In recent years, a large number of commercial airplanes equipped with sensors have been contributing to collecting the atmospheric composition and meteorological data. Based on a worldwide commercial airplane dataset, we identify an inherent limitation of such an airplane-based sensing system as airplanes fly through airspace with spatiotemporal dynamic. As a result, the sensed data is coarse-grained and incomplete. Moreover, naive methods such as installing sensors on more airplanes, cannot effectively address the issue. In this paper, we propose a collaborative airspace sensing strategy auSense to improve the sensing performance. In auSense, UAVs (Unmanned Aerial Vehicles) are employed to sense the airspace when the airspace is not sensed by airplanes. Particularly, auSense leverages the existing worldwide commercial airplane tracking infrastructure and fine-grained trajectory data to estimate whether and when one airspace will be sensed by airplanes. Then auSense schedules the UAV paths to improve the sensing performance. We implement auSense in different airspaces based on a 1-month airplane trajectory data. The experiments demonstrate the superior performance of our data-driven UAV and commercial airplane collaborative strategy over the ground truth and other benchmarks.

中文翻译：

---

auSense：商用飞机和无人机的协同空域感知

近年来，大量配备传感器的商用飞机为收集大气成分和气象数据做出了贡献。基于全球商用飞机数据集，我们确定了这种基于飞机的传感系统的固有局限性，因为飞机以时空动态飞过空域。结果，感测到的数据是粗粒度且不完整的。此外，在更多飞机上安装传感器等幼稚的方法无法有效解决问题。在本文中，我们提出了一种协作空域感知策略 auSense 来提高感知性能。在 auSense 中，当飞机无法感知空域时，会使用 UAV（无人驾驶飞行器）来感知空域。特别，auSense 利用现有的全球商用飞机跟踪基础设施和细粒度轨迹数据来估计飞机是否以及何时会感知一个空域。然后 auSense 调度无人机路径以提高传感性能。我们根据 1 个月的飞机轨迹数据在不同的空域中实施 auSense。实验证明了我们的数据驱动无人机和商用飞机协作策略在地面实况和其他基准测试中的优越性能。