Force Policy: Learning Hybrid Force-Position Control Policy under Interaction Frame for Contact-Rich Manipulation

2026-02-25

AI Summary

This work addresses the challenge in contact-rich manipulation tasks where existing methods struggle to simultaneously satisfy global task objectives guided by vision and local stability driven by force feedback. The authors propose Force Policy, a novel approach that decouples the control strategy into two components: global visual guidance and local high-frequency force-position hybrid control. This decoupling is enabled by an instantaneous interaction frame recoverable from demonstrations, which is dynamically estimated during contact to ensure stable interaction. Crucially, the method learns not only control parameters but also the underlying control structure itself. Experimental results demonstrate that Force Policy significantly outperforms strong baselines across diverse real-world contact tasks, achieving enhanced robustness in contact establishment, improved force regulation accuracy, and reliable generalization to novel objects with varying geometries and physical properties.

Technology Category

• Intelligent Robots: Manipulation
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Computer Vision: Low Level & Physics-based Vision

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
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems

Abstract

Contact-rich manipulation demands human-like integration of perception and force feedback: vision should guide task progress, while high-frequency interaction control must stabilize contact under uncertainty. Existing learning-based policies often entangle these roles in a monolithic network, trading off global generalization against stable local refinement, while control-centric approaches typically assume a known task structure or learn only controller parameters rather than the structure itself. In this paper, we formalize a physically grounded interaction frame, an instantaneous local basis that decouples force regulation from motion execution, and propose a method to recover it from demonstrations. Based on this, we address both issues by proposing Force Policy, a global-local vision-force policy in which a global policy guides free-space actions using vision, and upon contact, a high-frequency local policy with force feedback estimates the interaction frame and executes hybrid force-position control for stable interaction. Real-world experiments across diverse contact-rich tasks show consistent gains over strong baselines, with more robust contact establishment, more accurate force regulation, and reliable generalization to novel objects with varied geometries and physical properties, ultimately improving both contact stability and execution quality. Project page: https://force-policy.github.io/

Problem

• contact-rich manipulation
• force-position control
• interaction frame
• vision-force integration
• policy learning

Innovation

• interaction frame
• hybrid force-position control
• contact-rich manipulation
• vision-force policy
• demonstration-based learning