SuReNav: Superpixel Graph-based Constraint Relaxation for Navigation in Over-constrained Environments
Keonyoung Koh, Moonkyeong Jung, Samuel Seungsup Lee, Daehyung Park · 2026
Read SuReNav: Superpixel Graph-based Constraint Relaxation for Navigation in Over-constrained Environments by Keonyoung Koh, Moonkyeong Jung, Samuel Seungsup Lee, Daehyung Park (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-06807.txt.
Description
Keonyoung Koh, Moonkyeong Jung, Samuel Seungsup Lee, and Daehyung Park tackle robot navigation in places where every rule cannot be obeyed at once. In semi-static environments, such as offices that change slowly, a planner must avoid all hard obstacles while straying as little as possible into merely risky areas. Older methods lean on preset costs for each area, which limits how well they generalize.
Their system, SuReNav, divides the map into superpixels, groups of similar cells, and treats each region's constraints as a graph. A graph neural network trained on human demonstrations decides which constraints can safely be relaxed, and the system interleaves relaxation, planning, and movement. It scored highest on human-likeness against strong baselines on two- and three-dimensional maps, and a quadruped robot demonstrated the approach in real city navigation.
Catalog record
| CREATOR | Keonyoung Koh, Moonkyeong Jung, Samuel Seungsup Lee, Daehyung Park |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 40 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-06807