Cyber-Physical Dataset for UAVs Under Normal Operations and Cyber-Attacks

Unmanned aerial vehicles (UAVs) are being used for various applications, but the associated cyber risks are also increasing. Machine learning techniques have been successfully adopted to develop intrusion detection systems (IDSs). However, none of the existing works published the cyber or physical datasets that have been used to develop the IDS, which hinders further research in this field. Hence, we developed a testbed that includes a UAV, controller, and data collection tools, to launch cyber-attacks (de-authentication denial-of-service (DoS) and replay attacks) on the UAV and collect cyber and physical data under normal and attack conditions, and we make this dataset publicly available.无人机（Unmanned Aerial Vehicles, UAVs）已被广泛应用于各类场景，但伴随而来的网络风险也日益攀升。机器学习技术已被成功用于开发入侵检测系统（Intrusion Detection Systems, IDSs）。然而，现有相关研究均未公开其用于开发IDS的网络或物理数据集，这极大掣肘了该领域的后续研究。为此，我们搭建了包含无人机、控制器及数据采集工具的测试平台，对无人机发起认证取消拒绝服务（Denial-of-Service, DoS）攻击与重放攻击两类网络攻击，并采集了正常工况与攻击工况下的网络与物理数据，同时将该数据集公开共享。