为采用航迹数据分析,优化现有的空域结构,介绍使用航迹数据聚类的分析方法。在模糊C均值聚类(FCM)的基础上,提出对FCM进行基于遗传模拟退火算法(GSAA)的改进和优化。首先分析广播式自动相关监视(ADS—B)航迹数据的特点,处理异常数据并使用模糊聚类法直接提取航迹特征点;然后运用FCM与优化后的FCM对航迹特征点进行聚类分析;最后用2种算法针对同一实例进行计算,并构建对应的平均航迹。结果表明:与FCM相比,基于GSAA的改进FCM算法的目榱甬辨信蝻/b1气,n%骆卷出JI、,占不△伯由士暑埔幸冰辨杯占催出曲估罾
To optimize the present airspace structure using analysis of trajectory data, a trajectory data clustering analysis method was worked out. The basic FCM was improved on basis of GSAA. The features of trajectory were analyzed and abnormal values were handled. The proper values were extracted, which could represent the actual trajectory, by fuzzy subtractive clustering. They were clustered separately by basic FCM and improved FCM based on GSAA, and average tracks were made. At last, an example was calculated by two algorithms. The simulation results show that using improved FCM, the objective value is reduced by 15.20% , comparing with that by FCM, and that with improved FCM, the clustering center would not to location where a large number of trajectory data points muster.