Yggdrasil speeds up 3D scene graph queries for real-time robots
Yggdrasil: a Layer-First 3D Scene Graph for Real-Time Querying
Robotics
Summary
Robots use 3D scene graphs to understand and interact with their surroundings, but existing graphs are slow to work with in real time. The authors created Yggdrasil, a new kind of 3D scene graph that organizes information efficiently so robots can quickly find what they need. Their approach makes queries up to 121 times faster, fitting well within the time budgets of real-time systems. This speed-up was demonstrated across different robotics tasks, making scene graph use much less of a bottleneck.
What this means in practice
- •For robotics developers: Implement Yggdrasil to enable real-time 3D scene querying for improved efficiency in perception and planning pipelines.
- •For embedded systems engineers: Use Yggdrasil to reduce computational overhead and latency of scene graph queries on limited hardware for mobile robots.
Authors
Arshia Akhavan, Ermanno Bartoli, Afnan Algharbi, Alireza Hoseinpur, Iolanda Leite, Bryan Donyanavard
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.