关键词: 空管安全管理, 威胁与差错管理模型, 模型本体, 信息抽取, 相关性分析

Abstract: To address the problems of identifying key information and managing potential risks in the hazard control list of the air traffic control (ATC) system, this paper proposes a correlation analysis method for ATC safety information based on the Threat and Error Management (TEM) model. Firstly, the TEM model ontology is constructed according to the internal classification of the TEM model. Then, information extraction techniques are employed to extract structured events from the text of the hazard control list. The deep autoencoder network model and an improved cosine similarity formula are used to match the text semantic similarity. Finally, a Bayesian network is applied to quantitatively analyze the probabilistic relationships among the TEM ontology categories, obtaining the correlations between threats, errors, and undesired aircraft states. Through a case study using data from a civil aviation ATC hazard control list, key information, entity words and event types strongly related to ATC operational safety are identified and extracted. Based on the extraction results, reasonable suggestions are proposed for enhancing the safety of various ATC operational phases

Key words: air traffic control (ATC) safety management, threat and error management (TEM) model, ontology, information extraction, correlation analysis