With the increasing activities of unmanned aircraft (UA) near airports, severe challenges have emerged regarding airport safety management. This paper presents a brand-new data-driven based approach to accurately design the geofencing for UAs around an airport, focusing on midair collision risk between commercial flights and small UAs which is considered as the most important hazard. In premise, prevalent trajectory extraction from real flight tracks is realized based on an advanced and general model, taking the advantage of describing turning legs which is more popular in terminal area. As a key solution, a Gauss-Laplace-composite distribution is provided for rigorous estimation of flight track distribution in line with the extracted prevalent trajectory. Based on those, airport geofencing is proposed consisting of an inner critical area and an outer buffer area. This approach can be adaptive for different traffic patterns of an airport. The buffer area is an extension range to help the execution of some defense measures. Results of an empirical study for setting up geofencing for small UAs around Chongqing Jiangbei International Airport (ZUCK) show the superiority of the proposed approach, which indicates that it can provide practical application, especially varying from different target level of safety.

随着机场附近无人飞机（UA）活动的增加，在机场安全管理方面出现了严峻的挑战。本文提出了一种全新的基于数据驱动的方法，可以为机场周围的UA准确设计地理围栏，重点关注商业航班和小型UA之间的空中碰撞风险，这被认为是最重要的危害。在此前提下，利用先进的通用模型实现了从真实飞行轨迹的普遍轨迹提取，并利用了描述在航站区更受欢迎的转弯腿的优势。作为关键解决方案，提供了高斯-拉普拉斯复合分布，以便根据提取的流行轨迹严格估计飞行轨迹的分布。基于这些，建议将机场地理围栏包括内部关键区域和外部缓冲区。这种方法可以适应机场的不同交通模式。缓冲区是扩展范围，有助于执行某些防御措施。为重庆江北国际机场（ZUCK）周围的小型UA建立地理围栏的实证研究结果表明，该方法具有优越性，表明它可以提供实际应用，尤其是在不同的安全目标级别之间。