PhD in Regulatory Compliance

Autonomous vehicles (AVs) are advancing rapidly, yet their safety assurance remains limited by the opacity of modern AI systems. Current machine learning–based decision pipelines often function as “black boxes,” making it difficult for regulators, engineers, and the public to understand or verify why an AV behaved in a particular way. This lack of transparency is a major barrier to certification, public confidence, and safe deployment on public roads. There is a pressing need for interpretable, verifiable methods that bridge the gap between complex autonomous behaviour and established safety and regulatory standards. This PhD project, co-funded by the Department for Transport, will investigate Digital Commentary Driving (DCD) as a novel approach to explaining and assessing AV decision-making. Inspired by human commentary driving, DCD provides a structured, machine-readable account of an agent’s perceptions, reasoning, and intent. The project will formalise this concept for AVs and explore how DCD can be used to evaluate behavioural safety, regulatory compliance, and alignment with human expectations. The research will: Analyse current AV safety assessment practices and compare them with methods used in other safety-critical sectors. Review international AV regulations and identify gaps in explainability and traceability. Develop a formal DCD model capable of representing AV reasoning and decision processes. Use the CARLA simulation platform to generate DCD-style data in high-risk or ambiguous driving scenarios. Build a proof-of-concept verification pipeline that maps DCD outputs to measurable safety metrics and relevant regulatory norms. Automate this pipeline to support scalable, simulation-based compliance checking. The outcomes will include a formal DCD framework, a prototype verification system for regulatory assessment, and high-quality publications. Working closely with the Driver and Vehicle Standards Agency (DVSA), the student will gain valuable insight into real-world regulatory processes, with opportunities for short-term placement. The project offers comprehensive training in safe AI, robotics, simulation, and transport regulation, preparing the student for an impactful career in academia, government, or the AV industry.