Combining Physics and Deep Learning to learn Continuous-Time Dynamics Models

• URL: http://arxiv.org/abs/2110.01894v1
• Date: Tue, 5 Oct 2021 09:30:56 GMT
• Title: Combining Physics and Deep Learning to learn Continuous-Time Dynamics Models
• Authors: Michael Lutter and Jan Peters
• Abstract summary: We introduce physics-inspired deep networks that combine first principles from physics with deep learning. Deep Lagrangian Networks (DeLaN) parametrize the system energy using two networks. Results show that the proposed approach obtains dynamics models that can be applied to physical systems for real-time control.
• Score: 31.842201180914756
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Deep learning has been widely used within learning algorithms for robotics. One disadvantage of deep networks is that these networks are black-box representations. Therefore, the learned approximations ignore the existing knowledge of physics or robotics. Especially for learning dynamics models, these black-box models are not desirable as the underlying principles are well understood and the standard deep networks can learn dynamics that violate these principles. To learn dynamics models with deep networks that guarantee physically plausible dynamics, we introduce physics-inspired deep networks that combine first principles from physics with deep learning. We incorporate Lagrangian mechanics within the model learning such that all approximated models adhere to the laws of physics and conserve energy. Deep Lagrangian Networks (DeLaN) parametrize the system energy using two networks. The parameters are obtained by minimizing the squared residual of the Euler-Lagrange differential equation. Therefore, the resulting model does not require specific knowledge of the individual system, is interpretable, and can be used as a forward, inverse, and energy model. Previously these properties were only obtained when using system identification techniques that require knowledge of the kinematic structure. We apply DeLaN to learning dynamics models and apply these models to control simulated and physical rigid body systems. The results show that the proposed approach obtains dynamics models that can be applied to physical systems for real-time control. Compared to standard deep networks, the physics-inspired models learn better models and capture the underlying structure of the dynamics.

Related papers

• Deepmechanics [0.0]
  Physics-informed deep learning models have emerged as powerful tools for learning dynamical systems.<n>We benchmark three prominent physics-informed architectures using the DeepChem framework.<n>We find that all benchmarked models struggle to maintain stability for chaotic or nonconservative systems.
  arXiv  Detail & Related papers  (2026-02-20T08:27:43Z)
• Physics Encoded Blocks in Residual Neural Network Architectures for Digital Twin Models [2.8720819157502344]
  This paper presents a generic approach based on a novel physics-encoded residual neural network architecture. Our method combines physics blocks as mathematical operators from physics-based models with learning blocks comprising feed-forward layers. Compared to conventional neural network-based methods, our method improves generalizability with substantially low data requirements.
  arXiv  Detail & Related papers  (2024-11-18T11:58:20Z)
• Physically Consistent Neural ODEs for Learning Multi-Physics Systems [0.0]
  In this paper, we leverage the framework of Irreversible port-Hamiltonian Systems (IPHS), which can describe most multi-physics systems. We propose Physically Consistent NODEs (PC-NODEs) to learn parameters from data. We demonstrate the effectiveness of the proposed method by learning the thermodynamics of a building from the real-world measurements.
  arXiv  Detail & Related papers  (2022-11-11T11:20:35Z)
• Human Trajectory Prediction via Neural Social Physics [63.62824628085961]
  Trajectory prediction has been widely pursued in many fields, and many model-based and model-free methods have been explored. We propose a new method combining both methodologies based on a new Neural Differential Equation model. Our new model (Neural Social Physics or NSP) is a deep neural network within which we use an explicit physics model with learnable parameters.
  arXiv  Detail & Related papers  (2022-07-21T12:11:18Z)
• Physics Embedded Neural Network Vehicle Model and Applications in Risk-Aware Autonomous Driving Using Latent Features [6.33280703577189]
  Non-holonomic vehicle motion has been studied extensively using physics-based models. In this paper, we seamlessly combine deep learning with a fully differentiable physics model to endow the neural network with available prior knowledge.
  arXiv  Detail & Related papers  (2022-07-16T12:06:55Z)
• ModLaNets: Learning Generalisable Dynamics via Modularity and Physical Inductive Bias [14.474273671369584]
  We propose a structural neural network framework with modularity and physical inductive bias. This framework models the energy of each element using modularity and then construct the target dynamical system via Lagrangian mechanics. We examine our framework for modelling double-pendulum or three-body systems with small training datasets.
  arXiv  Detail & Related papers  (2022-06-24T14:54:25Z)
• Physics-Inspired Temporal Learning of Quadrotor Dynamics for Accurate Model Predictive Trajectory Tracking [76.27433308688592]
  Accurately modeling quadrotor's system dynamics is critical for guaranteeing agile, safe, and stable navigation. We present a novel Physics-Inspired Temporal Convolutional Network (PI-TCN) approach to learning quadrotor's system dynamics purely from robot experience. Our approach combines the expressive power of sparse temporal convolutions and dense feed-forward connections to make accurate system predictions.
  arXiv  Detail & Related papers  (2022-06-07T13:51:35Z)
• Which priors matter? Benchmarking models for learning latent dynamics [70.88999063639146]
  Several methods have proposed to integrate priors from classical mechanics into machine learning models. We take a sober look at the current capabilities of these models. We find that the use of continuous and time-reversible dynamics benefits models of all classes.
  arXiv  Detail & Related papers  (2021-11-09T23:48:21Z)
• Constructing Neural Network-Based Models for Simulating Dynamical Systems [59.0861954179401]
  Data-driven modeling is an alternative paradigm that seeks to learn an approximation of the dynamics of a system using observations of the true system. This paper provides a survey of the different ways to construct models of dynamical systems using neural networks. In addition to the basic overview, we review the related literature and outline the most significant challenges from numerical simulations that this modeling paradigm must overcome.
  arXiv  Detail & Related papers  (2021-11-02T10:51:42Z)
• Neural Networks with Physics-Informed Architectures and Constraints for Dynamical Systems Modeling [19.399031618628864]
  We develop a framework to learn dynamics models from trajectory data. We place constraints on the values of the outputs and the internal states of the model. We experimentally demonstrate the benefits of the proposed approach on a variety of dynamical systems.
  arXiv  Detail & Related papers  (2021-09-14T02:47:51Z)
• ST-PCNN: Spatio-Temporal Physics-Coupled Neural Networks for Dynamics Forecasting [15.265694039283106]
  We propose a physics-coupled neural network model to learn parameters governing the physics of the system. A-temporal physics-coupled neural network (ST-PCNN) model is proposed to achieve three goals. Experiments, using simulated and field-collected ocean data, validate that ST-PCNN outperforms existing physics-informed models.
  arXiv  Detail & Related papers  (2021-08-12T19:34:00Z)
• Physics-Integrated Variational Autoencoders for Robust and Interpretable Generative Modeling [86.9726984929758]
  We focus on the integration of incomplete physics models into deep generative models. We propose a VAE architecture in which a part of the latent space is grounded by physics. We demonstrate generative performance improvements over a set of synthetic and real-world datasets.
  arXiv  Detail & Related papers  (2021-02-25T20:28:52Z)
• Learning Stable Deep Dynamics Models [91.90131512825504]
  We propose an approach for learning dynamical systems that are guaranteed to be stable over the entire state space. We show that such learning systems are able to model simple dynamical systems and can be combined with additional deep generative models to learn complex dynamics.
  arXiv  Detail & Related papers  (2020-01-17T00:04:45Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.