arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2603.08342cs.RO
收藏

PhaForce: 基于慢规划与快修正的相位调度视觉-力政策学习用于接触密集 manipulation

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

Mingxin Wang, Zhirun Yue, Renhao Lu, Yizhe Li, Zihan Wang, Guoping Pan, Kangkang Dong, Jun Cheng, Yi Cheng, Houde Liu

AI总结:

PhaForce 通过慢规划与快修正机制,实现接触密集 manipulation 中力与视觉的协同控制,提升任务成功率与接触质量。

摘要:

接触密集 manipulation 需要不仅视觉主导的任务语义,还需要闭环反应到力/扭矩(F/T)瞬态。然而,生成性视觉-运动政策通常受限于低频更新,因为推理延迟和动作分块,未能充分利用 F/T 作为控制率反馈。此外,现有力感知方法往往连续注入力,并且无差别地应用,缺乏一个明确的机制来调度在不同任务阶段何时/多少/在哪里应用力。我们提出 PhaForce,一种相位调度的视觉-力政策,通过统一的接触/相位计划协调低频分块级规划和高频残差修正。PhaForce 包括(i)一个接触感知的相位预测器(CAP)估计接触概率和相位信念,(ii)一个慢扩散规划器,通过双门视觉-力融合和正交残差注入进行双重门控,以保留视觉语义同时基于力进行条件化,以及(iii)一个快速校正器,将控制率相位路由残差应用于可解释的校正子空间中,以进行分块内的微调。在多个真实的机器人接触密集任务中,PhaForce 实现了 86% 的平均成功率(比基线高 40 个百分点),同时显著提高了接触质量,通过调节交互力并表现出对 OOD 几何偏移的鲁棒适应性。

英文摘要:

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.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑