OrbiSim: World Models as Differentiable Physics Engines for Embodied Intelligence

TL;DR

OrbiSim introduces a fully differentiable physics engine for embodied intelligence, enabling advanced control, physical inference, and improved simulation fidelity in robotics.

Contribution

It presents a unified, physically-grounded world model that supports end-to-end differentiability, surpassing prior models in predictive accuracy and control capabilities.

Findings

Outperforms state-of-the-art world models in predictive fidelity.

Enables gradient-based policy optimization with sparse rewards.

Supports differentiable contact modeling and physical inference.

Abstract

We present OrbiSim, a novel robotic simulation paradigm that redefines world models as a fully differentiable physics engine for embodied intelligence. Unlike prior world models that focus on unconstrained imagination in latent or visual domains, OrbiSim establishes a unified, physically-grounded pathway that bridges structured scene assets, neural dynamics, and downstream reinforcement learning. By enabling end-to-end differentiability throughout the entire simulation loop -- spanning from explicit state transitions to visual observation generation -- OrbiSim supports tasks traditionally intractable for classical simulators, such as differentiable contact modeling, gradient-based policy optimization under sparse rewards, and intuitive physical inference. Empirical results demonstrate that OrbiSim significantly outperforms state-of-the-art world models in both predictive fidelity and control performance. Furthermore, its consistent responsiveness to asset configurations and physical parameters suggests its potential as a differentiable tool for enhancing robot simulation and policy training.