Modality-Augmented Fine-Tuning of Foundation Robot Policies for Cross-Embodiment Manipulation on GR1 and G1
This work addresses cross-embodiment manipulation for humanoid robots, presenting an incremental improvement through targeted modality design and multi-modal fine-tuning.
This paper tackles the problem of adapting foundation robot policies to different humanoid embodiments by proposing a modality-augmented fine-tuning framework, which improves success rates from 51% to 63% on GR1 and from 48% to 94% on G1 compared to baselines.
This paper presents a modality-augmented fine-tuning framework designed to adapt foundation robot policies to diverse humanoid embodiments. We validate our approach across two distinct settings: (i) the GR1 embodiment, utilizing public datasets where we introduce post-processed modalities, including binary contact signals and ZoeDepth-generated metric depth; and (ii) the Unitree G1 embodiment, for which we contribute a novel multi-modal dataset incorporating cuRobo motion planning, inverse kinematics, and ground-truth contact-force measurements. Our experiments demonstrate that modality augmentation consistently enhances policy performance across different embodiments. Specifically, for the GR1, integrating contact-state cues and RGB-D fusion improves online success rates from 51% to 63%. Furthermore, in the G1 "Pick Apple to Bowl" task, our contact-augmented model achieves a success rate of 94%, significantly outperforming the 48% achieved by standard fine-tuning and the 0% baseline of zero-shot transfer. These results highlight that lightweight post-processing effectively strengthens policies for GR1, while high-quality multi-modal data is crucial for reliable transfer to the Unitree G1. Consequently, this work establishes a unified, data-centric pathway for extending foundation robot policies through targeted modality design and multi-modal fine-tuning.