Researchers propose framework for safer self-driving cars

A team of researchers from Oxford, NASA, and industry partners has published a groundbreaking framework to address the critical but often overlooked psychological safety challenges in autonomous vehicles (AVs).

The study, titled 'Engineering Psychological Safety in Autonomous Vehicles,' introduces AV-PsySafe, a systems-theoretic framework that extends the traditional STAMP (Systems-Theoretic Accident Model and Processes) methodology to incorporate psychological constructs like trust, perceived control, and predictability. The framework includes a hazard analysis method to systematically identify psychological risks and introduces the Psychological Safety Integrity Level (PsySIL) for risk prioritization. Validated through real-world AV scenarios, the approach demonstrates consistent practical applicability and provides actionable insights for practitioners in AV development.

The research addresses a key barrier to AV adoption: while physical safety has seen significant advances, psychological safety—how users *feel* about interacting with AVs—remains understudied. The team’s framework bridges this gap by formalizing psychological risk assessment alongside physical safety, offering a unified approach for developing more trustworthy and human-centered autonomous systems.