Neural Topological SLAM for Visual Navigation

TL;DR

This paper introduces a neural topological SLAM approach for visual navigation that uses semantic and geometric representations to improve long-horizon image-goal navigation in unseen environments, achieving significant performance gains.

Contribution

The paper presents a novel topological representation combining semantics and geometry, along with supervised algorithms for building and maintaining these maps in noisy conditions.

Findings

Over 50% improvement over existing methods

Effective representations capture structural regularities

Supports long-horizon navigation in unseen environments

Abstract

This paper studies the problem of image-goal navigation which involves navigating to the location indicated by a goal image in a novel previously unseen environment. To tackle this problem, we design topological representations for space that effectively leverage semantics and afford approximate geometric reasoning. At the heart of our representations are nodes with associated semantic features, that are interconnected using coarse geometric information. We describe supervised learning-based algorithms that can build, maintain and use such representations under noisy actuation. Experimental study in visually and physically realistic simulation suggests that our method builds effective representations that capture structural regularities and efficiently solve long-horizon navigation problems. We observe a relative improvement of more than 50% over existing methods that study this task.