The research area for autonomous vehicle's anomaly detection using machine learning hadn't considered using an Echo State Network with Federated Learning to create a fast hybrid-machine learning algorithm for the classification of various attacks that happen with autonomous vehicles.
The project uses datasets from these two places:
- Car Hacking which consists of five files such as
DoS_dataset.csv,Fuzzy_dataset.csv,gear_dataset.csv,RPM_dataset.csv, andnormal_run_data.7z. - CAN Intrusion which consists of four files such as
Attack_free_dataset.txt,DoS_attack_dataset.txt,Fuzzy_attack_dataset.txt, andImpersonation_attack_dataset.txt.
To gain access to these datasets, your e-mail and reasoning for wanting them will have to be shared with the authors by using a google form.