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

TL;DR

This paper formalizes a physically grounded interaction frame, an instantaneous local basis that decouples force regulation from motion execution, and proposes a method to recover it from demonstrations, ultimately improving both contact stability and execution quality.

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/

Figures (19)

• Figure 1: Interaction Frame for Example Contact-Rich Tasks.
• Figure 2: Force Policy and Dual-Policy Asynchronous Scheduler.(Left)Force Policy consists of a global vision policy and a local force policy. The global policy provides task-level visual context and global actions, while the local policy predicts interaction structure and local actions to realize hybrid force-position control during contact. (Right) The dual-policy asynchronous scheduler switches between the two policies and reduces latency and jerk via model-agnostic chunk alignment using dynamic time warping (DTW) dtw.
• Figure 3: Tasks. We design three tasks spanning two categories (polishing and insertion) to evaluate different policies for contact-rich manipulation. The descriptions on the right highlight the key challenges of each task compared to similar tasks in prior literature. All tasks require highly accurate force regulation to be successfully completed. We randomize object placement within the workspace area during both data collection and evaluation for each task.
• Figure 3: Force Regulation Evaluation on the Push and Flip Task.(Left) Pushing the heavy object requires approximately 45N, while demonstrations apply about 15N pushing force to flip the target object; we then measure its pushed distance $d$ as an indicator of force regulation. (Right) Statistics of the pushed distance for each method.
• Figure 4: Visualization of Effective Forces during Deployment and from Demonstrations on the Scrape off Sticker (Hard) Task. All baselines fail to replicate the force behavior from demonstrations, resulting in degraded performance. Force Policy closely imitates the effective force in demonstrations, achieving higher success rates.

Theorems & Definitions (5)

• Theorem 1: Unified Spatial Basis
• Corollary 1: Spectral-Geometric Isomorphism
• Proposition 1: Intent Alignment
• Theorem 2: Co-axial Twist to an Applied Wrench
• Proposition 2: Co-axial Wrench to an Applied Twist