One-Shot Real-to-Sim via End-to-End Differentiable Simulation and Rendering

TL;DR

This paper presents a novel end-to-end differentiable approach for learning comprehensive world models for robots from minimal observations, enabling joint optimization of geometry, appearance, and physical properties for improved simulation and rendering.

Contribution

It introduces a new rigid object representation with differentiable geometry and appearance fields, allowing joint optimization of scene properties from sparse data.

Findings

Successfully learns simulation- and rendering-ready world models from a single robot action sequence.

Demonstrates effectiveness in both simulated and real environments.

Enables end-to-end optimization of geometry, appearance, and physical properties.

Abstract

Identifying predictive world models for robots in novel environments from sparse online observations is essential for robot task planning and execution in novel environments. However, existing methods that leverage differentiable programming to identify world models are incapable of jointly optimizing the geometry, appearance, and physical properties of the scene. In this work, we introduce a novel rigid object representation that allows the joint identification of these properties. Our method employs a novel differentiable point-based geometry representation coupled with a grid-based appearance field, which allows differentiable object collision detection and rendering. Combined with a differentiable physical simulator, we achieve end-to-end optimization of world models, given the sparse visual and tactile observations of a physical motion sequence. Through a series of world model identification tasks in simulated and real environments, we show that our method can learn both simulation- and rendering-ready world models from only one robot action sequence. The code and additional videos are available at our project website: https://tianyi20.github.io/rigid-world-model.github.io/