论文部分内容阅读
为了准确掌握终端区空中交通流复杂多变的空间分布特征,有效评估、优化进离场程序,基于重采样技术研究了终端区三维真实飞行轨迹的聚类问题,提出了一种计算速度快、可扩展性强、可信度高的聚类方法.首先,结合重采样和主成分分析方法,将高维轨迹数据在保留飞行特征的前提下映射到低维空间;其次,基于Mean Shift方法建立飞行轨迹聚类分析与异常轨迹提取模型;最后,利用终端区的真实飞行轨迹数据进行实例验证,并分析模型中各个参数对聚类结果的影响.研究结果表明:该方法耗时0.004 s得到累计贡献率为96.16%的主成分,较好地逼近原始飞行轨迹数据;相较于层次聚类法,本文方法得到的飞行轨迹聚类结果具有更高的可信度,能够准确对应机场标准进场航线设置,并将相似度较低的飞行轨迹提取为异常轨迹.
In order to accurately grasp the spatial and temporal distribution characteristics of air traffic flow in the terminal area, to effectively evaluate and optimize the departure procedure, the clustering problem of the real three-dimensional trajectory in the terminal area was studied based on the resampling technique. A fast, High scalability and high reliability.Firstly, based on the method of resampling and principal component analysis, the trajectories of high-dimensional trajectories are mapped to low-dimensional space with the preserved flight characteristics.Secondly, based on the Mean Shift method Flight trajectory clustering analysis and abnormal trajectory extraction model.Finally, the real flight trajectory data of the terminal area are used to verify the model, and the influence of each parameter on the clustering results is analyzed.The results show that this method takes 0.004 s to accumulate Compared with the Hierarchical clustering method, the clustering results of flight trajectory obtained by this method have higher credibility and can accurately correspond to the airport standard approach Route setting, and extract the flight trajectories with low similarity as abnormal trajectories.