A High-Fidelity Digital Twin for Robotic Manipulation Based on 3D Gaussian Splatting

TL;DR

This paper introduces a fast, high-fidelity digital twin framework for robotic manipulation using 3D Gaussian Splatting, enabling real-time, photorealistic scene reconstruction and collision model generation from sparse RGB data.

Contribution

The authors develop a practical system that constructs detailed digital twins within minutes, integrating semantic labelling and collision-ready geometry for robotic tasks.

Findings

High-quality digital twins reconstructed in minutes from sparse RGB inputs.

Enhanced geometric accuracy improves robot manipulation robustness.

System successfully demonstrated on a Franka Emika Panda robot for pick-and-place tasks.

Abstract

Developing high-fidelity, interactive digital twins is crucial for enabling closed-loop motion planning and reliable real-world robot execution, which are essential to advancing sim-to-real transfer. However, existing approaches often suffer from slow reconstruction, limited visual fidelity, and difficulties in converting photorealistic models into planning-ready collision geometry. We present a practical framework that constructs high-quality digital twins within minutes from sparse RGB inputs. Our system employs 3D Gaussian Splatting (3DGS) for fast, photorealistic reconstruction as a unified scene representation. We enhance 3DGS with visibility-aware semantic fusion for accurate 3D labelling and introduce an efficient, filter-based geometry conversion method to produce collision-ready models seamlessly integrated with a Unity-ROS2-MoveIt physics engine. In experiments with a Franka Emika Panda robot performing pick-and-place tasks, we demonstrate that this enhanced geometric accuracy effectively supports robust manipulation in real-world trials. These results demonstrate that 3DGS-based digital twins, enriched with semantic and geometric consistency, offer a fast, reliable, and scalable path from perception to manipulation in unstructured environments.