民航空管电子进程单数据分析技术研究

发布时间：2018-01-11 23:07

本文关键词：民航空管电子进程单数据分析技术研究　出处：《中国民航大学》2017年硕士论文　论文类型：学位论文

【摘要】：在民航空管的管制工作中,管制员利用电子进程单实时掌握航空器运行的状态。它能够记录航班管制指令,记录航班关键数据,协助管制员的管制协调和移交。中国民航总局空管局关于下发《电子进程单应用》(咨询通告)的通知中提出电子进程单系统应能够记录管制指令和管制工作过程,故在使用过程中会产生大量的操作日志数据,但是目前还没有相关电子进程单系统日志的研究分析。而储存的这些真实的日志记录以及电子进程单数据,为分析管制员是否存在异常行为提供数据支持,另一方面,为安全事故的调查提供证据。为了能够发现民航空管管制员的异常行为和便于空管事故的调查,根据数据预处理、序列模式发现和行为相似度计算等数据分析方法,结合民航空管塔台电子进程单日志数据量庞大,结构复杂,大量冗余等特点,在实验的基础上,研究管制员行为异常的检测方法。基于空管的塔台管制的电子进程单系统的操作日志,提出一个管制员行为的异常检测算法。首先采用自定义的数据预处理方法对电子进程单日志进行预处理,具体包括数据清洗、管制席识别、管制员值班时间匹配、数据整合转存等;然后从数据规模、行为序列特点以及实现难易程度等方面考虑,采用Apriori All算法得到五类管制席的行为序列模式,最后采用递归式相关函数法计算管制员的行为序列与其所属管制席的序列模式的相似度,通过与阈值比较,以此来判断管制员的行为的异常情况,最后在算法的基础上提出了一个管制员行为异常检测系统设计。
[Abstract]:In the control work of civil aviation air traffic control, the controller uses electronic process sheet to grasp the status of aircraft operation in real time. It can record flight control instructions and record flight critical data. Assist in the coordination and transfer of controllers. CAAC (Advisory notice). It was stated in the notification that the electronic process single system should be able to record control directives and control processes. Therefore, a large number of operation log data will be generated in the process of use, but there is no research and analysis of the related electronic process single system log at present. These real log records and the electronic process single data are stored. In order to find the abnormal behavior of the air traffic controllers and facilitate the investigation of the air traffic control accidents, this paper provides data support for the analysis of the abnormal behavior of the controllers, on the other hand, provides evidence for the investigation of the safety accidents. According to the data preprocessing, sequence pattern discovery, behavior similarity calculation and other data analysis methods, combined with the civil aviation air traffic control tower electronic process single log data volume large, complex structure, a large number of redundancy and so on. On the basis of the experiment, the detection method of the abnormal behavior of the controller is studied. The operation log of the electronic process single system based on the air traffic control tower is studied. This paper proposes an anomaly detection algorithm for controller behavior. Firstly, a self-defined data preprocessing method is used to preprocess electronic process single log, including data cleaning and controller identification. Controller on duty time matching, data integration and storage; Then from the aspects of data scale, behavior sequence characteristics and the degree of difficulty in implementation, the Apriori All algorithm is used to obtain the behavior sequence pattern of five kinds of regulators. Finally, the recursive correlation function method is used to calculate the similarity between the controller's behavior sequence and the sequence pattern of the controller's controller, and to judge the abnormal behavior of the controller by comparing with the threshold. Finally, a controller behavior anomaly detection system is designed based on the algorithm.
【学位授予单位】：中国民航大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：V355.1;TP311.13

【参考文献】