Pilot factor is worth considering when analyzing the causes of civil aviation accidents. This study introduces a data-driven Bayesian network (BN) approach to investigating the joint causal effects of pilot and other factors on civil aviation safety. A total number of 163 individual pilot-related accidents in the National Transportation Safety Board (NTSB) aviation accident database from 2008 to 2020 are analyzed, focusing on eliciting the causal effects of various potential risk factors, including pilot factors, on civil aviation accidents. The modeling of the interdependency among the risk influencing factors (RIFs) and their causal contributory effect on the accident outcome is structured by a tree augmented network (TAN) and validated by sensitivity analysis. The novelty of this study is to incorporate pilot factors derived from the civil aviation accident database into risk analysis, combined with other external factors. The results indicate that weather conditions and flight phases are more correlated with casualty types of civil aviation accidents than pilot action and decision, and three other pilot factors only contribute to fatal injury in civil aviation accidents.

Flight safety is a key issue in the aviation industry. Recently, with the prevalence of flight data recording systems, some deep learning-based studies have been devoted to predicting safety incidents based on flight data. However, these studies, although they exhibit higher prediction accuracy, have largely neglected the interpretability analysis of safety incidents which is of great concern to airlines and pilots. To address this issue, we define flight safety prediction as a multiscale time series classification problem and propose an interpretable model named IMTCN to provide both accurate predictions and high interpretability of flight safety. First, multiple temporal convolutional networks (TCNs) are utilized to capture local representations and long effective histories from multivariate flight data. Because different flight parameters are collected with diverse sampling frequencies, multiple TCNs are used to handle these parameters separately. Then, we creatively adapt the class activation mapping (CAM) method, which has been used for interpretation in image classification, and combine it with the TCN to provide flight data interpretability. The established model can pinpoint key flight parameters and corresponding moments that contribute most to safety incidents. Experimental results on a real-world dataset with 37,943 Airbus A320 aircraft flights show that our model outperforms the baselines on the task of exceedance classification and prediction 2 seconds and 4 seconds in advance, and case studies demonstrate its superb interpretability for flight safety analysis.

In the face of explosive growth of wireless data and electronic devices, cognitive radio (CR) technology is an effective solution to identify spectrum resources underutilized by the primary user (PU). Moreover, cooperative spectrum sensing (CSS) is the key function of CR technology to enhance the detection performance by exploiting spatial diversity via the observations of spatially located sensor nodes (SNs). However, in a classical cooperative paradigm, each SN is always required to report individual local decision about the PU status to the fusion center (FC) for the global decision-making. Such a static reporting model regardless of the surrounding environment is not conducive to achieving the cooperative gain and adapting to a large cognitive wireless sensor network (CWSN). In view of this, this paper proposes an efficient CSS to optimize the reporting time and the achievable throughput of CWSNs, in which consists of selective reporting, multi-bit quantization, data fusion, and dynamic parameter. First of all, SNs only report some necessary observations by means of selective reporting, and then encodes local observations into multi-bit binary codes and send them to the FC. Further, the received local observations are aggregated into a final decision about the PU status. To be specific, the bit number of binary data and the selective reporting range as a parameter are dynamically adjusted at each SN. At last, simulation results demonstrate that the proposed CSS algorithm significantly not only reduces reporting time, but also improves the achievable throughput of CWSNs while maintaining the CSS performance.

Trajectory prediction (TP) is a vital operation in air traffic control systems for flight monitoring and tracking. The approach phase of general aviation (GA) aircraft is more of a visual approach, which is related to the safety of the flight and whether to go around. Therefore, it is important to accurately predict the flight trajectory of the approach phase. Based on the historical flight trajectories of GA aircraft, a TP model is proposed with deep learning after feature extraction in this study, and the hybrid model combines a time convolution network and an improved transformer model. First, feature extraction of the spatiotemporal dimension is performed on the preprocessed flight data by using TCN; then, the extracted features are executed by adopting the Informer model for TP. The performance of the novel architecture is verified by experiments based on real flight trajectory data. The results show that the proposed TCN-Informer architecture performs better according to various evaluation metrics, which means that the prediction accuracies of the hybrid model are better than those of the typical prediction models widely used today. Moreover, it has been verified that the proposed method can provide valuable suggestions for decision-making regarding whether to go around during the approach.

Understanding the structure of the Chinese railway network (CRN) is crucial for maintaining its efficiency and planning its future development. To advance our knowledge of CRN, we modeled CRN as a complex weighted network and explored the structural characteristics of the network via statistical evaluations and spatial analysis. Our results show CRN as a small-world network whose train flow obeys power-law decaying, demonstrating that CRN is a mature transportation infrastructure with a scale-free structure. CRN also shows significant spatial heterogeneity and hierarchy in its regionally uneven train flow distribution. We then examined the nodal centralities of CRN using four topological measures: degree, strength, betweenness, and closeness. Nodal degree is positively correlated with strength, betweenness, and closeness. Unlike the common feature of a scale-free network, the most connected nodes in CRN are not necessarily the most central due to underlying geographical, political, and socioeconomic factors. We proposed an integrated measure based on the four centrality measures to identify the global role of each node and the multilayer structure of CRN and confirm that stable connections hold between different layers of CRN.

Excessive workload caused by weather conditions may increase pilot errors and flight risk. The assessment of pilot workload using pilots' physiological data is a potential method to address this issue. We conducted a flight simulator study involving 18 cadet pilots with real flight experiences ranging from 232 to 250 hours and used non-invasive functional Near-Infrared Spectroscopy (fNIRS) to measure the pilots' brain activity. Pilots' subjective National Aeronautics and Space Administration-task load index (NASA-TLX) ratings of 0000-0000 © 2022 IEEE workload were also recorded. The tested flight maneuvers included a total of 54 takeoff and climb tasks under different weather conditions. Over 1100 features covering three hemoglobin signals and four brain cortexes were extracted from the fNIRS data. A statistical analysis was carried out on the subjective ratings and fNIRS features, and a weighting analysis was applied to the selected fNIRS features that were highly sensitive to pilot workload levels. A Convolutional Neural Network (CNN) with an attention mechanism (CNN-Attention) was built as a classifier for assessing pilot workload and compared with CNN, Deep Neural Network (DNN), eXtreme Gradient Boosting (XGBoost), and Random Forest (RF) models. The results suggested that pilot workload was highly associated with three hemoglobin measures as well as the activities of different brain regions, including the prefrontal cortex, motor cortex, and occipital cortex, and the CNN-Attention performed best. The findings from this study could provide a reference for optimizing pilot training systems and improving flight safety under different weather conditions