HSGNav: Hypothesis Scene Graphs for Multi-Scale Lifelong Navigation under Object Relocation
HSGNav: Hypothesis Scene Graphs for Multi-Scale Lifelong Navigation under Object Relocation
This research presents HSGNav, a navigation system that separates persistent furniture entities from revisable surface-object hypotheses to handle object relocation and multi-scale search in lifelong home environments. The authors introduce FAROS-Bench to evaluate mid-episode memory invalidation and multi-scale goal finding, showing that HSGNav competes with prior methods on GOAT-Bench while substantially outperforming them on the new benchmark.
The paper introduces HSGNav, a navigation framework using Hypothesis Scene Graphs to handle object relocation and multi-scale search in lifelong home environments, validated on the new FAROS-Bench benchmark alongside existing GOAT-Bench results.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
- paper: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
Source：arXiv Robotics — research abstracts · arxiv.org