PhaForce: Phase-Scheduled Visual-Force Policy Learning with Slow Planning and Fast Correction for Contact-Rich Manipulation

TL;DR

PhaForce introduces a phase-scheduled visual-force policy that combines slow planning with fast correction, enabling more effective contact-rich manipulation by coordinating vision and force feedback across task phases.

Contribution

The paper presents a novel phase-scheduled policy with a contact-aware predictor, dual-gated visual-force fusion, and a fast residual corrector for improved contact-rich manipulation.

Findings

Achieves 86% success rate on real-robot tasks, outperforming baselines by 40 percentage points.

Improves contact quality by regulating interaction forces.

Demonstrates robustness to out-of-distribution geometric shifts.

Abstract

Contact-rich manipulation requires not only vision-dominant task semantics but also closed-loop reactions to force/torque (F/T) transients. Yet, generative visuomotor policies are typically constrained to low-frequency updates due to inference latency and action chunking, underutilizing F/T for control-rate feedback. Furthermore, existing force-aware methods often inject force continuously and indiscriminately, lacking an explicit mechanism to schedule when / how much / where to apply force across different task phases. We propose PhaForce, a phase-scheduled visual--force policy that coordinates low-rate chunk-level planning and high-rate residual correction via a unified contact/phase schedule. PhaForce comprises (i) a contact-aware phase predictor (CAP) that estimates contact probability and phase belief, (ii) a Slow diffusion planner that performs dual-gated visual--force fusion with orthogonal residual injection to preserve vision semantics while conditioning on force, and (iii) a Fast corrector that applies control-rate phase-routed residuals in interpretable corrective subspaces for within-chunk micro-adjustments. Across multiple real-robot contact-rich tasks, PhaForce achieves an average success rate of 86% (+40 pp over baselines), while also substantially improving contact quality by regulating interaction forces and exhibiting robust adaptability to OOD geometric shifts.