Generation of Uncertainty-Aware High-Level Spatial Concepts in Factorized 3D Scene Graphs via Graph Neural Networks

TL;DR

This paper introduces a learning-based method for automatically inferring and integrating high-level spatial concepts into 3D scene graphs for SLAM, enhancing indoor navigation and mapping without manual heuristics.

Contribution

It proposes an online learning approach to infer spatial concepts as factors in SLAM, reducing manual design and improving generalization across environments.

Findings

Room detection improved by 20.7% in simulation

Trajectory estimation improved by 19.2% in simulation

Room detection improved by 5.3% on real sites

Abstract

Enabling robots to autonomously discover high-level spatial concepts (e.g., rooms and walls) from primitive geometric observations (e.g., planar surfaces) within 3D Scene Graphs is essential for robust indoor navigation and mapping. These graphs provide a hierarchical metric-semantic representation in which such concepts are organized. To further enhance graph-SLAM performance, Factorized 3D Scene Graphs incorporate these concepts as optimization factors that constrain relative geometry and enforce global consistency. However, both stages of this process remain largely manual: concepts are typically derived using hand-crafted, concept-specific heuristics, while factors and their covariances are likewise manually designed. This reliance on manual specification limits generalization across diverse environments and scalability to new concept classes. This paper presents a novel learning-based method that infers spatial concepts online from observed vertical planes and introduces them as optimizable factors within a SLAM backend, eliminating the need to handcraft concept generation, factor design, and covariance specification. We evaluate our approach in simulated environments with complex layouts, improving room detection by 20.7% and trajectory estimation by 19.2%, and further validate it on real construction sites, where room detection improves by 5.3% and map matching accuracy by 3.8%. Results confirm that learned factors can improve their handcrafted counterparts in SLAM systems and serve as a foundation for extending this approach to new spatial concepts.