Yggdrasil: a Layer-First 3D Scene Graph for Real-Time Querying

2026-09-29

AI Summary

This study addresses the high query latency of existing 3D scene graphs, which impedes real-time robotic perception. We propose a hierarchy-first, general-purpose 3D scene graph framework that introduces a novel hierarchical structure optimized for efficient generation. By constructing a universal graph model comprising nodes, edges, and layers with refined underlying data structures, the approach balances generation efficiency with consumption speed while natively supporting positional-semantic queries across indoor, outdoor, flat, and hierarchical representations. Experimental results demonstrate that the framework achieves microsecond-level response times, accelerating queries by 121× with latencies of only 2–127 microseconds. Furthermore, it eliminates up to 99% of scene graph processing overhead across three mainstream pipelines, substantially expediting retrieval in downstream tasks.

Abstract

Robotic agents use 3D scene graphs (3DSG) to perform tasks ranging from scene understanding to scene interaction. Although an extensive body of work addresses scene graph generation, little attention has been paid to optimizing the graph for consumption, which leaves state-of-the-art perception pipelines to work around their own scene graph and to pay a latency cost that does not fit the real-time budget a perception loop runs on. We present Yggdrasil, the first 3D scene graph designed to be efficient for both generation and consumption: a layer-first hierarchical graph built from generic nodes, edges, and layers, which expresses the representations existing pipelines already produce, indoor or outdoor, flat or hierarchical, while natively answering the positional and semantic queries downstream tasks issue. Against a published DSG baseline, Yggdrasil answers queries up to $121\times$ faster, and every query we measure falls between 2 and 127 microseconds, three to five orders of magnitude inside the 200 microsecond keyframe budget a 3DSG consumer lives in, on both a workstation and embedded class device. We integrate Yggdrasil into three published pipelines spanning human trajectory prediction, object-goal navigation, and human-aware motion planning, where it removes up to 99% of the time each spends on its scene graph. The implementation, benchmark harness, and all three integrations are available online.

Problem

• 3D scene graph
• real-time querying
• latency optimization
• robotic perception
• graph consumption

Innovation

• 3D Scene Graph
• Layer-First Hierarchy
• Real-Time Querying
• Scene Representation
• Robotic Perception