Contact-Aware Imitation Learning Through Contact Factorization

2026-10-07

AI Summary

This study addresses the problem of force signal fluctuations and policy generalization failures in contact-rich manipulation caused by surface geometry, normal, and friction perturbations. To this end, we propose FACE (Factorized Contact), a framework that introduces a novel contact-factorized imitation learning paradigm to decouple task intent from environmental contact dynamics. Specifically, FACE infers local contact parameters through a learned normal estimator and an online friction estimator, and encodes force observations within a normalized relative coordinate frame, thereby enabling execution-time adaptation without parameter updates. Real-robot experiments demonstrate that the proposed framework exhibits superior zero-shot robust generalization under unseen surface properties and geometric variations.

Abstract

Generalizable contact-rich manipulation requires robots to preserve intended task behavior while adapting its physical realization to changing contact conditions. However, interaction forces can vary substantially with small changes in surface geometry, orientation, and friction, making policies trained directly on raw force measurements difficult to transfer beyond demonstrated conditions. We introduce FACE, a contact-factorized imitation learning framework that separates intended task behavior from environment-dependent contact factors. Our representation expresses interaction forces in normalized, contact-relative coordinates, while a learned contact-normal estimator and an online friction estimator infer the local contact normal and effective friction scale. Together, these estimators enable force observations to be encoded and policy outputs to be decoded into physical motion and force commands during execution. In this way, FACE adapts execution to current contact conditions while preserving the intended task behavior, without updating the policy parameters. We evaluate FACE on real-robot contact-rich manipulation under unseen variations in surface properties and geometry, demonstrating robust generalization across contact conditions through controlled comparisons with variants that adapt prior approaches to our setting. Videos and additional materials can be found on the project page: https://rcilab.khu.ac.kr/face.

Problem

• imitation learning
• contact-rich manipulation
• generalization
• interaction forces
• policy transfer

Innovation

• Imitation Learning
• Contact Factorization
• Contact-rich Manipulation
• Generalization
• Force Representation