Safe Learning-Based Control of Autonomous Vehicles

• Collaborators: Mohammad Hassan Asemani,
• August 2026 - Present

Ensuring safety in autonomous driving is challenging in dynamic environments where risk levels continuously change. Conventional Control Barrier Functions typically rely on fixed safety boundaries and do not explicitly adapt to varying environmental risk. In this research, we aim to develop a risk-aware CBF framework that integrates reinforcement learning to adapt safety boundaries based on the estimated level of environmental risk.