TA-VLA: Elucidating the Design Space of Torque-aware Vision-Language-Action Models
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866908528120168448 |
|---|---|
| author | Zhang, Zongzheng Xu, Haobo Yang, Zhuo Yue, Chenghao Lin, Zehao Gao, Huan-ang Wang, Ziwei Zhao, Hao |
| author_facet | Zhang, Zongzheng Xu, Haobo Yang, Zhuo Yue, Chenghao Lin, Zehao Gao, Huan-ang Wang, Ziwei Zhao, Hao |
| contents | Many robotic manipulation tasks require sensing and responding to force signals such as torque to assess whether the task has been successfully completed and to enable closed-loop control. However, current Vision-Language-Action (VLA) models lack the ability to integrate such subtle physical feedback. In this work, we explore Torque-aware VLA models, aiming to bridge this gap by systematically studying the design space for incorporating torque signals into existing VLA architectures. We identify and evaluate several strategies, leading to three key findings. First, introducing torque adapters into the decoder consistently outperforms inserting them into the encoder.Third, inspired by joint prediction and planning paradigms in autonomous driving, we propose predicting torque as an auxiliary output, which further improves performance. This strategy encourages the model to build a physically grounded internal representation of interaction dynamics. Extensive quantitative and qualitative experiments across contact-rich manipulation benchmarks validate our findings. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_07962 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | TA-VLA: Elucidating the Design Space of Torque-aware Vision-Language-Action Models Zhang, Zongzheng Xu, Haobo Yang, Zhuo Yue, Chenghao Lin, Zehao Gao, Huan-ang Wang, Ziwei Zhao, Hao Robotics Many robotic manipulation tasks require sensing and responding to force signals such as torque to assess whether the task has been successfully completed and to enable closed-loop control. However, current Vision-Language-Action (VLA) models lack the ability to integrate such subtle physical feedback. In this work, we explore Torque-aware VLA models, aiming to bridge this gap by systematically studying the design space for incorporating torque signals into existing VLA architectures. We identify and evaluate several strategies, leading to three key findings. First, introducing torque adapters into the decoder consistently outperforms inserting them into the encoder.Third, inspired by joint prediction and planning paradigms in autonomous driving, we propose predicting torque as an auxiliary output, which further improves performance. This strategy encourages the model to build a physically grounded internal representation of interaction dynamics. Extensive quantitative and qualitative experiments across contact-rich manipulation benchmarks validate our findings. |
| title | TA-VLA: Elucidating the Design Space of Torque-aware Vision-Language-Action Models |
| topic | Robotics |
| url | https://arxiv.org/abs/2509.07962 |