Safe and Explainable AI

Connected autonomous vehicle (CAV) technology prototypes have existed for some time, however questions around the safety of this technology from the public and industry who want to commercialise these technologies have blocked it from developing. AI technologies are at the heart of the CAVs especially in the perception sub-systems and are inherently non-deterministic in nature. Arguing for their safety and verifying AI-based systems requires the development of novel techniques. This project will focus on creating methods to argue for safety of AI systems in perception, planning and act subsystems of CAVs.

As the use of AI proliferates in our daily lives, such research questions will impact other sectors too, e.g. healthcare, Industry 4.0 and vehicle electrification.