Real2Sim: A Physics-driven and Editable Gaussian Splatting Framework for Autonomous Driving Scenes

TL;DR

Real2Sim is a physics-aware, editable scene synthesis framework for autonomous driving that combines Gaussian Splatting with a differentiable physics solver, enabling realistic, high-fidelity, and customizable scenario generation.

Contribution

It introduces a unified approach integrating 4D Gaussian Splatting with a differentiable physics engine for dynamic, editable, and realistic driving scene synthesis.

Findings

Validated on Waymo dataset showing high-quality rendering and reconstruction.

Demonstrated scene editing and physics simulation capabilities.

Enabled generation of challenging scenarios like collisions and post-impact trajectories.

Abstract

Reliable autonomous driving relies on large-scale, well-labeled data and robust models. However, manual data collection is resource-intensive, and traditional simulation suffers from a persistent reality gap. While recent generative frameworks and radiance-field methods improve visual fidelity, they still struggle with temporal and spatial consistency and cannot ensure physics-aware behavior, limiting their applicability to driving scenario generation. To address these challenges, we propose Real2Sim, an unified framework that combines 4D Gaussian Splatting (4DGS) with a differentiable Material Point Method (MPM) solver. Real2Sim explicitly reconstructs dynamic driving scenes as temporally continuous Gaussian primitives, supports instance-level editing, and simulates realistic object-object and object-environment interactions. This framework enables physics-aware, high-fidelity synthesis of diverse, editable scenarios, including challenging corner cases such as collisions and post-impact trajectories. Experiments on the Waymo Open Dataset validate Real2Sim's capabilities in rendering, reconstruction, editing, and physics simulation, demonstrating its potential as a scalable tool for data generation in downstream tasks such as perception, tracking, trajectory prediction, and end-to-end policy learning.