Effective Task Planning with Missing Objects using Learning-Informed Object Search
Raihan Islam Arnob * , Max Merlin † , Abhishek Paudel * , Benned Hedegaard † , George Konidaris † , and Gregory J. Stein * · 2026
Read Effective Task Planning with Missing Objects using Learning-Informed Object Search by Raihan Islam Arnob * , Max Merlin † , Abhishek Paudel * , Benned Hedegaard † , George Konidaris † , and Gregory J. Stein * (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-2602-11468.txt.
Description
Robot task planners usually assume every needed object is already known. Raihan Islam Arnob and five co-authors build a planning frame for missing objects using learning-informed object search actions. Each such action is a policy to find and retrieve one object, informed by model-based expected costs.
High-level planning treats those actions as fixed, interleaving search and execution into plans that are effective, sound, and complete despite doubt. The design stays compatible with current full-knowledge solvers. Tests in simulated homes and in the real world beat non-learned and learned rivals on recall and meal preparation tasks. The work lets planners ask what is needed and when to look for it.
Catalog record
| CREATOR | Raihan Islam Arnob * , Max Merlin † , Abhishek Paudel * , Benned Hedegaard † , George Konidaris † , and Gregory J. Stein * |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 35 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-11468