Gaussian Splatting as a Unified Representation for Autonomy in Unstructured Environments

TL;DR

This paper proposes Gaussian splatting as a unified, efficient representation capturing geometric, photometric, and semantic information for autonomous navigation in large-scale unstructured outdoor environments.

Contribution

It introduces Gaussian splatting as a novel unified representation that supports real-time, large-scale, task-driven robot navigation in complex outdoor settings.

Findings

Gaussian splatting effectively captures complex environment structures.

Semantic embedding enhances task-driven navigation capabilities.

The approach is suitable for real-time autonomous navigation in large-scale environments.

Abstract

In this work, we argue that Gaussian splatting is a suitable unified representation for autonomous robot navigation in large-scale unstructured outdoor environments. Such environments require representations that can capture complex structures while remaining computationally tractable for real-time navigation. We demonstrate that the dense geometric and photometric information provided by a Gaussian splatting representation is useful for navigation in unstructured environments. Additionally, semantic information can be embedded in the Gaussian map to enable large-scale task-driven navigation. From the lessons learned through our experiments, we highlight several challenges and opportunities arising from the use of such a representation for robot autonomy.