用物理约束神经网络精准建模水下车辆动态，提升预测长期准确性
Modelling of Underwater Vehicles using Physics-Informed Neural Networks with Control
- 引入控制输入的物理信息神经网络，融合动力学方程与数据驱动
- 在模拟水下车辆上实现比传统方法更优的长时序预测性能
- 开源框架支持多种配置测试，适合机器人与控制系统研究者
物理信息神经网络（PINNs）通过将物理规律融入数据驱动模型，提升泛化能力与样本效率。本文提出开源的带控制的物理信息神经网络（PINC）框架，用于建模水下车辆的动力学行为。PINC以初始状态、控制动作和时间作为输入，扩展了传统PINNs的能力，实现训练域外的物理一致动态演进。通过测试不同损失函数、梯度加权策略及超参数配置，验证结果表明，在模拟水下车辆上的长时序预测精度显著优于非物理信息基线模型。
原文摘要 · Abstract (English)
Physics-informed neural networks (PINNs) integrate physical laws with data-driven models to improve generalization and sample efficiency. This work introduces an open-source implementation of the Physics-Informed Neural Network with Control (PINC) framework, designed to model the dynamics of an underwater vehicle. Using initial states, control actions, and time inputs, PINC extends PINNs to enable physically consistent transitions beyond the training domain. Various PINC configurations are tested, including differing loss functions, gradient-weighting schemes, and hyperparameters. Validation on a simulated underwater vehicle demonstrates more accurate long-horizon predictions compared to a non-physics-informed baseline
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。