Habitat-GS: A High-Fidelity Navigation Simulator with Dynamic Gaussian Splatting

TL;DR

Habitat-GS is a high-fidelity, real-time navigation simulator that integrates dynamic Gaussian Splatting for photorealistic rendering and human avatars, enhancing agent training and generalization in realistic environments.

Contribution

The paper introduces Habitat-GS, a novel simulator combining 3D Gaussian Splatting with dynamic avatars, supporting photorealistic rendering and human-aware navigation capabilities.

Findings

Agents trained on 3DGS scenes show improved cross-domain generalization.

Gaussian avatars enable effective human-aware navigation.

The system scales well with scene complexity and avatar count.

Abstract

Training embodied AI agents depends critically on the visual fidelity of simulation environments and the ability to model dynamic humans. Current simulators rely on mesh-based rasterization with limited visual realism, and their support for dynamic human avatars, where available, is constrained to mesh representations, hindering agent generalization to human-populated real-world scenarios. We present Habitat-GS, a navigation-centric embodied AI simulator extended from Habitat-Sim that integrates 3D Gaussian Splatting scene rendering and drivable gaussian avatars while maintaining full compatibility with the Habitat ecosystem. Our system implements a 3DGS renderer for real-time photorealistic rendering and supports scalable 3DGS asset import from diverse sources. For dynamic human modeling, we introduce a gaussian avatar module that enables each avatar to simultaneously serve as a photorealistic visual entity and an effective navigation obstacle, allowing agents to learn human-aware behaviors in realistic settings. Experiments on point-goal navigation demonstrate that agents trained on 3DGS scenes achieve stronger cross-domain generalization, with mixed-domain training being the most effective strategy. Evaluations on avatar-aware navigation further confirm that gaussian avatars enable effective human-aware navigation. Finally, performance benchmarks validate the system's scalability across varying scene complexity and avatar counts.