CANdid - An Open-Access Annotated Dataset of Vehicle CAN Bus Traffic

Learn about a groundbreaking open-access dataset for automotive cybersecurity research in this 16-minute conference presentation from VehicleSec '25. Discover how researchers from the University of Adelaide and Defence Science and Technology Group Australia have created CANdid, a comprehensive collection of Controller Area Network (CAN) bus traffic data captured from ten different vehicles operating under both real-world and controlled conditions. Explore the challenges of understanding CAN data formats that vary across vehicle manufacturers and how this dataset addresses the limitations of existing publicly available CAN datasets. Examine the unique features of CANdid, including synchronized video footage of driver actions, GPS information, and clearly labeled vehicle operations that enable detailed analysis of automotive network communications. Follow two practical case studies demonstrating the dataset's applications: automatic identification of vehicle components using labeled actions, and machine learning-based detection of vehicle turning maneuvers from raw CAN data. Understand how this comprehensive dataset enables researchers to gain deeper insights into CAN implementation variations across manufacturers and advance automotive cybersecurity research with access to diverse, well-documented real-world vehicle network data.