基于划分检测模型的终端区异常轨迹检测方法  

Detection Method of Trajectory Outliers in Terminal Airspace Based on a Partition-and-Detect Model

作　　者：任杰[1] 韩邦村[1] 

机构地区：[1]中国民航大学空中交通管理学院,天津300300

出　　处：《航空计算技术》2013年第6期35-38,共4页Aeronautical Computing Technique

基　　金：国家自然科学基金重点项目资助(61039001);中央高校基本科研业务费专项资助项目(3122013SY03)

摘　　要：针对异常轨迹对轨迹聚类效果的影响以及给进离场程序管制适用性的量化分析提供分析数据的考虑,在结合数据挖掘中异常检测理论的基础上,利用划分检测框架及将距离和密度结合起来,设计了一种有效检测算法,运用到终端区飞行轨迹的异常检测当中,检测出飞行轨迹中的异常轨迹,从而改善轨迹的聚类效果。实例仿真结果表明:方法可以准确地检测出异常轨迹。For trajectory outliers effect on trajectory clustering and provide data to air traffic control ser- viceability of current terminal arrival/departure procedures to quantitative analysis. Based on the theory of data mining in anomaly detection, designs an effective detection method by used partition- and- detect framework together with Euclidean distance and density. Apply this method to detect trajectory outliers in terminal area, thereby improving the trajectory clustering results. Application results show that this method can accurately detect the abnormal trajectories.

关 键 词：空中交通管理 空中交通流 异常轨迹检测 划分一检测模型 欧式距离 密度 

分 类 号：V328[航空宇航科学与技术—人机与环境工程]