arXiv：cs.LG· Xun Huang·· 4 小时前AI 评分36
四旋翼飞控中的统计湍流与高保真扰动场：风场保真度对强化学习控制器鲁棒性的影响
Statistical Turbulence and High-Fidelity Disturbance Fields for Quadrotor Flight Control
AI 导读
一项针对四旋翼强化学习控制的研究对比了从无风、离散1-cosine阵风到统计湍流、合成相干结构及大气边界层大涡模拟五种扰动保真度，在0-12 m/s风速范围内对PPO智能体做全交叉训练-测试评估。结果显示训练-测试矩阵出人意料地平坦，最廉价的结构化训练风（离散阵风域随机化）在所有测试列中排名第一，该结论在27 g风敏平台上得到复现；机制诊断表明限制鲁棒性的是控制权限而非风场真实度。
正文
Abstract:Reinforcement-learning quadrotor controllers are usually trained under simplified wind models, yet the impact of wind-field fidelity, as opposed to magnitude, on policy robustness remains unquantified. This paper compares five disturbance-fidelity levels, from wind-free flight and discrete 1-cosine gusts through statistical turbulence and synthetic coherent structures to large-eddy-simulation fields of the atmospheric boundary layer, in a full cross-fidelity train test evaluation of proximal policy optimization (PPO) agents, with cascaded PID and geometric SE(3) controllers as training-free references, over a 0-12 m/s wind sweep. Before any controller comparison is made, all disturbance data are validated: every synthetic generator is checked quantitatively against its analytical or certification-standard reference, and the large-eddy-simulation fields against the imposed log law. On a racing-class quadrotor in hover, the train test matrix is remarkably flat, and the cheapest structured training wind, which is discrete-gust domain randomization, ranks first in every test column, a ranking replicated on a wind-sensitive 27 g platform; a once-tuned geometric controller brackets the learned PPO policies at zero crash rate. Mechanism diagnostics show that control authority, not wind realism, bounds robustness, so wind-fidelity investment should scale with platform wind sensitivity.
| Subjects: | Fluid Dynamics (physics.flu-dyn); Machine Learning (cs.LG); Robotics (cs.RO) |
| Cite as: | arXiv:2610.06874 [physics.flu-dyn] |
| (or arXiv:2610.06874v1 [physics.flu-dyn] for this version) | |
| https://doi.org/10.48550/arXiv.2610.06874 arXiv-issued DOI via DataCite |
Submission history
From: Xun Huang Prof [view email]
[v1]
Wed, 9 Sep 2026 05:14:20 UTC (2,649 KB)
来源：arXiv：cs.LG · arxiv.org