Risk Identification And Analysis Based On Operational Data Of Air Traffic Control

In recent years, the domestic civil aviation industry has always showed a sustained and rapid development momentum, the air traffic flow increasing, at the same time facing limited special system of air traffic management, and the ability of existing safeguards, would inevitably put forward higher requirements to civil aviation ATC operation, so the research topic of ATC safe operation is gradually becoming the focus of the civil aviation scholars. Taking into account the civil aviation ATC operation work is achieved by staff, man-machine interaction information, therefore starting from the historical data records of interaction information, by means of data mining techniques being widely used in other industries to conduct the safe operation research of air traffic control and to identify potential law behind historical operating data, provides supporting suggestion for the ATC decision-making.In this paper, starting from the historical radar track data, make a preliminary exploration for the problem of risk scenario identification in operation process of aircraft by means of data mining techniques. First, for the original radar track data, according to the characteristics of the data for effective data preprocessing, form the target trajectory data; then quickly search flight conflicts from the target trajectory data, and generate sample data set of risk scenarios after the effective description for flight conflicts; and use classical K-means partitioning method for space division of risk scenarios, forming a plurality of flight conflict areas, for each dense regions identify the typical risk scenarios based on DBSCAN clustering, discovering those risk scenarios with the higher frequency; and finally, make the statistical analysis from a macro point of conflict over time, altitude, speed changes, then from the microscopic make the conflict risk analysis, respectively based on the ellipsoid distances model for the analysis of conflict risk scenarios without ATC intervention, and based survival analysis model to model controllers conflict detection process, and at the same time build a simple model to resolve the conflict and achieve the indirect risk analysis of conflict risk scenarios under ATC intervention.