Extending Lagrangian and Hamiltonian Neural Networks with Differentiable Contact Models
arXiv:2102.06794
Abstract
The incorporation of appropriate inductive bias plays a critical role in learning dynamics from data. A growing body of work has been exploring ways to enforce energy conservation in the learned dynamics by encoding Lagrangian or Hamiltonian dynamics into the neural network architecture. These existing approaches are based on differential equations, which do not allow discontinuity in the states and thereby limit the class of systems one can learn. However, in reality, most physical systems, such as legged robots and robotic manipulators, involve contacts and collisions, which introduce discontinuities in the states. In this paper, we introduce a differentiable contact model, which can capture contact mechanics: frictionless/frictional, as well as elastic/inelastic. This model can also accommodate inequality constraints, such as limits on the joint angles. The proposed contact model extends the scope of Lagrangian and Hamiltonian neural networks by allowing simultaneous learning of contact and system properties. We demonstrate this framework on a series of challenging 2D and 3D physical systems with different coefficients of restitution and friction. The learned dynamics can be used as a differentiable physics simulator for downstream gradient-based optimization tasks, such as planning and control.
References in corpus (19)
- PyTorch: An Imperative Style, High-Performance Deep Learning Library
- Interaction Networks for Learning about Objects, Relations and Physics
- Hamiltonian Neural Networks
- A Compositional Object-Based Approach to Learning Physical Dynamics
- Symplectic ODE-Net: Learning Hamiltonian Dynamics with Control
- Differentiable Convex Optimization Layers
- Neural Jump Stochastic Differential Equations
- Dissecting Neural ODEs
- gradSim: Differentiable simulation for system identification and visuomotor control
- Dissipative SymODEN: Encoding Hamiltonian Dynamics with Dissipation and Control into Deep Learning
- Fast and Feature-Complete Differentiable Physics for Articulated Rigid Bodies with Contact
- Simplifying Hamiltonian and Lagrangian Neural Networks via Explicit Constraints
- DensePhysNet: Learning Dense Physical Object Representations via Multi-step Dynamic Interactions
- LagNetViP: A Lagrangian Neural Network for Video Prediction
- Learning Physical Constraints with Neural Projections
- Neural Ordinary Differential Equations for Intervention Modeling
- Unsupervised Learning of Lagrangian Dynamics from Images for Prediction and Control
- Benchmarking Energy-Conserving Neural Networks for Learning Dynamics from Data
- Derivatives of partial eigendecomposition of a real symmetric matrix for degenerate cases