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.

Accurate and efficient object detection is essential for aerial airport surveillance, playing a critical role in aviation safety and the advancement of autonomous operations. Although recent deep learning approaches have achieved notable progress, significant challenges persist, including severe object occlusion, extreme scale variation, dense panoramic clutter, and the detection of very small targets. In this study, we introduce a novel and unified detection framework designed to address these issues comprehensively. Our method integrates a Normalized Gaussian Wasserstein Distance loss for precise probabilistic bounding box regression, Dilation-wise Residual modules for improved multi-scale feature extraction, a Hierarchical Screening Feature Pyramid Network for effective hierarchical feature fusion, and DualConv modules for lightweight yet robust feature representation. Extensive experiments conducted on two public airport surveillance datasets, ASS1 and ASS2, demonstrate that our approach yields substantial improvements in detection accuracy. Specifically, the proposed method achieves an improvement of up to 14.6 percentage points in mean Average Precision (mAP@0.5) compared to state-of-the-art YOLO variants, with particularly notable gains in challenging small-object categories such as personnel detection. These results highlight the effectiveness and practical value of the proposed framework in advancing aviation safety and operational autonomy in airport environments.

Short-horizon emergency re-optimization of airport apron ground-support operations works when its objective carries timing stability, and backfires when it does not. Using a new benchmark—89 days of Amsterdam and Chicago ADS-B ground traces including transponder-equipped ramp service vehicles, with semi-synthetic task and fault layers: 3540 counterfactual-controlled fault episodes and 740 test reschedule instances with CP-SAT references (85% proved optimal)—we score every recovery over the 30–120 min emergency horizon using a full-day shadow replay against a matched no-fault counterfactual, with date-level paired inference. CP-SAT with a start-time stability term, or with a lexicographic rule that returns the most plan-preserving plan attaining the first-stage window objective value, matches plan-preserving repair on realized disruption on vehicle-removal faults (median 1.6–2.4 vs. 2.7 ks) and lowers absolute realized tardiness on 19/19 test dates. The same model minimizing only in-window weighted tardiness is worse end-to-end than greedy earliest-due-date dispatch on 19/19 dates (exact $p=3.8\times {10}^{-6}$). For instances certified as optimal at every churn budget, the unconstrained solve reaches the window optimum of the minimal-churn solve yet realizes 2.6 times more disruption; learned re-optimizers trained on the same objective are worse still, and rolling-horizon repetition amplifies the harm. The harm travels through re-timing and re-sequencing rather than reassignment, so an assignment-only churn penalty does not help. The typical instance has no in-window headroom over greedy dispatch (median learning space 0). The ordering holds across ten generator settings and service-time noise. The benchmark will be made publicly available.