SafePath: Automated Extraction of Temporal Logic Based Trajectory Embeddings for Enhanced Autonomous Vehicle Safety
PI Madhur Behl, Ph.D.
This research presents SafePath, a new method for evaluating and comparing autonomous vehicle safety by analyzing trajectory data.
SafePath aims to verify the correctness of AVs’ navigation algorithms and compare how different AVs respond in similar traffic situations, ultimately providing valuable insights for regulators and the industry, improving AV development, and enhancing safety in complex scenarios. The study will use real-world data from Waymo to evaluate SafePath’s effectiveness, with the goal of contributing to the overall safety of AVs and similar autonomous systems.
