GraphMapper: Efficient Visual Navigation by Scene Graph Generation

TL;DR

GraphMapper is a method that enables autonomous agents to efficiently learn and utilize 3D scene graph representations for navigation and other tasks, improving interaction efficiency and scene understanding.

Contribution

We introduce GraphMapper, a novel approach that simultaneously learns scene graph representations and navigation policies, enhancing efficiency and versatility in autonomous navigation.

Findings

Fewer interactions needed for effective navigation.

Scene graph representations improve downstream task performance.

GraphMapper can be integrated with existing systems for better efficiency.

Abstract

Understanding the geometric relationships between objects in a scene is a core capability in enabling both humans and autonomous agents to navigate in new environments. A sparse, unified representation of the scene topology will allow agents to act efficiently to move through their environment, communicate the environment state with others, and utilize the representation for diverse downstream tasks. To this end, we propose a method to train an autonomous agent to learn to accumulate a 3D scene graph representation of its environment by simultaneously learning to navigate through said environment. We demonstrate that our approach, GraphMapper, enables the learning of effective navigation policies through fewer interactions with the environment than vision-based systems alone. Further, we show that GraphMapper can act as a modular scene encoder to operate alongside existing Learning-based solutions to not only increase navigational efficiency but also generate intermediate scene representations that are useful for other future tasks.