Maps with meaning: 3D scene graphs, maps you can query in language, and maps that keep track of what changed. This is the biggest subfield in the diary, mostly because scene graphs moved fast between 2019 and 2026. The scene graph explainer has the history.
Concepts
- 3D scene graphs: history, the two camps, and what teams would need.
- Open-vocabulary mapping: how language gets onto a map, and what it costs.
- Dynamic scenes: change detection, absence, and forgetting.
- Language-grounded planning, filed under planning.
- Collaborative SLAM, filed under multi-robot systems.
Comparisons
Entities
Datasets, platforms, tools and groups that come up in this subfield: MIT SPARK Lab, CLIP, ScanNet, Replica, 3RScan, Matterport3D, Kimera-Multi. The entity index has the rest.
Open questions
- How do robots agree on a node when one says “chair”, another “grey office chair”, and a third has only a lidar cluster?
- What counts as evidence that something is gone when there is no wall behind it?
- How should a scene graph forget, over months?
Papers
| Paper | Year | Venue | Domain | Threads |
|---|---|---|---|---|
| USS-Nav | 2026 | arXiv preprint | air | perception |
| Hickory | 2026 | arXiv preprint | land | perception, long-term |
| PSG-Nav | 2026 | ICML 2026 | land | perception, long-term |
| LOST-3DSG | 2026 | arXiv preprint | land | long-term, perception |
| DAAAM | 2026 | CVPR 2026 (per the code README; arXiv lists no venue) | land | long-term, perception |
| TrackGraph | 2026 | arXiv 2026 | land, air | perception, long-term |
| Hydra++ | 2026 | IROS 2026 (accepted, per arXiv comments) | land | perception |
| SceneGraphVLM | 2026 | arXiv preprint | general | perception |
| BiMoSG | 2026 | arXiv preprint | land | perception |
| 3DSG survey | 2026 | arXiv preprint | general | perception, long-term |
| vS-Graphs | 2026 | IEEE RA-L, vol. 11, no. 8, August 2026 | land | perception |
| Picasso | 2026 | arXiv preprint | general | perception |
| SuperMap | 2026 | RSS 2026 | land | long-term, perception |
| CAPruner | 2026 | ACL 2026 | general | perception |
| AAPD model | 2025 | Royal Society Open Science 2025 | land | robustness, long-term |
| FunGraph | 2025 | IROS 2025 | land | perception |
| CURB-SG | 2024 | ICRA 2024 | land | perception, long-term |
| LSG | 2024 | arXiv preprint | land | perception |
| xFLIE | 2024 | arXiv preprint | land, air | perception |
| OpenGraph | 2024 | arXiv preprint | land | perception |
| Hydra (IJRR) | 2024 | The International Journal of Robotics Research (IJRR), 2024 | land | perception, long-term |
| Open3DSG | 2024 | CVPR 2024 | land | perception |
| Clio | 2024 | IEEE RA-L 2024 | land | perception, long-term |
| Khronos | 2024 | RSS 2024 | land | long-term, perception |
| HOV-SG | 2024 | RSS 2024 | land | perception |
| Hydra-Multi | 2023 | IROS 2023 | land | perception, robustness |
| LLM-guided scene graph planning | 2023 | arXiv preprint | land | perception, robustness |
| SayPlan | 2023 | CoRL 2023 | land | perception, long-term |
| ConceptGraphs | 2023 | ICRA 2024 | land | perception |
| 3D VSG | 2023 | ICRA 2023 | land | long-term, perception |
| Common-sense spatial knowledge (workshop) | 2023 | RSS 2023 Workshop on Robot Representations for Scene Understanding, Reasoning and Planning | land | perception |
| Language-enabled spatial ontologies | 2023 | IEEE RA-L 2024 | land | perception |
| Mono-Hydra | 2023 | ISPRS Annals (GSW 2023) | land, air | perception |
| Taskography | 2022 | CoRL 2021 | land | perception |
| S-Graphs | 2022 | IEEE RA-L 2022 (with IROS 2022) | land | perception |
| Hydra | 2022 | RSS 2022 | land | perception, long-term |
| DSG navigation policies | 2021 | ICRA 2022 | land | perception, long-term |
| Kimera | 2021 | IJRR 2021 | land | perception, robustness, long-term |
| 3D Dynamic Scene Graphs | 2020 | RSS 2020 | land | perception, long-term |
| 3DSSG | 2020 | CVPR 2020 | land | perception, long-term |
| 3D Scene Graph (Armeni) | 2019 | ICCV 2019 | land | perception |
| 3-D Scene Graph (Kim) | 2019 | IEEE Transactions on Cybernetics | land | perception |
| Paper | Year | Venue | Domain | Threads |
|---|---|---|---|---|
| Benthic long-term localisation | 2026 | Frontiers in Robotics and AI 2026 | sea | long-term, perception |
| T2-Z | 2026 | arXiv preprint | general | long-term |
| HEROIC | 2026 | arXiv preprint | air, land | perception |
| SGM-SLAM | 2026 | arXiv preprint | land | perception |
| PERSEPHONE perception | 2026 | ICARCV 2026 | land | perception |
| CoSMO | 2026 | CoRL 2026 | land | robustness, perception |
| EvoNav-Bench | 2026 | arXiv preprint | land | long-term |
| LTR² | 2026 | IEEE Transactions on Robotics (accepted) | land | long-term, perception |
| MemoGuard | 2026 | IEEE Embedded Systems Letters (ESWEEK CODES 2026) | land | long-term, robustness |
| RIWM | 2026 | arXiv preprint (perspective) | general | long-term, robustness |
| H²-EMV | 2026 | arXiv preprint | land | long-term |
| Kairos | 2026 | arXiv preprint | land | long-term, perception |
| Worth Remembering | 2026 | arXiv preprint | land | long-term |
| SEAM | 2026 | arXiv preprint | land | long-term |
| OASIS-Map | 2026 | arXiv preprint | land | long-term |
| Outdoor open-set 3DSGs | 2026 | IEEE Transactions on Field Robotics (accepted) | land | perception, long-term |
| Mono-Hydra++ | 2026 | arXiv preprint | land | perception |
| FOCUS | 2026 | arXiv preprint | land | long-term |
| ChronoGraph | 2026 | arXiv preprint | land | perception, long-term |
| UQ-DAAAM | 2026 | arXiv preprint | land | long-term, perception |
| 3DSG semantic uncertainty | 2026 | arXiv preprint | land | perception, robustness |
| V-JEPA 2 | 2025 | arXiv preprint | general | long-term |
| ASHiTA | 2025 | CVPR 2025 | land | perception |
| Multi-robot language-grounded planning | 2025 | arXiv preprint | land | perception |
| ROMAN | 2024 | RSS 2025 | land, air | perception, robustness |
| Eiffel Tower dataset | 2023 | IJRR 2023 | sea | long-term, perception |
| Neural Trees | 2021 | NeurIPS 2021 | general | robustness, perception |
Other subfields: perception, mapping and localisation, planning and navigation, control, robot learning, multi-robot systems, safety and verification, security, Human-robot interaction, field robotics.