arXiv：cs.LG· Miguel Leiva-V\'elez, Nicolas Gaston-Rozado, Andrea Scorsoglio, Hodei Urrutxua, V\'ictor Rodr\'iguez-Fern\'andez·· 3 小时前
对抗强化学习如何提升行星着陆的鲁棒性
Adversarial Reinforcement Learning for Robust Planetary Landing
AI 导读
研究用对抗强化学习发现行星着陆任务中的挑战性异常条件，由一个次级智能体生成物理约束扰动来削弱着陆控制器性能，构成非对称序贯零和博弈。训练分三阶段：基线控制器学习、随机扰动鲁棒性增强，以及控制器与扰动生成器交替优化的迭代对抗阶段。三自由度着陆环境初步实验显示，对抗扰动揭示了传统随机化遗漏的脆弱性，并初步表明该方法有望提升鲁棒性与泛化能力。
正文
Abstract:Deep reinforcement learning (DRL) policies for planetary landing and other aerospace guidance, navigation, and control tasks are commonly trained under randomized disturbances to improve robustness. However, typical stochastic perturbation sampling assumes independent variations, whereas real mission disturbances often exhibit spatial and temporal correlation and cannot be reproduced through independent random draws. This work investigates adversarial reinforcement learning to discover challenging off-nominal conditions. A secondary agent generates physically constrained perturbations that degrade the performance of a landing controller, forming an asymmetric sequential zero-sum game. Training proceeds in three stages: baseline controller learning, robustness enhancement with stochastic perturbations, and an iterative adversarial phase in which the controller and disturbance generator are alternately optimized. Preliminary experiments in a three-degree-of-freedom landing environment show that adversarial disturbances reveal vulnerabilities missed by conventional randomization and provide early indications that, with a more extensive training campaign, the approach can yield improved robustness and stronger generalization under previously unseen perturbation regimes.
| Subjects: | Machine Learning (cs.LG); Systems and Control (eess.SY) |
| Cite as: | arXiv:2610.20252 [cs.LG] |
| (or arXiv:2610.20252v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.20252 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Miguel Leiva Mr. [view email]
[v1]
Wed, 14 Oct 2026 14:23:47 UTC (1,228 KB)
来源：arXiv：cs.LG · arxiv.org