Project 3.1 – Future Safety Assurance

Verification and validation of CAV safety

How can the safety of CAVs be effectively verified and validated to operate in rural contexts?

The verification and validation suitable for urban contexts must adapt to the fundamentally different rural contexts.

This project will define novel context-aware deep learning methods such as Generative Adversarial Networks (GANs) based data generation algorithms, which are suitable to support a wide range of context changes and transformations seeded by recorded real situations. The algorithms will be modular to verify and validate different sub-components of CAVs in physical and virtual testing environments, extending Project 1.4. It will identify metrics and references for Safe System behaviour and performance, covering both technology and human factors as discovered from UC1-3.