Salvato in:
Dettagli Bibliografici
Autori principali: Marques, Joao Marcos Correia, Dengler, Nils, Zaenker, Tobias, Mucke, Jesper, Wang, Shenlong, Bennewitz, Maren, Hauser, Kris
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:https://arxiv.org/abs/2502.20606
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866915349750874112
author Marques, Joao Marcos Correia
Dengler, Nils
Zaenker, Tobias
Mucke, Jesper
Wang, Shenlong
Bennewitz, Maren
Hauser, Kris
author_facet Marques, Joao Marcos Correia
Dengler, Nils
Zaenker, Tobias
Mucke, Jesper
Wang, Shenlong
Bennewitz, Maren
Hauser, Kris
contents Searching for objects in cluttered environments requires selecting efficient viewpoints and manipulation actions to remove occlusions and reduce uncertainty in object locations, shapes, and categories. In this work, we address the problem of manipulation-enhanced semantic mapping, where a robot has to efficiently identify all objects in a cluttered shelf. Although Partially Observable Markov Decision Processes~(POMDPs) are standard for decision-making under uncertainty, representing unstructured interactive worlds remains challenging in this formalism. To tackle this, we define a POMDP whose belief is summarized by a metric-semantic grid map and propose a novel framework that uses neural networks to perform map-space belief updates to reason efficiently and simultaneously about object geometries, locations, categories, occlusions, and manipulation physics. Further, to enable accurate information gain analysis, the learned belief updates should maintain calibrated estimates of uncertainty. Therefore, we propose Calibrated Neural-Accelerated Belief Updates (CNABUs) to learn a belief propagation model that generalizes to novel scenarios and provides confidence-calibrated predictions for unknown areas. Our experiments show that our novel POMDP planner improves map completeness and accuracy over existing methods in challenging simulations and successfully transfers to real-world cluttered shelves in zero-shot fashion.
format Preprint
id arxiv_https___arxiv_org_abs_2502_20606
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Map Space Belief Prediction for Manipulation-Enhanced Mapping
Marques, Joao Marcos Correia
Dengler, Nils
Zaenker, Tobias
Mucke, Jesper
Wang, Shenlong
Bennewitz, Maren
Hauser, Kris
Robotics
Machine Learning
Searching for objects in cluttered environments requires selecting efficient viewpoints and manipulation actions to remove occlusions and reduce uncertainty in object locations, shapes, and categories. In this work, we address the problem of manipulation-enhanced semantic mapping, where a robot has to efficiently identify all objects in a cluttered shelf. Although Partially Observable Markov Decision Processes~(POMDPs) are standard for decision-making under uncertainty, representing unstructured interactive worlds remains challenging in this formalism. To tackle this, we define a POMDP whose belief is summarized by a metric-semantic grid map and propose a novel framework that uses neural networks to perform map-space belief updates to reason efficiently and simultaneously about object geometries, locations, categories, occlusions, and manipulation physics. Further, to enable accurate information gain analysis, the learned belief updates should maintain calibrated estimates of uncertainty. Therefore, we propose Calibrated Neural-Accelerated Belief Updates (CNABUs) to learn a belief propagation model that generalizes to novel scenarios and provides confidence-calibrated predictions for unknown areas. Our experiments show that our novel POMDP planner improves map completeness and accuracy over existing methods in challenging simulations and successfully transfers to real-world cluttered shelves in zero-shot fashion.
title Map Space Belief Prediction for Manipulation-Enhanced Mapping
topic Robotics
Machine Learning
url https://arxiv.org/abs/2502.20606