Recognition Method of Aircraft Group Based on Hierarchical Clustering

In the research based on the air traffic complexity assessment of the controller's workload, the problem of the reduction of the controller load caused by the reduction of the traffic flow structure has gradually attracted people's attention. In the existing research, the cognitive complexity of the controller is reduced through the division of aircraft groups to reduce the workload. The traditional k-means division algorithm has the disadvantages that it is unable to distinguish aircrafts with different headings and clustered positions and the number of groups needs to be specified in advance. Aiming at the shortcomings of the k-means algorithm, this paper uses hierarchical clustering algorithm, by adding heading factors and improving the selection method of clustering results, to realize the automatic identification of the aircraft group in the traffic flow within the sector. The group division evaluation index is formed from the angle of position cohesion and heading profile coefficient, which is used to comprehensively reflect the effect of group division. Finally, the actual radar data is used for group division and evaluation and analysis, and the effectiveness of the method in this paper is verified by comparison with the k-means method.