Safety-Critical Scenario Generation Via Reinforcement Learning Based Editing

arXiv:2306.14131v3 Announce Type: replace
Abstract: Generating safety-critical scenarios is essential for testing and verifying the safety of autonomous vehicles. Traditional optimization techniques suffer from the curse of dimensionality and limit the search space to fixed parameter spaces. To address these challenges, we propose a deep reinforcement learning approach that generates scenarios by sequential editing, such as adding new agents or modifying the trajectories of the existing agents. Our framework employs a reward function consisting of both risk and plausibility objectives. …

abstract arxiv autonomous autonomous vehicles challenges cs.lg cs.ro dimensionality editing optimization reinforcement reinforcement learning safety safety-critical search space spaces testing the curse of dimensionality type vehicles via