Ensuring safety in autonomous long-haul trucking means moving beyond hardware reliability to address behavioral safety, quantifiable risk estimation, and robust validation frameworks. This session explores how SOTIF and complementary standards like ISO 21448 and UL 4600 can be operationalized in real-world AV programs. With a focus on high-resolution safety verification for evasive maneuvers and risk-driven validation criteria, the session outlines how to build convincing, standards-aligned safety cases that enable safe deployment of L4 autonomous systems on public roads.

• How SOTIF frameworks can be scaled and applied to behavioral safety challenges unique to long-haul autonomous trucking, including evasive maneuver logic and complex highway scenarios
• Developing quantifiable, scenario-based validation metrics for collision risk and system performance, aligned with standards like ISO 21448, UL 4600, and ISO 26262
• Leveraging simulation, real-world testing, and edge-case modeling to define and verify AV safety goals within a robust, standards-aligned safety case

• Where do current standards fall short in addressing the complexity of AI-based safety and validation frameworks?
• How can SOTIF principles be effectively integrated with machine learning systems to manage edge cases and unknowns?
• What best practices are emerging around layered safety architectures and intelligent redundancy?
• In what ways do cybersecurity and functional safety intersect in AD system design — and where are the critical gaps?
• How can global harmonization of AI safety standards accelerate safer deployment of automated driving technologies?