Go fetch!-dynamic grasps using boston dynamics spot with external robotic arm

S Zimmermann, R Poranne… - 2021 IEEE International …, 2021‏ - ieeexplore.ieee.org
We combine Boston Dynamics Spot® with a light-weight, external robot arm to perform
dynamic gras** maneuvers. While Spot is a reliable, robust and easy-to-control mobile …

Neural networks with physics-informed architectures and constraints for dynamical systems modeling

F Djeumou, C Neary, E Goubault… - … for Dynamics and …, 2022‏ - proceedings.mlr.press
Effective inclusion of physics-based knowledge into deep neural network models of
dynamical systems can greatly improve data efficiency and generalization. Such a priori …

Combining physics and deep learning to learn continuous-time dynamics models

M Lutter, J Peters - The International Journal of Robotics …, 2023‏ - journals.sagepub.com
Deep learning has been widely used within learning algorithms for robotics. One
disadvantage of deep networks is that these networks are black-box representations …

Lagrangian operator inference enhanced with structure-preserving machine learning for nonintrusive model reduction of mechanical systems

H Sharma, DA Najera-Flores, MD Todd… - Computer Methods in …, 2024‏ - Elsevier
Complex mechanical systems often exhibit strongly nonlinear behavior due to the presence
of nonlinearities in the energy dissipation mechanisms, material constitutive relationships, or …

Investigating compounding prediction errors in learned dynamics models

N Lambert, K Pister, R Calandra - arxiv preprint arxiv:2203.09637, 2022‏ - arxiv.org
Accurately predicting the consequences of agents' actions is a key prerequisite for planning
in robotic control. Model-based reinforcement learning (MBRL) is one paradigm which relies …

Deconstructing the inductive biases of hamiltonian neural networks

N Gruver, M Finzi, S Stanton, AG Wilson - arxiv preprint arxiv:2202.04836, 2022‏ - arxiv.org
Physics-inspired neural networks (NNs), such as Hamiltonian or Lagrangian NNs,
dramatically outperform other learned dynamics models by leveraging strong inductive …

Compositional learning of dynamical system models using port-Hamiltonian neural networks

C Neary, U Topcu - Learning for Dynamics and Control …, 2023‏ - proceedings.mlr.press
Many dynamical systems—from robots interacting with their surroundings to large-scale
multi-physics systems—involve a number of interacting subsystems. Toward the objective of …

Encoding physical constraints in differentiable newton-euler algorithm

G Sutanto, A Wang, Y Lin, M Mukadam… - … for Dynamics and …, 2020‏ - proceedings.mlr.press
Abstract The recursive Newton-Euler Algorithm (RNEA) is a popular technique in robotics for
computing the dynamics of robots. The computed dynamics can then be used for torque …

Preserving Lagrangian structure in data-driven reduced-order modeling of large-scale dynamical systems

H Sharma, B Kramer - Physica D: Nonlinear Phenomena, 2024‏ - Elsevier
This work presents a nonintrusive physics-preserving method to learn reduced-order models
(ROMs) of Lagrangian systems, which includes nonlinear wave equations. Existing intrusive …

Learning modular simulations for homogeneous systems

J Gupta, S Vemprala, A Kapoor - Advances in Neural …, 2022‏ - proceedings.neurips.cc
Complex systems are often decomposed into modular subsystems for engineering
tractability. Although various equation based white-box modeling techniques make use of …