ForceVLA: Enhancing VLA Models with a Force-aware MoE for Contact-rich Manipulation

Cewu Lu (Shanghai Jiao Tong University) · Panpan Cai (Shanghai Jiaotong University) · Wenqiang Zhang (Fudan University) · Yan Song (University College London, University of London) · Jiawen Yu (Fudan University) · Hairuo Liu (Shanghai Jiaotong University) · Qiaojun Yu (Shanghai Jiaotong University) · Jieji Ren (Shanghai Jiaotong University) · Ce Hao (National University of Singapore) · Haitong Ding (Shanghai University) · Guangyu Huang (Xi'an Jiaotong University) · Guofan Huang (Fudan University)
6-axis force feedbackaction decodingcontact dynamicscontext-aware routingdexterous manipulationforce-aware mixture-of-expertsforce-torque signalsforcevla-datamodality-specific expertsphysically intelligent robotic controlpretrained visual-language embeddingsproprioceptionrobotic manipulationsynchronized visionvision-language-action

Vision-Language-Action (VLA) models have advanced general-purpose robotic manipulation by leveraging pretrained visual and linguistic representations. However, they struggle with contact-rich tasks that require fine-grained control involving force, especially under visual occlusion or dynamic uncertainty. To address these limitations, we propose \textbf{ForceVLA}, a novel end-to-end manipulation framework that treats external force sensing as a first-class modality within VLA systems. ForceVLA introduces \textbf{FVLMoE}, a force-aware Mixture-of-Experts fusion module that dynamically integrates pretrained visual-language embeddings with real-time 6-axis force feedback during action decoding. This enables context-aware routing across modality-specific experts, enhancing the robot's ability to adapt to subtle contact dynamics. We also introduce \textbf{ForceVLA-Data}, a new dataset comprising synchronized vision, proprioception, and force-torque signals across five contact-rich manipulation tasks. ForceVLA improves average task success by 23.2\% over strong $\pi_0$-based baselines, achieving up to 80\% success in tasks such as plug insertion. Our approach highlights the importance of multimodal integration for dexterous manipulation and sets a new benchmark for physically intelligent robotic control. Code and data will be released at https://sites.google.com/view/forcevla2025/.