When an autonomous vehicle accelerates, yields, overtakes, or stops, it implicitly makes a claim: “this behavior is legal.” But how can we prove that claim across different legal frameworks and driving contexts? In this talk, we dive into the emerging concept of runtime traffic-law evaluation: representing legislation as a structured ruleset, evaluating it continuously in the vehicle, and generating human-understandable compliance explanations. We discuss the challenges of legal interpretation, conflicting constraints, country-specific variants, and ODD-driven applicability. The result is a blueprint for building autonomous driving systems that are not only performant, but also transparent and legally accountable.

• What “traffic-law compliance” means for autonomous driving, and how it complements functional safety (FuSa) and overall system safety
• Why traffic law is a key scalability challenge for AD across countries, jurisdictions, and ODDs
• How legislation can be translated into structured, machine-readable rulesets (“legislation as code”)
• How runtime traffic-law evaluation works in-vehicle to assess vehicle state and maneuvers in real time
• How to generate explainable compliance outputs (e.g., proof trees, logging) for engineers and regulators