Learning Robust Satellite Attitude Dynamics with Physics-Informed Normalising Flow

📄 arXiv: 2508.07841v3 📥 PDF

作者: Carlo Cena, Mauro Martini, Marcello Chiaberge

分类: cs.LG, eess.SY

发布日期: 2025-08-11 (更新: 2025-10-27)


💡 一句话要点

提出基于物理信息的神经网络以提升卫星姿态控制的鲁棒性

🎯 匹配领域: 支柱一:机器人控制 (Robot Control)

关键词: 卫星姿态控制 模型预测控制 物理信息神经网络 深度学习 鲁棒性 数据驱动模型 自注意力机制

📋 核心要点

  1. 现有的纯数据驱动模型在面对训练域外输入时,泛化能力和稳定性较差,难以满足航天器姿态控制的需求。
  2. 本文提出将物理信息神经网络(PINNs)与航天器姿态动态学习相结合,以提高模型的鲁棒性和稳定性。
  3. 实验结果表明,物理信息的引入使得模型的平均相对误差降低了27.08%,并在MPC框架中显著提升了控制精度和收敛速度。

📝 摘要(中文)

姿态控制是航天器操作的基本方面。模型预测控制(MPC)作为一种强大的策略,依赖于准确的系统动态模型来优化控制动作。然而,在物理模型不完整或难以推导的情况下,纯数据驱动模型在泛化和稳定性方面存在挑战。为了解决这些问题,本文探讨了将物理信息神经网络(PINNs)与航天器姿态动态学习相结合的优势。通过使用具有自注意力机制的Real NVP神经网络架构,训练了多种模型,并比较了纯数据驱动方法和物理信息变体的性能。结果显示,物理信息的引入显著提高了模型的性能,尤其是在MPC框架中,PINN模型在控制精度和鲁棒性方面优于纯数据驱动模型,收敛时间改善高达62%。

🔬 方法详解

问题定义:本文旨在解决航天器姿态控制中,现有纯数据驱动模型在泛化和稳定性方面的不足,尤其是在面对训练域外输入时的表现不佳。

核心思路:通过将物理信息神经网络(PINNs)与数据驱动学习相结合,利用物理模型的知识来增强模型的鲁棒性和稳定性,从而提升姿态控制的效果。

技术框架:整体架构包括使用Real NVP神经网络,结合自注意力机制进行训练,分为两种训练策略:纯数据驱动基线和物理信息变体。

关键创新:最重要的创新在于将物理信息与深度学习相结合,显著提高了模型在控制任务中的表现,尤其是在面对噪声和摩擦等不确定性时的鲁棒性。

关键设计:在网络结构上,采用了Real NVP架构,并设计了适应性损失函数以平衡物理信息与数据驱动学习的贡献,确保模型在训练过程中的稳定性和准确性。

📊 实验亮点

实验结果显示,物理信息的引入使得模型的平均相对误差降低了27.08%。在MPC框架中,PINN模型在控制精度和鲁棒性方面优于纯数据驱动模型,收敛时间改善高达62%。

🎯 应用场景

该研究的潜在应用领域包括航天器的自主控制、卫星姿态调整以及其他需要高精度控制的机器人系统。通过提升模型的鲁棒性和稳定性,能够在复杂环境中实现更可靠的操作,具有重要的实际价值和未来影响。

📄 摘要(原文)

Attitude control is a fundamental aspect of spacecraft operations. Model Predictive Control (MPC) has emerged as a powerful strategy for these tasks, relying on accurate models of the system dynamics to optimize control actions over a prediction horizon. In scenarios where physics models are incomplete, difficult to derive, or computationally expensive, machine learning offers a flexible alternative by learning the system behavior directly from data. However, purely data-driven models often struggle with generalization and stability, especially when applied to inputs outside their training domain. To address these limitations, we investigate the benefits of incorporating Physics-Informed Neural Networks (PINNs) into the learning of spacecraft attitude dynamics, comparing their performance with that of purely data-driven approaches. Using a Real-valued Non-Volume Preserving (Real NVP) neural network architecture with a self-attention mechanism, we trained several models on simulated data generated with the Basilisk simulator. Two training strategies were considered: a purely data-driven baseline and a physics-informed variant to improve robustness and stability. Our results demonstrate that the inclusion of physics-based information significantly enhances the performance in terms of the mean relative error with the best architectures found by 27.08%. These advantages are particularly evident when the learned models are integrated into an MPC framework, where PINN-based models consistently outperform their purely data-driven counterparts in terms of control accuracy and robustness, and achieve improved settling times when compared to traditional MPC approaches, yielding improvements of up to 62%, when subject to observation noise and RWs friction.