GBPP: Grasp-Aware Base Placement Prediction for Robots via Two-Stage Learning

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Chen, Jizhuo, Liu, Diwen, Wang, Jiaming, Soh, Harold
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866916952012750848
author Chen, Jizhuo
Liu, Diwen
Wang, Jiaming
Soh, Harold
author_facet Chen, Jizhuo
Liu, Diwen
Wang, Jiaming
Soh, Harold
contents GBPP is a fast learning based scorer that selects a robot base pose for grasping from a single RGB-D snapshot. The method uses a two stage curriculum: (1) a simple distance-visibility rule auto-labels a large dataset at low cost; and (2) a smaller set of high fidelity simulation trials refines the model to match true grasp outcomes. A PointNet++ style point cloud encoder with an MLP scores dense grids of candidate poses, enabling rapid online selection without full task-and-motion optimization. In simulation and on a real mobile manipulator, GBPP outperforms proximity and geometry only baselines, choosing safer and more reachable stances and degrading gracefully when wrong. The results offer a practical recipe for data efficient, geometry aware base placement: use inexpensive heuristics for coverage, then calibrate with targeted simulation.
format Preprint
id arxiv_https___arxiv_org_abs_2509_11594
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle GBPP: Grasp-Aware Base Placement Prediction for Robots via Two-Stage Learning
Chen, Jizhuo
Liu, Diwen
Wang, Jiaming
Soh, Harold
Robotics
Artificial Intelligence
GBPP is a fast learning based scorer that selects a robot base pose for grasping from a single RGB-D snapshot. The method uses a two stage curriculum: (1) a simple distance-visibility rule auto-labels a large dataset at low cost; and (2) a smaller set of high fidelity simulation trials refines the model to match true grasp outcomes. A PointNet++ style point cloud encoder with an MLP scores dense grids of candidate poses, enabling rapid online selection without full task-and-motion optimization. In simulation and on a real mobile manipulator, GBPP outperforms proximity and geometry only baselines, choosing safer and more reachable stances and degrading gracefully when wrong. The results offer a practical recipe for data efficient, geometry aware base placement: use inexpensive heuristics for coverage, then calibrate with targeted simulation.
title GBPP: Grasp-Aware Base Placement Prediction for Robots via Two-Stage Learning
topic Robotics
Artificial Intelligence
url https://arxiv.org/abs/2509.11594