Research papers · Posted 2023.09.01 · Preprint
Suicidal Pedestrian: Generation of Safety-Critical Scenarios for Autonomous Vehicles
Suicidal Pedestrian: Generation of Safety-Critical Scenarios for Autonomous Vehicles
Overview
The study creates a reinforcement-learning pedestrian agent in CARLA that searches for vehicle collisions. Custom rewards and collision-oriented metrics expose pedestrian-related decision failures in two imitation-learning driving algorithms.
Author-supplied source information
Author-supplied information on arXiv.
6 pages; 5 figures; 2 tables
Version: v1