基于异常因子的航空器飞行轨迹异常检测研究

On the abnormal detection of the aircraft flight trajectory based on the abnormal factor statistics

为保障终端区航空器飞行安全,提高空域运行效率,提出了一种基于异常因子的航空器飞行轨迹异常检测方法。首先,给出了航空器异常轨迹的定义和分类,同时考虑轨迹的时空属性,采用基于时间比的自顶向下算法进行轨迹分割;其次,建立了基于密度聚类的轨迹段相似性模型,找到每类簇的中心轨迹;最后,利用中心轨迹计算每个轨迹簇的信息熵作为异常因子,同时取航空器飞行距离为另一异常因子,得到轨迹异常值,通过比较异常值的大小及其分布特点实现轨迹异常检测。结果表明,该方法可以有效识别异常轨迹,且对比发现以轨迹信息熵和飞行距离作为异常因子时,异常检测效果更好。

In order to ensure the flight safety of the aircraft in the terminal area and improve the airspace operation efficiency,the paper intends to propose a method based on the anomaly factor to detect the abnormal flight trajectory of the aircraft.According to the said abnormal flight state of the aircraft in the authentic flying status-in-situ,it would be possible to give the definition and classification of the abnormal trajectory of the aircraft under way.At the same time,it is also possible for us to work out the original trajectory data from a large number of incomplete trajectories,as well as the takeoff and landing flight inclinations nearby the airport,so as to predict and preprocess the trajectories and the flight standard trajectories.And,then,the top-down algorithm can be done based on the time ratio to segment the trajectory by taking account of the space-time attribute of the trajectory.In addition,it would be also possible to lay out the similarity model of the trajectory segment based on the density clustering to detect the centeral trajectory of the trajectory cluster in accordance with the same flight mode.And,so,finally,the information entropy of each trajectory cluster can be calculated and gained as the anomaly factor by using the center trajectory so as to determine the flight distance of the aircraft as another anomaly in comparison with the size and distribution features of the outliers.Thus,the results of our study can show that:the method we have proposed can effectively identify the abnormal trajectory,whereas the effect of the anomaly detection can be found more comprehensivley in case the trajectory information entropy and flight distance can be taken as the abnormal factors.And,therefore,in so doing,the controller can also help to confirm that in the actual flight,the abnormal flight can be identified as seen in this paper with greater safety risk,which can thus lay an important foundation for regulating the pilot’s operation and getting rid of reckless events in the later stage.