ForceVLA2: Unleashing Hybrid Force-Position Control with Force Awareness for Contact-Rich Manipulation

2026-03-16

AI Summary

This work addresses the limitations of existing contact-intensive manipulation approaches, which predominantly rely on position control and lack explicit perception and regulation of interaction forces, resulting in insufficient stability, precision, and robustness. To overcome these challenges, we propose the first end-to-end vision-language-action framework that integrates explicit force-aware tasks. Our method dynamically fuses multimodal inputs—including visual observations, language instructions, proprioceptive states, and force signals—through force-aware prompting and a cross-scale Mixture-of-Experts (MoE) mechanism, enabling closed-loop hybrid force-position control. Evaluated across five contact-rich manipulation tasks, the proposed approach achieves success rates 48.0% and 35.0% higher than those of Pi0 and Pi0.5, respectively, while significantly mitigating common failure modes such as robotic arm overload and unstable contact interactions.

Technology Category

• Intelligent Robots: Manipulation
• Computer Vision: Multi-modal Vision
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Embodied intelligence for contact-rich manipulation has predominantly relied on position control, while explicit awareness and regulation of interaction forces remain under-explored, limiting stability, precision, and robustness in real-world tasks. We propose ForceVLA2, an end-to-end vision-language-action framework that equips robots with hybrid force-position control and explicit force awareness. ForceVLA2 introduces force-based prompts into the VLM expert to construct force-aware task concepts across stages, and employs a Cross-Scale Mixture-of-Experts (MoE) in the action expert to adaptively fuse these concepts with real-time interaction forces for closed-loop hybrid force-position regulation. To support learning and evaluation, we construct ForceVLA2-Dataset, containing 1,000 trajectories over 5 contact-rich tasks, including wiping, pressing, and assembling, with multi-view images, task prompts, proprioceptive state, and force signals. Extensive experiments show that ForceVLA2 substantially improves success rates and reliability in contact-rich manipulation, outperforming pi0 and pi0.5 by 48.0% and 35.0%, respectively, across the 5 tasks, and mitigating common failure modes such as arm overload and unstable contact, thereby actively advancing force-aware interactive physical intelligence in VLAs. The project page is available at https://sites.google.com/view/force-vla2/home.

Problem

• contact-rich manipulation
• force awareness
• hybrid force-position control
• embodied intelligence
• interaction forces

Innovation

• force-aware manipulation
• hybrid force-position control
• vision-language-action (VLA)
• Mixture-of-Experts (MoE)
• contact-rich tasks