摘要
针对现航空器异常行为的检测,主要集中在位置异常和速度异常,缺少对航空器垂直方向上异常的检测,提出对航空器爬升下降率异常检测的研究。由参数估计对航空器爬升下降率异常进行定义;分析航空器爬升下降率影响因素,确定航空器爬升下降率异常影响因子。基于分层抽样子集训练随机森林分类器,用随机森林分类器模型对测试集进行检测,确定测试样本集所属类别。实验结果表明,上述模型在航空器爬升率检测方面具有一定可行性。
In view of the detection of abnormal behavior of the aircraft, it mainly focuses on positional anomalies and speed anomalies, and lacks the detection of anomalies in the vertical direction of the aircraft. Therefore, the research on the abnormal detection of aircraft climb and descent rate is proposed. The parameter estimation was used to define the aircraft’s climb rate reduction anomaly;the factors affecting the aircraft’s climb rate were analyzed, and the aircraft’s abnormal climb rate was determined;the random forest classifier was trained based on the stratified sampling subset, and the random forest classifier model was used to test the set. A test was performed to determine the category of the test sample. The experimental results show that the model has certain feasibility in aircraft climb rate detection.
作者
李楠
靳辉辉
樊瑞
LI Nan;JIN Hui-hui;FAN Rui(College of Air Traffic Management,Civil Aviation University of China,Tianjin 300300,China)
出处
《计算机仿真》
北大核心
2020年第5期25-29,共5页
Computer Simulation
基金
国家自然科学基金民航联合研究基金(U1533112)
国家重点研发计划课题(2016YFB050).
关键词
终端区
爬升下降率
异常检测
置信区间
随机森林
Terminal area
Climbing rate
Anomaly detection
Confidence interval
Random forest