摘要
为了实现智能化处理海量航班数据,精细描述终端区运行特性,研究了基于谱聚类的终端区飞行轨迹分析方法。在分析终端区航班飞行特点的基础上,提出基于航向因子修正的欧式距离轨迹相似度计算方法;利用高斯核函数对相似度矩阵进行平滑降噪处理,实现了函数中规模参数的自动化选取;采用改进谱聚类算法对终端区飞行轨迹进行聚类划分。利用广州白云机场进行实例验证,结果表明,方法能够有效处理终端区轨迹数据。
In order to process massive flight data intelligently, descript operational characteristics of the terminal area,the paper proposed trajectory analysis method in terminal area based on spectral clustering. First,based on the analysis of flight characteristics in terminal area ,the paper proposed similarity calculation method based on the Euclidean distance which corrected by course correction factor;then it used Gaussian kernel to smooth the similarity matrix and reduce noise,the scale parameter can be selected automatically;Finally ,the paper used improved spectral clustering algorithm to cluster terminal area flight paths.With the actual data of Guangzhou Airport,the method is verified to process trajectory in terminal area effectively.
出处
《航空计算技术》
2015年第5期46-50,共5页
Aeronautical Computing Technique
基金
国家科技重大支撑计划项目资助(2011BAH24B09)
关键词
空中交通
终端区
飞行轨迹
谱聚类
air traffic
terminal area
trajectory
spectral clustering