NeuralSim: Augmenting Differentiable Simulators with Neural Networks

@article{Heiden2021NeuralSimAD,
  title={NeuralSim: Augmenting Differentiable Simulators with Neural Networks},
  author={Eric Heiden and David Millard and Erwin Coumans and Yizhou Sheng and Gaurav S. Sukhatme},
  journal={2021 IEEE International Conference on Robotics and Automation (ICRA)},
  year={2021},
  pages={9474-9481}
}
Differentiable simulators provide an avenue for closing the sim-to-real gap by enabling the use of efficient, gradient-based optimization algorithms to find the simulation parameters that best fit the observed sensor readings. Nonetheless, these analytical models can only predict the dynamical behavior of systems for which they have been designed. In this work, we study the augmentation of a novel differentiable rigid-body physics engine via neural networks that is able to learn nonlinear… 
Sparse-Input Neural Network Augmentations for Differentiable Simulators
TLDR
This work studies the augmentation of a differentiable rigid-body physics engine via neural networks that is able to learn nonlinear relationships between dynamic quantities and can thus learn effects not accounted for in the traditional simulator.
Accelerated Policy Learning with Parallel Differentiable Simulation
TLDR
A high-performance differentiable simulator and a new policy learning algorithm (SHAC) that can effectively leverage simulation gradients, even in the presence of non-smoothness are presented.
DiSECt: A Differentiable Simulator for Parameter Inference and Control in Robotic Cutting
TLDR
DiSECt: the first differentiable simulator for cutting soft materials is presented, showing that the simulator can be calibrated to match resultant forces and deformation from a state-of-the-art commercial solver and real-world cutting datasets, with generality across cutting velocities and object instances.
Fundamental Challenges in Deep Learning for Stiff Contact Dynamics
  • Mihir Parmar, Mathew Halm, Michael Posa
  • Biology
    2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)
  • 2021
TLDR
Emp empirical evidence that learning an accurate model in the first place can be confounded by contact is presented, as modern deep learning approaches are not designed to capture this non-smoothness of contact.
A research framework for writing differentiable PDE discretizations in JAX
TLDR
A simple design pattern is proposed to construct a library of differentiable operators and discretizations, by representing operators as mappings between families of continuous functions, parametrized by finite vectors, which is demonstrated on an acoustic optimization problem.
DiSECt: A Differentiable Simulation Engine for Autonomous Robotic Cutting
TLDR
DiSECt: the first differentiable simulator for cutting soft materials is presented, which can be calibrated to match resultant forces and deformation fields from a state-of-the-art commercial solver and real-world cutting datasets, with generality across cutting velocities and object instances.
Probabilistic Inference of Simulation Parameters via Parallel Differentiable Simulation
TLDR
This paper demonstrates how the inference technique can identify symmetries between the parameters and provide highly accurate predictions on an underactuated mechanism where a 7-DOF robot arm excites an object with an unknown mass configuration.
Robot Learning From Randomized Simulations: A Review
TLDR
A comprehensive review of sim-to-real research for robotics, focusing on a technique named “domain randomization” which is a method for learning from randomized simulations.
Real-time Model Predictive Control and System Identification Using Differentiable Physics Simulation
TLDR
A differentiable physics simulation framework that performs online system identification and optimal control simultaneously, using the incoming observations from the target environment in real time, and devise an algorithm to assess the confidence of the estimated parameters using numerical analysis of the dynamic equations.
Differentiable and Learnable Robot Models
TLDR
A library that focuses on the ability to combine data driven methods with analytical rigid body computations and implements both differentiable and learnable models of the kinematics and dynamics of robots in Pytorch.
...
1
2
3
4
...

References

SHOWING 1-10 OF 64 REFERENCES
Interactive Differentiable Simulation
TLDR
This work introduces Interactive Differentiable Simulation (IDS), a differentiable physics engine that allows for efficient, accurate inference of physical properties of rigid-body systems and exhibits orders of magnitude improvements in sample efficiency over model-free reinforcement learning algorithms on challenging nonlinear control domains.
End-to-End Differentiable Physics for Learning and Control
TLDR
This paper demonstrates how to perform backpropagation analytically through a physical simulator defined via a linear complementarity problem, and highlights the system's ability to learn physical parameters from data, efficiently match and simulate observed visual behavior, and readily enable control via gradient-based planning methods.
Sim-to-Real Transfer with Neural-Augmented Robot Simulation
TLDR
This work introduces a method for training a recurrent neural network on the differences between simulated and real robot trajectories and then using this model to augment the simulator, which can be used to learn control policies that transfer significantly better to real environments than policies learned on existing simulators.
Differentiable Simulation for Physical System Identification
TLDR
This letter proposes an extension of the staggered projections algorithm for more accurate solutions of the problem of contacts with friction and introduces a differentiable simulator and an efficient way to compute the analytical derivatives of the involved optimization problems.
Differentiable Physics Models for Real-world Offline Model-based Reinforcement Learning
TLDR
This work demonstrates experimentally that for the offline model-based reinforcement learning setting, physics-based models can be beneficial compared to high-capacity function approximators if the mechanical structure is known and generalizes the approach of physics parameter identification from modeling holonomic multi-body systems to systems with nonholonomic dynamics using end-to-end automatic differentiation.
Augmenting Physical Simulators with Stochastic Neural Networks: Case Study of Planar Pushing and Bouncing
TLDR
This paper builds an efficient, generalizable physical simulator with universal uncertainty estimates for two scenarios, planar pushing and ball bouncing, by augmenting an analytical rigid-body simulator with a neural network that learns to model uncertainty as residuals.
MuJoCo: A physics engine for model-based control
TLDR
A new physics engine tailored to model-based control, based on the modern velocity-stepping approach which avoids the difficulties with spring-dampers, which can compute both forward and inverse dynamics.
Scalable Differentiable Physics for Learning and Control
TLDR
A scalable framework for differentiable physics that can support a large number of objects and their interactions and outperforms derivative-free and model-free baselines by at least an order of magnitude.
Real2Sim Transfer using Differentiable Physics
TLDR
This work introduces a new layer in the deep learning toolbox that imposes a strong inductive bias to generate physically accurate predictions of rigid-body dynamics and allows for the automatic inference of system parameters given an ad-hoc model description.
Deep Lagrangian Networks: Using Physics as Model Prior for Deep Learning
TLDR
The proposed DeLaN network can learn the equations of motion of a mechanical system with a deep network efficiently while ensuring physical plausibility and exhibits substantially improved and more robust extrapolation to novel trajectories and learns online in real-time.
...
1
2
3
4
5
...