空管威胁、差错与意外状态的贝叶斯模型分析研究

Study on Threat,Error and Undesired States in Air Traffic Control Based on Bayesian Method

为准确理解民航空中交通管理中各类人为差错数据的可靠性和相关性,为安全决策提供量化依据,对空管威胁、差错和意外状态数据进行贝叶斯模型分析。该方法在先验概率基础上,通过研究空中意外状态发生条件下威胁与差错各自发生的后验概率,获得上述3类安全信息的相关性大小排序。针对收集到的625起数据样本进行了研究,结果表明:与空中意外状态的发生相关性最大的3类差错依次为:管制员对机组的通讯差错、信息差错以及管制员对管制员的通讯差错;与空中意外状态的发生相关性最大的3类威胁依次为:程序威胁、相邻单位威胁和空中交通威胁。空管运行单位的安全管控应重点关注管制员/机组间的通讯差错,并针对各种可能出现的威胁预先制定应对程序。

To accurately understand the reliability and correlations of all kinds of human error data in civil air traffic control and thus provide quantitative basis for decision-making, an analysis of the data about the threat, error and unexpected state in air traffic control was conducted by using Bayesian model. This analysis, based on the prior probability, obtained the sequence of the correlations of the three kinds of safety information above through studying the respective posterior probabilities of threats and errors under unexpected states. Investigation on 625 sample data shows that the three kinds of errors most correlated to unexpected state are communication error of ATCs （ air traffic controllers） to pilots, information error, and communication error of ATCs to ATCs in order; and that the three kinds of threats most correlated to unexpected state are procedure threat, threat from adjacent units, and air traffic threat in order. Therefore, the safety control of ATC units should lay their emphasis on the communication errors between ATC controller and pilots and on making corresponding procedures against all kinds of oossible threats.