Sensor attack online classification for UAVs using machine learning

Xiaomin Wei, Yizhen Xu, Haibin Zhang, Cong Sun, Xinghua Li, Fenghua Huang, Jianfeng Ma

Abstract: Highlights•The first work to classify sensor attacks in real-time for UAVs. The accuracy is 89.38%.•Design and develop multiple sensor injection attack methods.•A multi-sensor attack classification using key features and heterogeneous data.•Deploy the model in UAV firmware with detection time 0.495 ms and storage 48 KB.