Whole-Body Mobile Manipulation using Offline Reinforcement Learning on Sub-optimal Controllers

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Jauhri, Snehal, Prasad, Vignesh, Chalvatzaki, Georgia
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915936501497856
author Jauhri, Snehal
Prasad, Vignesh
Chalvatzaki, Georgia
author_facet Jauhri, Snehal
Prasad, Vignesh
Chalvatzaki, Georgia
contents Mobile Manipulation (MoMa) of articulated objects, such as opening doors, drawers, and cupboards, demands simultaneous, whole-body coordination between a robot's base and arms. Classical whole-body controllers (WBCs) can solve such problems via hierarchical optimization, but require extensive hand-tuned optimization and remain brittle. Learning-based methods, on the other hand, show strong generalization capabilities but typically rely on expensive whole-body teleoperation data or heavy reward engineering. We observe that even a sub-optimal WBC is a powerful structural prior: it can be used to collect data in a constrained, task-relevant region of the state-action space, and its behavior can still be improved upon using offline reinforcement learning. Building on this, we propose WHOLE-MoMa, a two-stage pipeline that first generates diverse demonstrations by randomizing a lightweight WBC, and then applies offline RL to identify and stitch together improved behaviors via a reward signal. To support the expressive action-chunked diffusion policies needed for complex coordination tasks, we extend offline implicit Q-learning with Q-chunking for chunk-level critic evaluation and advantage-weighted policy extraction. On three tasks of increasing difficulty using a TIAGo++ mobile manipulator in simulation, WHOLE-MoMa significantly outperforms WBC, behavior cloning, and several offline RL baselines. Policies transfer directly to the real robot without finetuning, achieving 80% success in bimanual drawer manipulation and 68% in simultaneous cupboard opening and object placement, all without any teleoperated or real-world training data.
format Preprint
id arxiv_https___arxiv_org_abs_2604_12509
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Whole-Body Mobile Manipulation using Offline Reinforcement Learning on Sub-optimal Controllers
Jauhri, Snehal
Prasad, Vignesh
Chalvatzaki, Georgia
Robotics
Computer Vision and Pattern Recognition
Mobile Manipulation (MoMa) of articulated objects, such as opening doors, drawers, and cupboards, demands simultaneous, whole-body coordination between a robot's base and arms. Classical whole-body controllers (WBCs) can solve such problems via hierarchical optimization, but require extensive hand-tuned optimization and remain brittle. Learning-based methods, on the other hand, show strong generalization capabilities but typically rely on expensive whole-body teleoperation data or heavy reward engineering. We observe that even a sub-optimal WBC is a powerful structural prior: it can be used to collect data in a constrained, task-relevant region of the state-action space, and its behavior can still be improved upon using offline reinforcement learning. Building on this, we propose WHOLE-MoMa, a two-stage pipeline that first generates diverse demonstrations by randomizing a lightweight WBC, and then applies offline RL to identify and stitch together improved behaviors via a reward signal. To support the expressive action-chunked diffusion policies needed for complex coordination tasks, we extend offline implicit Q-learning with Q-chunking for chunk-level critic evaluation and advantage-weighted policy extraction. On three tasks of increasing difficulty using a TIAGo++ mobile manipulator in simulation, WHOLE-MoMa significantly outperforms WBC, behavior cloning, and several offline RL baselines. Policies transfer directly to the real robot without finetuning, achieving 80% success in bimanual drawer manipulation and 68% in simultaneous cupboard opening and object placement, all without any teleoperated or real-world training data.
title Whole-Body Mobile Manipulation using Offline Reinforcement Learning on Sub-optimal Controllers
topic Robotics
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2604.12509