VITA: Vehicle severity-aware control for automated driving

Research Project, 2027 – 2028

Trustworthiness is a prerequisite for the societal acceptance of autonomous vehicles (AVs): for humans to delegate life-critical decisions to an algorithm, that algorithm must be transparent about its priorities and demonstrably aligned with human safety values. Yet current AV path planning frameworks optimize for collision avoidance alone, leaving injury-related priorities undefined precisely in the scenarios where they matter most. When a collision cannot be fully avoided, existing frameworks have no mechanism to reason about or discriminate between collision consequences.  

This project develops a severity-aware Model Predictive Control (MPC) framework for AV path planning that explicitly minimizes injury consequences rather than simply avoiding collisions.

The primary stakeholders are automotive industry, who gain deployable planning framework and safety assessment linked directly to injury outcomes.