REVIEW 6 cited by
Dojo: A Differentiable Physics Engine for Robotics
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
We present Dojo, a differentiable physics engine for robotics that prioritizes stable simulation, accurate contact physics, and differentiability with respect to states, actions, and system parameters. Dojo models hard contact and friction with a nonlinear complementarity problem with second-order cone constraints. We introduce a custom primal-dual interior-point method to solve the second order cone program for stable forward simulation over a broad range of sample rates. We obtain smooth gradient approximations with this solver through the implicit function theorem, giving gradients that are useful for downstream trajectory optimization, policy optimization, and system identification applications. Specifically, we propose to use the central path parameter threshold in the interior point solver as a user-tunable design parameter. A high value gives a smooth approximation to contact dynamics with smooth gradients for optimization and learning, while a low value gives precise simulation rollouts with hard contact. We demonstrate Dojo's differentiability in trajectory optimization, policy learning, and system identification examples. We also benchmark Dojo against MuJoCo, PyBullet, Drake, and Brax on a variety of robot models, and study the stability and simulation quality over a range of sample frequencies and accuracy tolerances. Finally, we evaluate the sim-to-real gap in hardware experiments with a Ufactory xArm 6 robot. Dojo is an open source project implemented in Julia with Python bindings, with code available at https://github.com/dojo-sim/Dojo.jl.
Forward citations
Cited by 6 Pith papers
-
React to Surprises: Stable-by-Design Neural Feedback Control and the Youla-REN
A Youla-REN policy class guarantees d-tube contraction and Lipschitzness for partially-observed nonlinear systems with disturbances, and covers all contracting and Lipschitz closed loops under certainty-equivalence observers.
-
NeuralActuator: Neural Actuation Modeling for Robot Dynamics and External Force Perception
A Transformer-based actuator model learns torque surrogates and external forces on low-cost servo arms from pose trajectories and motor telemetry, improving force estimation and behavior-cloning control.
-
Few-Shot Neural Differentiable Simulator: Real-to-Sim Rigid-Contact Modeling
Few-shot contact-parameter identification plus MuJoCo data scaling trains a fully differentiable mesh GNN that matches real rigid-contact trajectories better than Brax.
-
DiffCoTune: Differentiable Co-Tuning for Cross-domain Robot Control
DiffCoTune co-tunes simulator and controller parameters via gradient-based alternating optimization, improving sim-to-real transfer with fewer than five deployment rollouts.
-
Shooting for Contact: Contact-Implicit Multiple Shooting for Dynamic Motion Retargeting
A differentiable MuJoCo simulator is embedded in a multiple-shooting optimizer to generate dynamically feasible, contact-consistent reference motions that accelerate motion-imitation RL training and transfer zero-shot...
-
From 2D to 3D Cognition: A Brief Survey of General World Models
A survey proposing a two-pillar, three-capability framework that organizes recent AI world models by their transition from 2D visual prediction to 3D cognition.
Discussion (0). Continue with ORCID to comment.