on large-scale ADS-B datasets, including simulated deceptive maneuvers across various flight stages such as departing, maneuvering, and landing. Experimental results show that RDPA excels in detecting spoofing attacks with high accuracy, achieving an F1-score of 99.09\%, significantly outperforming existing methods. This highlights RDPA’s potential as a crucial tool for enhancing air traffic security.

The different rail transit systems, such as regional railway and urban rail transit, cooperate to form the syncretic railway network (SRN). With the rapid development of SRN, the limited transportation capacity is inadequate to meet the booming passenger flow during peak hours, where cascading failures caused by large passenger flow become a threat to SRN. This paper adopts a state equation to depict the failure‐restoration process and investigates detailed restoration strategies considering the characteristics of recoverable and repeatable failures of stations. Specifically, three different restoration strategies—spontaneous restoration strategy (SRS), active restoration strategy (ARS), and hybrid restoration strategy (HRS)—are proposed, and the varying effects of restoration time, restoration probability, restoration objective, and restoration priority for SRN with ARS are compared. These restoration strategies are applied to the actual SRN in Chengdu, where it is found that HRS has a better effect than other strategies. Furthermore, stations in the metro network with higher passenger flow allocate more restoration resources to improve the robustness of SRN, while the restoration effect of SRN increases noticeably with the restoration coefficient and the reserve coefficient. The restoration strategies presented in this paper can improve the safety management of SRN.

Identifying the key nodes in complicated networks is an essential topic. A number of methods have been developed in recent years to solve this issue more effectively. Multi-attribute ranking is a widely used and efficient method to increase the accuracy of identifying the key nodes. Using k-shell iteration information and propagation threshold differences, we thoroughly analyze the node’s position attribute and the propagation attribute to offer a hybrid influence method based on information entropy. The two attributes will be weighted using the information entropy weighting method, and then the nodes’ influence ranking will be calculated. Correlation experiments in nine different networks were carried out based on the Susceptible–Infected–Recovered (SIR) model. Among these, we use the imprecision function, Kendall’s correlation coefficient, and the complementary cumulative distribution function to validate the suggested method. The experimental results demonstrate that our suggested method outperforms previous node ranking methods in terms of monotonicity, relevance, and accuracy and performs well to achieve a more accurate ranking of nodes in the network.

The identification of key nodes in complex networks is a hot topic. Therefore, it attracts increasing attention from different fields, like airline networks and social networks. To identify the key nodes in complex network, we suggest an improved gravity model method that takes propagation features into account. Relevant experiments were carried out in four actual airline networks based on the Susceptible Infected Recovered (SIR) model. First, we analyze the correlation between the proposed method and other benchmark methods.Then, Kendall’s correlation coefficient and the imprecision function were used as evaluation metrics to analyze and validate the proposed method. Empirical results reveal that the suggested method outperforms previous benchmark methods in terms of precision and effectiveness for identifying key nodes, especially in the US air network, where Kendall’s tau achieves a 107% improvement compared to the gravity centrality method.

The front-page news in authoritative newspapers usually represents extremely significant national policies. Accurately classifying the front-page news from an amount of news helps us to quickly acquire and deeply understand the changing political and economic situations. In this paper, we propose a front-page news classification model StackText based on stacking the textual context and attribute information of news. In the proposed model, we first balance the classification size in the training set through the weighted random sampling algorithm, then construct the context and attribute feature vectors through the textual embedding and statistical analysis of news, and finally pretrain a classifier StackNet based on a neural network to realize the front-page news classification in the testing set. Taking People’s Daily as the experimental example, we compare the proposed model with benchmark methods based on statistical learning and deep learning. Based on the news sets in four stages of People’s Daily, the experimental results of the front-page news classification show that the proposed model achieves the highest average accuracy and a better balance between precision and recall, which is verified by the corresponding -score.