TCRL: Temporal-Coupled Adversarial Training for Robust Constrained Reinforcement Learning in Worst-Case Scenarios
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918336899579904 |
|---|---|
| author | Xu, Wentao Yao, Zhongming Li, Weihao Song, Zhenghang Song, Yumeng Li, Tianyi Li, Yushuai |
| author_facet | Xu, Wentao Yao, Zhongming Li, Weihao Song, Zhenghang Song, Yumeng Li, Tianyi Li, Yushuai |
| contents | Constrained Reinforcement Learning (CRL) aims to optimize decision-making policies under constraint conditions, making it highly applicable to safety-critical domains such as autonomous driving, robotics, and power grid management. However, existing robust CRL approaches predominantly focus on single-step perturbations and temporally independent adversarial models, lacking explicit modeling of robustness against temporally coupled perturbations. To tackle these challenges, we propose TCRL, a novel temporal-coupled adversarial training framework for robust constrained reinforcement learning (TCRL) in worst-case scenarios. First, TCRL introduces a worst-case-perceived cost constraint function that estimates safety costs under temporally coupled perturbations without the need to explicitly model adversarial attackers. Second, TCRL establishes a dual-constraint defense mechanism on the reward to counter temporally coupled adversaries while maintaining reward unpredictability. Experimental results demonstrate that TCRL consistently outperforms existing methods in terms of robustness against temporally coupled perturbation attacks across a variety of CRL tasks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_13040 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | TCRL: Temporal-Coupled Adversarial Training for Robust Constrained Reinforcement Learning in Worst-Case Scenarios Xu, Wentao Yao, Zhongming Li, Weihao Song, Zhenghang Song, Yumeng Li, Tianyi Li, Yushuai Machine Learning Constrained Reinforcement Learning (CRL) aims to optimize decision-making policies under constraint conditions, making it highly applicable to safety-critical domains such as autonomous driving, robotics, and power grid management. However, existing robust CRL approaches predominantly focus on single-step perturbations and temporally independent adversarial models, lacking explicit modeling of robustness against temporally coupled perturbations. To tackle these challenges, we propose TCRL, a novel temporal-coupled adversarial training framework for robust constrained reinforcement learning (TCRL) in worst-case scenarios. First, TCRL introduces a worst-case-perceived cost constraint function that estimates safety costs under temporally coupled perturbations without the need to explicitly model adversarial attackers. Second, TCRL establishes a dual-constraint defense mechanism on the reward to counter temporally coupled adversaries while maintaining reward unpredictability. Experimental results demonstrate that TCRL consistently outperforms existing methods in terms of robustness against temporally coupled perturbation attacks across a variety of CRL tasks. |
| title | TCRL: Temporal-Coupled Adversarial Training for Robust Constrained Reinforcement Learning in Worst-Case Scenarios |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2602.13040 |