chawlaps/ATFMTraj

ATFMTraj是一个用于航空交通管理的飞机轨迹分类数据集，旨在解决该领域轨迹数据通常缺乏标签的问题，从而支持分类和聚类模型的探索。数据集通过参考航空出版物进行标注，涵盖三个机场：仁川国际机场（RKSI，包括到达RKSIa和离场RKSId数据）、斯德哥尔摩阿兰达机场（ESSA，仅到达数据）和苏黎世机场（LSZH，仅到达数据）。数据来源包括Opensky数据库、瑞典民用航空交通控制数据集和相关研究数据集。预处理步骤包括清洗不完整轨迹、将地理坐标转换为笛卡尔坐标、以1秒间隔重采样、平滑处理、归一化坐标以及填充不等长轨迹。标注过程基于K-means聚类和手动调整，涉及到达走廊、跑道标签、进近程序等，最终标签由跑道、IAF和STAR组合定义。数据集适用于机器学习任务，如轨迹分类和表示学习。ATFMTraj is an aircraft trajectory classification dataset for air traffic management, designed to address the lack of labels in trajectory data within this field, thereby enabling exploration of classification and clustering models. The dataset is annotated by referring to aeronautical publications and covers three airports: Incheon International Airport (RKSI, including arrival RKSIa and departure RKSId data), Stockholm Arlanda Airport (ESSA, arrival data only), and Zurich Airport (LSZH, arrival data only). Data sources include the Opensky database, the Swedish Civil Air Traffic Control dataset, and related research datasets. Preprocessing steps involve cleaning incomplete trajectories, converting geographic coordinates to Cartesian coordinates, resampling at 1-second intervals, smoothing, normalizing coordinates, and padding trajectories of unequal lengths. The annotation process is based on K-means clustering and manual adjustments, covering arrival corridors, runway labeling, approach procedures, etc., with final labels defined by combinations of runways, IAFs, and STARs. The dataset is suitable for machine learning tasks such as trajectory classification and representation learning.