Anomaly Detection, Verification and Validation via a Mini DrivingCube Demonstrator (ADMIRE)

Purpose and goal

The increasing integration of self-driving functions and connectivity in modern vehicles create more vulnerabilities, and thus more risks for cyber security attacks. Therefore, verification and validation of vehicles from a cybersecurity perspective requires new approaches. In the ADMIRE project, we conducted a study on the use of AI in conjunction with HiL (Hardware in the Loop) tests to identify attacks based on zero-day exploits and with safety implications in the European transport infrastructure.

Expected results and effects

The project has set up a demonstrator in the form of a 1/10 Mini DrivingCube, which has potential to move to a larger scale. ADMIRE has demonstrated the feasibility and value of continued research on: * Scaling down the HiL test system to increase the number of those that can be realistically deployed. * Establish scaled-down HiL test systems that mimic field conditions more closely. * AI and statistical models that link safety to safety. Together with additional partners in three more countries, the project has submitted a European project proposal on the theme of AI and cybersecurity.

Planned approach and implementation

The project plan and work packages have worked well, but the ideas and ambitions would have been sufficient for a larger project where we would have benefited from engaging more persons with additional competencies to reach even further. We also note that the topics of personal safety and cyber security lead to organizational challenges as the issues are often handled by different departments within the same organization. This has become clear, not least when we worked with several organizations to write the European project proposal.