Safety-Critical Scenario Generation Via Reinforcement Learning Based Editing
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2023
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913256480702464 |
|---|---|
| author | Liu, Haolan Zhang, Liangjun Hari, Siva Kumar Sastry Zhao, Jishen |
| author_facet | Liu, Haolan Zhang, Liangjun Hari, Siva Kumar Sastry Zhao, Jishen |
| contents | 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. The plausibility objective leverages generative models, such as a variational autoencoder, to learn the likelihood of the generated parameters from the training datasets; It penalizes the generation of unlikely scenarios. Our approach overcomes the dimensionality challenge and explores a wide range of safety-critical scenarios. Our evaluation demonstrates that the proposed method generates safety-critical scenarios of higher quality compared with previous approaches. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2306_14131 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | Safety-Critical Scenario Generation Via Reinforcement Learning Based Editing Liu, Haolan Zhang, Liangjun Hari, Siva Kumar Sastry Zhao, Jishen Machine Learning Robotics 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. The plausibility objective leverages generative models, such as a variational autoencoder, to learn the likelihood of the generated parameters from the training datasets; It penalizes the generation of unlikely scenarios. Our approach overcomes the dimensionality challenge and explores a wide range of safety-critical scenarios. Our evaluation demonstrates that the proposed method generates safety-critical scenarios of higher quality compared with previous approaches. |
| title | Safety-Critical Scenario Generation Via Reinforcement Learning Based Editing |
| topic | Machine Learning Robotics |
| url | https://arxiv.org/abs/2306.14131 |