MobileVLA-R1 2.0: RL-Enhanced Reasoning for Mobile Robot Control

Hugging Face Daily Papers Papers

Summary

MobileVLA-R1 2.0 is an RL-enhanced vision-language-action framework that couples structured reasoning with mobile robot control, achieving improvements in long-horizon instruction following on real-world platforms.

Grounding natural-language instructions into reliable and executable actions remains a fundamental challenge for vision-language-action (VLA) systems on mobile robots, due to the persistent gap between high-level semantic reasoning and low-level locomotion and manipulation control. Existing approaches often rely on implicit reasoning or monolithic action prediction, making it difficult to maintain coherent long-horizon decision making while producing precise and adaptable robot actions. To address this challenge, we propose MobileVLA-R1 2.0, an RL-enhanced VLA framework that explicitly couples structured embodied reasoning with executable mobile robot control. The framework learns multi-granularity reasoning over embodied trajectories through supervised Chain-of-Thought (CoT) alignment and reinforcement learning, improving reasoning-to-action consistency beyond purely behavioral supervision. To support both locomotion and manipulation, we further introduce a reasoning-conditioned action decoder that maps multimodal reasoning representations to task-level action targets, which are subsequently translated into embodiment-specific commands by robot controllers. This design provides a unified perception-reasoning-action interface while decoupling high-level action generation from robot-specific actuation. We conduct extensive evaluations on language-guided navigation, quadruped control, and humanoid mobile manipulation, covering VLN-CE, QUARD, and real-world deployments on Unitree Go2 and G1 robots. MobileVLA-R1 2.0 consistently outperforms strong VLA baselines, achieving an average 1.6 point improvement in SR on VLN-CE and a 10.0 point improvement in full-task success on real-world G1 mobile manipulation tasks over MobileVLA-R1, while demonstrating robust long-horizon instruction following and closed-loop execution across different robotic platforms.
Original Article
View Cached Full Text

Cached at: 09/15/26, 02:39 AM

Paper page - MobileVLA-R1 2.0: RL-Enhanced Reasoning for Mobile Robot Control

Source: https://huggingface.co/papers/2609.06251

Abstract

MobileVLA-R1 2.0 couples structured reasoning with robot control via chain-of-thought alignment and reinforcement learning to improve long-horizon mobile robot instruction following.

Grounding natural-language instructions into reliable and executable actions remains a fundamental challenge forvision-language-action(VLA) systems on mobile robots, due to the persistent gap between high-level semantic reasoning and low-level locomotion and manipulation control. Existing approaches often rely on implicit reasoning or monolithic action prediction, making it difficult to maintain coherent long-horizon decision making while producing precise and adaptable robot actions. To address this challenge, we propose MobileVLA-R1 2.0, an RL-enhanced VLA framework that explicitly couples structuredembodied reasoningwith executablemobile robot control. The framework learns multi-granularity reasoning over embodied trajectories through supervisedChain-of-Thought(CoT) alignment andreinforcement learning, improving reasoning-to-action consistency beyond purely behavioral supervision. To support both locomotion and manipulation, we further introduce areasoning-conditioned action decoderthat mapsmultimodal reasoningrepresentations to task-level action targets, which are subsequently translated into embodiment-specific commands by robot controllers. This design provides a unified perception-reasoning-action interface while decoupling high-level action generation from robot-specific actuation. We conduct extensive evaluations on language-guided navigation, quadruped control, and humanoid mobile manipulation, covering VLN-CE, QUARD, and real-world deployments on Unitree Go2 and G1 robots. MobileVLA-R1 2.0 consistently outperforms strong VLA baselines, achieving an average 1.6 point improvement in SR on VLN-CE and a 10.0 point improvement in full-task success on real-world G1 mobile manipulation tasks over MobileVLA-R1, while demonstrating robust long-horizon instruction following and closed-loop execution across different robotic platforms.

View arXiv pageView PDFProject pageGitHub6Add to collection

Get this paper in your agent:

hf papers read 2609\.06251

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2609.06251 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2609.06251 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2609.06251 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

From Foundation to Application: Improving VLA Models in Practice

Papers with Code Trending

This paper presents LingBot-VLA 2.0, which enhances VLA foundation models for robotics by improving generalization across tasks and embodiments, expanding action space to whole-body degrees of freedom, and incorporating predictive dynamics modeling for better temporal reasoning.