Agentic Decision-Making and Road Safety

This chapter examines how autonomous vehicles can evolve into agentic, human-aligned decision-making systems capable of operating safely within complex traffic environments. It synthesizes advances in perception, prediction, planning, ethical reasoning, explainable AI, and fail-safe engineering to present a unified architecture for accountable autonomous mobility. Emphasis is placed on designing resilient, transparent, and ethically grounded AV decision pipelines that mitigate uncertainty, anticipate human behavior, and prevent cascading failures. The chapter further identifies gaps in existing AV safety, reliability, and governance frameworks and proposes an integrative model that strengthens road safety outcomes while supporting future research, regulatory development, and industry implementation.