With the development of Vehicle-to-Everything (V2X) technology, connected autonomous vehicles (CAVs) share basic safety messages containing speed and location data to improve traffic efficiency and safety. However, it also exposes CA V s to cyberattack threats. Therefore, this paper proposes a cyberattack detection method for CA V s based on hierarchical vector Transformer. Firstly, various cyberattacks including DoSA, FDIA, and SybilA are injected into the publicly available Argoverse dataset, constructing a cyberattack dataset in V2X environment. Secondly, a cyberattack detection method based on hierarchical vector Transformers model is proposed, where the network hierarchically extracts vehicle location information, detects abnormal vehicles, identifies cyberattack types. Finally, the detection performance of the method under different cyberattack densities is verified in the constructed dataset. Experiments demonstrate that the proposed method has satisfactory detection performance and can meet real-time requirements.
Hierarchical Vector Transformer-Based Cyberattack Detection for Connected and Autonomous Vehicles via Cloud Platform
25.10.2024
866439 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
AIAA | 2017
|