Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2409.18343 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914958444331008 |
|---|---|
| author | Peng, Zhenghao Luo, Wenjie Lu, Yiren Shen, Tianyi Gulino, Cole Seff, Ari Fu, Justin |
| author_facet | Peng, Zhenghao Luo, Wenjie Lu, Yiren Shen, Tianyi Gulino, Cole Seff, Ari Fu, Justin |
| contents | A major challenge in autonomous vehicle research is modeling agent behaviors, which has critical applications including constructing realistic and reliable simulations for off-board evaluation and forecasting traffic agents motion for onboard planning. While supervised learning has shown success in modeling agents across various domains, these models can suffer from distribution shift when deployed at test-time. In this work, we improve the reliability of agent behaviors by closed-loop fine-tuning of behavior models with reinforcement learning. Our method demonstrates improved overall performance, as well as improved targeted metrics such as collision rate, on the Waymo Open Sim Agents challenge. Additionally, we present a novel policy evaluation benchmark to directly assess the ability of simulated agents to measure the quality of autonomous vehicle planners and demonstrate the effectiveness of our approach on this new benchmark. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2409_18343 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Improving Agent Behaviors with RL Fine-tuning for Autonomous Driving Peng, Zhenghao Luo, Wenjie Lu, Yiren Shen, Tianyi Gulino, Cole Seff, Ari Fu, Justin Artificial Intelligence I.2.6; I.2.9 A major challenge in autonomous vehicle research is modeling agent behaviors, which has critical applications including constructing realistic and reliable simulations for off-board evaluation and forecasting traffic agents motion for onboard planning. While supervised learning has shown success in modeling agents across various domains, these models can suffer from distribution shift when deployed at test-time. In this work, we improve the reliability of agent behaviors by closed-loop fine-tuning of behavior models with reinforcement learning. Our method demonstrates improved overall performance, as well as improved targeted metrics such as collision rate, on the Waymo Open Sim Agents challenge. Additionally, we present a novel policy evaluation benchmark to directly assess the ability of simulated agents to measure the quality of autonomous vehicle planners and demonstrate the effectiveness of our approach on this new benchmark. |
| title | Improving Agent Behaviors with RL Fine-tuning for Autonomous Driving |
| topic | Artificial Intelligence I.2.6; I.2.9 |
| url | https://arxiv.org/abs/2409.18343 |